channel_count
int64
1
1
characteristics_ch1
stringlengths
6
9.77k
contact_address
stringlengths
1
229
contact_city
stringlengths
1
47
contact_country
stringclasses
71 values
contact_institute
stringlengths
2
128
contact_name
stringlengths
4
74
contact_zip
stringclasses
1 value
data_processing
stringlengths
4
10.6k
extract_protocol_ch1
stringlengths
6
14.3k
geo_accession
stringlengths
9
10
instrument_model
stringclasses
13 values
last_update_date
stringlengths
11
11
library_selection
stringclasses
6 values
library_source
stringclasses
5 values
library_strategy
stringclasses
1 value
molecule_ch1
stringclasses
7 values
organism_ch1
stringclasses
1 value
platform_id
stringclasses
13 values
relation
stringlengths
0
2.61k
sample
stringlengths
9
10
series_id
stringlengths
8
29
singlecellprobability
float64
0
1
source_name_ch1
stringlengths
1
253
status
stringlengths
21
21
submission_date
stringlengths
11
11
taxid_ch1
stringclasses
1 value
title
stringlengths
1
120
type
stringclasses
1 value
sra
stringlengths
47
49
biosample
stringlengths
51
51
1
cell type: pancreatic progenitors cells
82 Walnut St. Unit 3
Brookline
USA
Harvard Medical School
Yuting,,Liu
All obtained reads from each sample were mapped against the human genome (hg19 build) with Bowtie/Tophat v2.0.2 which allows mapping across splice sites by reads segmentation (Trapnell et al., 2012). All programs were performed with default setting (unless otherwise specified).,The unique mapped reads (69-91% of total reads) were subsequently assembled into transcripts guided by reference annotation (Gencode v14 and Refseq gene models) with Cufflinks v2.0.2 (Trapnell et al., 2012). Expression level of each gene was quantified with normalized FPKM (fragments per kilobase of exon per million mapped fragments).,For the analyses of lncRNAs, cufflinks was used to estimate the transcript level FPKM values with the Gencode v14 lincRNA annotation (Derrien et al., 2012), and human body map lincRNA annotations (Cabili et al., 2011).,processed data file for protein-coding genes:,processed_FPKM_genes_human_pancreatic_lineage_140112.txt,processed data file for lncRNAs:,processed_FPKM_lncRNAs_human_pancreatic_lineage_140112.txt,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include FPKM values for each sample.
Total RNA was purified from 1,000 sorted cells using ZR RNA microprep kit (Zymo Research, UAS). The cDNA synthesis and amplification was performed with the SMARTer ultra-low input RNA kit (Clontech, US).,The amplified cDNA (10-40 ng) was then fragmented by Covaris S2 sonicator (Covaris, US) and converted to sequencing libraries following the Illumina’s construction protocol for low input DNA (Illumina, US). Barcoded libraries were pooled and sequenced in Illumina Hiseq 2000 instrument (pair-end 100bp).
GSM1093232
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248487,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974785
GSM1093232
GSE44875
0.558144
pancreatic progenitors cells
Public on Jun 01 2014
Mar 05 2013
9606
PP1
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248487
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974785
1
cell type: pancreatic progenitors cells
82 Walnut St. Unit 3
Brookline
USA
Harvard Medical School
Yuting,,Liu
All obtained reads from each sample were mapped against the human genome (hg19 build) with Bowtie/Tophat v2.0.2 which allows mapping across splice sites by reads segmentation (Trapnell et al., 2012). All programs were performed with default setting (unless otherwise specified).,The unique mapped reads (69-91% of total reads) were subsequently assembled into transcripts guided by reference annotation (Gencode v14 and Refseq gene models) with Cufflinks v2.0.2 (Trapnell et al., 2012). Expression level of each gene was quantified with normalized FPKM (fragments per kilobase of exon per million mapped fragments).,For the analyses of lncRNAs, cufflinks was used to estimate the transcript level FPKM values with the Gencode v14 lincRNA annotation (Derrien et al., 2012), and human body map lincRNA annotations (Cabili et al., 2011).,processed data file for protein-coding genes:,processed_FPKM_genes_human_pancreatic_lineage_140112.txt,processed data file for lncRNAs:,processed_FPKM_lncRNAs_human_pancreatic_lineage_140112.txt,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include FPKM values for each sample.
Total RNA was purified from 1,000 sorted cells using ZR RNA microprep kit (Zymo Research, UAS). The cDNA synthesis and amplification was performed with the SMARTer ultra-low input RNA kit (Clontech, US).,The amplified cDNA (10-40 ng) was then fragmented by Covaris S2 sonicator (Covaris, US) and converted to sequencing libraries following the Illumina’s construction protocol for low input DNA (Illumina, US). Barcoded libraries were pooled and sequenced in Illumina Hiseq 2000 instrument (pair-end 100bp).
GSM1093233
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248488,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974786
GSM1093233
GSE44875
0.15643
pancreatic progenitors cells
Public on Jun 01 2014
Mar 05 2013
9606
PP2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248488
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974786
1
cell type: human pancreatic alpha cells
82 Walnut St. Unit 3
Brookline
USA
Harvard Medical School
Yuting,,Liu
All obtained reads from each sample were mapped against the human genome (hg19 build) with Bowtie/Tophat v2.0.2 which allows mapping across splice sites by reads segmentation (Trapnell et al., 2012). All programs were performed with default setting (unless otherwise specified).,The unique mapped reads (69-91% of total reads) were subsequently assembled into transcripts guided by reference annotation (Gencode v14 and Refseq gene models) with Cufflinks v2.0.2 (Trapnell et al., 2012). Expression level of each gene was quantified with normalized FPKM (fragments per kilobase of exon per million mapped fragments).,For the analyses of lncRNAs, cufflinks was used to estimate the transcript level FPKM values with the Gencode v14 lincRNA annotation (Derrien et al., 2012), and human body map lincRNA annotations (Cabili et al., 2011).,processed data file for protein-coding genes:,processed_FPKM_genes_human_pancreatic_lineage_140112.txt,processed data file for lncRNAs:,processed_FPKM_lncRNAs_human_pancreatic_lineage_140112.txt,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include FPKM values for each sample.
Total RNA was purified from 1,000 sorted cells using ZR RNA microprep kit (Zymo Research, UAS). The cDNA synthesis and amplification was performed with the SMARTer ultra-low input RNA kit (Clontech, US).,The amplified cDNA (10-40 ng) was then fragmented by Covaris S2 sonicator (Covaris, US) and converted to sequencing libraries following the Illumina’s construction protocol for low input DNA (Illumina, US). Barcoded libraries were pooled and sequenced in Illumina Hiseq 2000 instrument (pair-end 100bp).
GSM1093235
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248490,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974788
GSM1093235
GSE44875
0.527332
human pancreatic alpha cells
Public on Jun 01 2014
Mar 05 2013
9606
alpha1
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248490
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974788
1
cell type: human pancreatic alpha cells
82 Walnut St. Unit 3
Brookline
USA
Harvard Medical School
Yuting,,Liu
All obtained reads from each sample were mapped against the human genome (hg19 build) with Bowtie/Tophat v2.0.2 which allows mapping across splice sites by reads segmentation (Trapnell et al., 2012). All programs were performed with default setting (unless otherwise specified).,The unique mapped reads (69-91% of total reads) were subsequently assembled into transcripts guided by reference annotation (Gencode v14 and Refseq gene models) with Cufflinks v2.0.2 (Trapnell et al., 2012). Expression level of each gene was quantified with normalized FPKM (fragments per kilobase of exon per million mapped fragments).,For the analyses of lncRNAs, cufflinks was used to estimate the transcript level FPKM values with the Gencode v14 lincRNA annotation (Derrien et al., 2012), and human body map lincRNA annotations (Cabili et al., 2011).,processed data file for protein-coding genes:,processed_FPKM_genes_human_pancreatic_lineage_140112.txt,processed data file for lncRNAs:,processed_FPKM_lncRNAs_human_pancreatic_lineage_140112.txt,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include FPKM values for each sample.
Total RNA was purified from 1,000 sorted cells using ZR RNA microprep kit (Zymo Research, UAS). The cDNA synthesis and amplification was performed with the SMARTer ultra-low input RNA kit (Clontech, US).,The amplified cDNA (10-40 ng) was then fragmented by Covaris S2 sonicator (Covaris, US) and converted to sequencing libraries following the Illumina’s construction protocol for low input DNA (Illumina, US). Barcoded libraries were pooled and sequenced in Illumina Hiseq 2000 instrument (pair-end 100bp).
GSM1093236
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248491,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974789
GSM1093236
GSE44875
0.734671
human pancreatic alpha cells
Public on Jun 01 2014
Mar 05 2013
9606
alpha2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248491
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974789
1
cell type: human pancreatic beta cells
82 Walnut St. Unit 3
Brookline
USA
Harvard Medical School
Yuting,,Liu
All obtained reads from each sample were mapped against the human genome (hg19 build) with Bowtie/Tophat v2.0.2 which allows mapping across splice sites by reads segmentation (Trapnell et al., 2012). All programs were performed with default setting (unless otherwise specified).,The unique mapped reads (69-91% of total reads) were subsequently assembled into transcripts guided by reference annotation (Gencode v14 and Refseq gene models) with Cufflinks v2.0.2 (Trapnell et al., 2012). Expression level of each gene was quantified with normalized FPKM (fragments per kilobase of exon per million mapped fragments).,For the analyses of lncRNAs, cufflinks was used to estimate the transcript level FPKM values with the Gencode v14 lincRNA annotation (Derrien et al., 2012), and human body map lincRNA annotations (Cabili et al., 2011).,processed data file for protein-coding genes:,processed_FPKM_genes_human_pancreatic_lineage_140112.txt,processed data file for lncRNAs:,processed_FPKM_lncRNAs_human_pancreatic_lineage_140112.txt,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include FPKM values for each sample.
Total RNA was purified from 1,000 sorted cells using ZR RNA microprep kit (Zymo Research, UAS). The cDNA synthesis and amplification was performed with the SMARTer ultra-low input RNA kit (Clontech, US).,The amplified cDNA (10-40 ng) was then fragmented by Covaris S2 sonicator (Covaris, US) and converted to sequencing libraries following the Illumina’s construction protocol for low input DNA (Illumina, US). Barcoded libraries were pooled and sequenced in Illumina Hiseq 2000 instrument (pair-end 100bp).
GSM1093237
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248492,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974790
GSM1093237
GSE44875
0.519434
human pancreatic beta cells
Public on Jun 01 2014
Mar 05 2013
9606
beta1
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248492
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974790
1
cell type: human pancreatic beta cells
82 Walnut St. Unit 3
Brookline
USA
Harvard Medical School
Yuting,,Liu
All obtained reads from each sample were mapped against the human genome (hg19 build) with Bowtie/Tophat v2.0.2 which allows mapping across splice sites by reads segmentation (Trapnell et al., 2012). All programs were performed with default setting (unless otherwise specified).,The unique mapped reads (69-91% of total reads) were subsequently assembled into transcripts guided by reference annotation (Gencode v14 and Refseq gene models) with Cufflinks v2.0.2 (Trapnell et al., 2012). Expression level of each gene was quantified with normalized FPKM (fragments per kilobase of exon per million mapped fragments).,For the analyses of lncRNAs, cufflinks was used to estimate the transcript level FPKM values with the Gencode v14 lincRNA annotation (Derrien et al., 2012), and human body map lincRNA annotations (Cabili et al., 2011).,processed data file for protein-coding genes:,processed_FPKM_genes_human_pancreatic_lineage_140112.txt,processed data file for lncRNAs:,processed_FPKM_lncRNAs_human_pancreatic_lineage_140112.txt,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include FPKM values for each sample.
Total RNA was purified from 1,000 sorted cells using ZR RNA microprep kit (Zymo Research, UAS). The cDNA synthesis and amplification was performed with the SMARTer ultra-low input RNA kit (Clontech, US).,The amplified cDNA (10-40 ng) was then fragmented by Covaris S2 sonicator (Covaris, US) and converted to sequencing libraries following the Illumina’s construction protocol for low input DNA (Illumina, US). Barcoded libraries were pooled and sequenced in Illumina Hiseq 2000 instrument (pair-end 100bp).
GSM1093238
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248493,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974791
GSM1093238
GSE44875
0.422463
human pancreatic beta cells
Public on Jun 01 2014
Mar 05 2013
9606
beta2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248493
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974791
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095127
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248556,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974931
GSM1095127
GSE44976
0.035638
Ctrl_RBM10.KD.24h_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
Ctrl_RBM10.KD.24h_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248556
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974931
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,stable cell line: stable cell lines inducibly expressing FLAG/HA-tagged RBM10,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095131
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248560,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974939
GSM1095131
GSE44976
0.001174
Ctrl_RBM10.OE.1_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
Ctrl_RBM10.OE.1_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248560
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974939
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,stable cell line: stable cell lines inducibly expressing FLAG/HA-tagged RBM10,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095132
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248561,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974941
GSM1095132
GSE44976
0.063725
Ctrl_RBM10.OE.2_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
Ctrl_RBM10.OE.2_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248561
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974941
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095133
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248562,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974943
GSM1095133
GSE44976
0.060617
RBM10.KD.24h_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
RBM10.KD.24h_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248562
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974943
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095134
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248563,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974945
GSM1095134
GSE44976
0.003833
RBM10.KD.48h.1_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
RBM10.KD.48h.1_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248563
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974945
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095135
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248564,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974947
GSM1095135
GSE44976
0.037335
RBM10.KD.48h.2_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
RBM10.KD.48h.2_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248564
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974947
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095136
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248565,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974949
GSM1095136
GSE44976
0.014024
RBM10.KD.72h_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
RBM10.KD.72h_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248565
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974949
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,stable cell line: stable cell lines inducibly expressing FLAG/HA-tagged RBM10,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095137
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248566,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974951
GSM1095137
GSE44976
0.005288
RBM10.OE.1_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
RBM10.OE.1_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248566
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974951
1
cell type: human embryonic kidney (HEK) 293 T-REx Flp-In cell lines,stable cell line: stable cell lines inducibly expressing FLAG/HA-tagged RBM10,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095138
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248567,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974953
GSM1095138
GSE44976
0
RBM10.OE.2_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
RBM10.OE.2_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248567
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974953
1
cell type: lymphoblastoid cell lines (LCLs) derived from healthy control,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095139
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248568,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974955
GSM1095139
GSE44976
0.00176
Normal.2_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
Normal.2_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248568
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974955
1
cell type: lymphoblastoid cell lines (LCLs) derived from patient,rna subset: poly A RNA
Wiesenstraße 14
Gießen
Germany
Technische Hochschule MIttelhessen
Andreas,,Gogol-Döring
RNA-seq reads were mapped with at most two mismatches to the human genome reference and a set of sequences consisting of all possible junctions between the exons of each Refseq gene. For each exon E of a RefSeq gene, we computed the number of reads which could only be mapped to E or exon junctions containing E with an overlap of at least 6bp ("spliced-in" event), and the number of reads which could be mapped only to exon junctions skipping E and overlap with both exons by at least 6bp ("spliced-out" event).,processed data files format and content: The .tab.txt files contain the number of reads supporting splicing-in and splicing-out events of exons. Each row of the table represents a single exon of a RefSeq gene. The columns of the list contains the chromosome ("chrom"), the starting position of the exon ("begin", 0-based), the ending position of the exon ("end", 0 based position of the first base after the exon), the name of the gene ("name"), a number which uniquely labels the exon within the gene ("exon"), the definition of the strand ("strand", either "+" or "-"), the number of reads supporting a splicing-in event ("in"), and the number of reads supporting a splicing-out event ("out"). Note that the first 3 columns conform to the bed file format.,Base calling was done on the Illumina HiSeq 2000 machine using RTA 1.12. We kept only reads passing the filtering.,Genome_build: hg19
For RNA-Seq, total RNA was extracted using Trizol reagent and poly A RNA was isolated from 1 ug of total RNA for sequencing libraries construction.,RNA-Seq libraries were prepared following Illumina TruSeq RNA Sample Prep Kit protocol.
GSM1095141
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248570,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974959
GSM1095141
GSE44976
0.003275
Patient_RNA-Seq
Public on Sep 03 2013
Mar 08 2013
9606
Patient_RNA-Seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248570
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974959
1
cell line: CHL-1, DKMG, JURKAT,cancer subtype: CHL-1-melanoma, DKMG-GBM, JURKAT-TALL
269 Campus Drive (CCSR 3245A)
Stanford
USA
Stanford University
Craig,Patrick,Giacomini
Samples were mapped to the human genome and RefSeq transcriptome using ELAND.,A custom C# script was used to extract all mate pairs with one read mapping to the candidate rearranged gene (identified by breakpoint analysis) and the other read mapping to a separate genomic locus.,A separate C# script screened mate pairs for single reads mapping to potential exon-exon fusion junctions of nominated gene fusions.,Genome_build: hg18,Supplementary_files_format_and_content: Our pipeline outputs a single file (GeneFusions.txt) after all samples are run through our pipeline. This file is a tab delimited text file with the first column representing the gene fusion name, the second column includes the sample from which the gene fusion was identified, the third column includes the chromosome of the 5' partner gene, the fourth column includes the chromosome of the 3' partner gene, the fifth column includes the exon position of the gene fusion junction, and the sixth column includes the number of mate pairs supporting the specific gene fusion.
mRNA Seq-8 Sample Prep Kit (Illumina)
GSM1097789
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248166,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01974539
GSM1097789
GSE45133,GSE45137
0.001798
Combined human cancer cell lines (CHL-1, DKMG, JURKAT)
Public on May 31 2013
Mar 13 2013
9606
Multi_C
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248166
https://www.ncbi.nlm.nih.gov/biosample/SAMN01974539
1
cell type: CD34+ cells,rna subtype: messenger RNA
Level 2, Lowy Cancer Research Centre, University of New South Wales
Sydney
Australia
UNIVERSITY OF NEW SOUTH WALES
Jason,,Wong
Base-calling: Illumina Real Time Analysis, scoring metric v1.5,Raw reads were prefiltered to remove all reads that have 3 or more bases with a quality score of less than Q13,All provided cleaned data have had adaptors clipped except the small RNA-seq data which still contains 5' adaptor: 5'-TGGAATTCTCGTATGCCGTCTTCTGCTTG,Alignments were all made against hg19. The following aligners were used. ChIP-seq: bwa (version 0.6.1-r104) default parameters , mRNA-seq: tophat (version 1.3.1) -g 25 -r 20 --segment-length 45 -G "hg19_refseq.gtf" --solexa1.3-quals, smallRNA: bowtie (version 0.12.5) -v 2 -m 25 --solexa1.3-quals --best --strata.,Bed files generated from BAM files using bedtools bamtobed (version 2.15.0). For ChIP-seq files, each bed entry is extended to 200bp in length. For RNA-seq the split option was used.,Genome_build: hg19,Supplementary_files_format_and_content: Bed files contain coordinates for all aligned reads
Bone marrow (BM) samples were harvested from normal volunteers with informed consent in accordance with local ethical guidelines. The CD34+ fraction was obtained by magnetic bead separation using an automated CliniMACS cell separation system (Miltenyi Biotec, Cologne). Cell purity was assessed by flow cytometry and was ≥ 98%. ChIP assays were performed with 15 x 10^6 per condition. Total RNA extraction and purification was performed using miRNeasy mini kits (QIAGEN). The small RNA fraction (<200nt) was collected in a separate fraction from total RNA, as per manufacturer’s standard instructions. Total RNA was amplified using the Ovation RNA-seq system V2 (NuGEN) prior to sequencing.,For ChIP, cells were harvested and washed with phosphate-buffered saline (PBS), followed by incubation of 1% (w/v) formaldehyde for 10 minutes at room temperature. To terminate the cross-link, cells were incubated with 0.125M glycine for 5 minutes. Cells were washed with PBS and lysed on ice in Cell lysis buffer (10mM Tris [pH 8.0], 10mM NaCl, 0.2% NP-40) for 10 minutes to recover nuclei. After centrifugation at 1500xg for 5 minutes, nuclei were lysed in Nucleus lysis buffer (50 mM Tris, 10mM EDTA, 1% SDS [pH 8.0]) on ice for 10 minutes. The lysate was diluted in IP dilution buffer (20mM Tris [pH 8.0], 2mM EDTA, 150mM NaCl, 1% Triton-X100, 0.01% SDS) and sonicated (Settings: High, 30sec pulses) using BioRuptor® sonicator (Diagenode, Liège, Belgium) to yield an average fragmentation size of approximately 200bp. The chromatin was pre-cleared with 100?g rabbit IgG for 1 hour followed by incubation with 100?l protein-G-agarose (Roche Applied Science, Penzberg, Germany) for 2 hours. For each sample, 300?l pre-cleared chromatin was removed (input for the subsequent qRT-PCR analysis) and the remaining chromatin was aliquoted, and incubated with the indicated antibody for 18 hours at 4°C. To collect immune complexes, 50?l protein G-agarose was added to the chromatin and incubated for additional 2 hours at 4°C. Protein-G-agarose pellets were washed at 5000xg; twice with 500?l IP wash buffer 1 (20mM Tris [pH 8.0], 2mM EDTAm 50mM NaCl, 1% Triton-X100, 0.1% SDS), once with IP wash buffer 2 (10mM Tris [pH 8.0], 1mM EDTA, 0.25M LiCl, 1% NP-40, 1% Sodium deoxycholate) and twice with TE (10mM Tris, 1mM EDTA [pH8.0]). Immuno-precipitated chromatin was eluted in 300?l Elution Buffer (100mM NaHCO3, 1% SDS) and reverse-cross link was obtained by incubation with RNase A and NaCl (0.3M final concentration) at 67°C for 18 hours followed by treatment with Proteinase K at 45°C for 2 hours. Input DNA (pre-cleared chromatin) was treated with RNase A and Proteinase K simultaneously.,Library construction was performed as per Illumina protocol Rev A 11257047
GSM1097887
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX248938,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01975666
GSM1097887
GSE45144
0.001978
CD34+ cells
Public on Sep 03 2013
Mar 13 2013
9606
CD34_mRNA-seq
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX248938
https://www.ncbi.nlm.nih.gov/biosample/SAMN01975666
1
knockdown chemistry: Gapmer,knockdown target: Scramble,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098180
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249074,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977945
GSM1098180
GSE45157
0.004565
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_Control_Gapmer_1
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249074
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977945
1
knockdown chemistry: Gapmer,knockdown target: Scramble,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098181
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249075,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977946
GSM1098181
GSE45157
0.003822
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_Control_Gapmer_2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249075
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977946
1
knockdown chemistry: Gapmer,knockdown target: linc-FIRRE,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098183
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249077,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977948
GSM1098183
GSE45157
0.003421
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_FIRRE_Gapmer_2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249077
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977948
1
knockdown chemistry: Gapmer,knockdown target: linc-FIRRE,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098184
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249078,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977949
GSM1098184
GSE45157
0.004334
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_FIRRE_Gapmer_3
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249078
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977949
1
knockdown chemistry: Gapmer,knockdown target: linc-FIRRE,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098185
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249079,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977950
GSM1098185
GSE45157
0
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_FIRRE_Gapmer_4
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249079
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977950
1
knockdown chemistry: siRNA,knockdown target: Scramble,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098186
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249080,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977951
GSM1098186
GSE45157
0.003661
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_Scramble_siRNA_1
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249080
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977951
1
knockdown chemistry: siRNA,knockdown target: Scramble,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098187
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249081,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977952
GSM1098187
GSE45157
0.003158
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_Scramble_siRNA_2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249081
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977952
1
knockdown chemistry: siRNA,knockdown target: hnRNPU,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098188
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249082,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977953
GSM1098188
GSE45157
0.002428
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_hnRNPU_siRNA_1
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249082
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977953
1
knockdown chemistry: siRNA,knockdown target: hnRNPU,cell line: HeLa
733 N. Broadway Avenue
Baltimore
USA
Johns Hopkins University
Loyal,A,Goff
Basecalls were performed using CASAVA 1.8,Reads were aligned using Tophat2 with default options,Expression quantification and differential analysis was done using Cuffdiff2 using UCSC reference transcriptome .gtf files (mouse or human) and Tophat2 aligned reads as input and default options.,Cuffdiff output files were processed using cummeRbund for indexing, exploration, and visualization of results.,Genome_build: hg19
1ml of Trizol and 200μl of chloroform were added and the mix was centrifuged at 4°C at 13,000 rpm for 15 minutes. The aqueous layer was cleaned and DNase treated on RNeasy Mini columns.,Illumina TruSeq RNA
GSM1098189
Illumina HiSeq 2000
May 15 2019
cDNA
transcriptomic
RNA-Seq
polyA RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249083,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01977954
GSM1098189
GSE45157
0.003661
HeLa Cells
Public on Jun 01 2013
Mar 14 2013
9606
HeLa_hnRNPU_siRNA_2
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249083
https://www.ncbi.nlm.nih.gov/biosample/SAMN01977954
1
age: 59,tissue: adipose tissue,log10 body mass index: 1.436353766,log10 basal metabolic rate (kcal): 1830,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.352532716,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.635899375,plasma free fatty acids under the curve ogtt (mmol/l * min): 23.7,fat mass (%): 22,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.491853096,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 132,log10 homair (insulin resistance index based on homa): 0.499381564,log10 homais (insulin secretion index based on homa): 2.316269962,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.60162548,insgenin (insulinogenic index): 2.017791932,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.470116353,log10 matsuda insulin sensitivity index: 0.608904289,muscle mass (%): 47.5,lg10 serum c-reactive protein (mg/l): 0.074450719,lg10 plasma adiponectin (mg/l): 0.799340549,ogtt fasting plasma free fatty acid (mmol/l): 0.39,ogtt 30 min plasma free fatty acid (mmol/l): 0.26,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 4.9,ogtt 30 min plasma glucose (mmol/l): 7.1,ogtt 120 min plasma glucose (mmol/l): 4.9,log10 il1 receptor antagonist (pg/ml): 2.186730376,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 1.161368002,ogtt 30 min plasma insulin (mu/l): 1.721810615,ogtt 120 min plasma insulin (mu/l): 1.862727528,log10 ogtt fasting plasma proinsulin (pm/l): 1.187520721,ogtt 30 min plasma proinsulin (pm/l): 1.383815366,ogtt 120 min plasma proinsulin (pm/l): 1.804139432,log10 bioimpedance: Resistance: 2.679427897,log10 bioimpedance (reactance): 1.819543936,waist to hip ratio: 0.985,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.447158031,log10 creatinine (umol/l): 1.944482672,log10 total cholesterol (mmol/l): 0.733197265,log10 ldl cholesterol (mmol/l): 0.5774918,log10 hdl cholesterol (mmol/l): 0.029383778,log10 total triglycerides (mmol/l): 0,log10 serum apoa1 (g/l): 0.096910013,log10 serum apob (g/l): 0.079181246,log10 urinary albumin excretion rate (ug/min): 0.726998728
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098196
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249095,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978023
GSM1098196
GSE45159
0.255963
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM388
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249095
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978023
1
age: 60,tissue: adipose tissue,log10 body mass index: 1.350674814,log10 basal metabolic rate (kcal): 2236,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.527231569,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.552601458,plasma free fatty acids under the curve ogtt (mmol/l * min): 25.95,fat mass (%): 15.7,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.518653156,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 109.5,log10 homair (insulin resistance index based on homa): 0.06279083,log10 homais (insulin secretion index based on homa): 1.769551079,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.278319089,insgenin (insulinogenic index): 1.917580524,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.186984572,log10 matsuda insulin sensitivity index: 1.01207028,muscle mass (%): 61.8,lg10 serum c-reactive protein (mg/l): -0.308918508,lg10 plasma adiponectin (mg/l): 1.068185862,ogtt fasting plasma free fatty acid (mmol/l): 0.38,ogtt 30 min plasma free fatty acid (mmol/l): 0.3,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.2,ogtt 30 min plasma glucose (mmol/l): 8,ogtt 120 min plasma glucose (mmol/l): 3.9,log10 il1 receptor antagonist (pg/ml): 1.87517706,log10 il1 beta (pg/ml): -0.494850022,log10 ogtt fasting plasma insulin (mu/l): 0.698970004,ogtt 30 min plasma insulin (mu/l): 1.639486489,ogtt 120 min plasma insulin (mu/l): 1.021189299,log10 ogtt fasting plasma proinsulin (pm/l): 1.056904851,ogtt 30 min plasma proinsulin (pm/l): 1.442479769,ogtt 120 min plasma proinsulin (pm/l): 1.494154594,log10 bioimpedance: Resistance: 2.688419822,log10 bioimpedance (reactance): 2.049218023,waist to hip ratio: 0.93814433,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.230448921,log10 creatinine (umol/l): 1.897627091,log10 total cholesterol (mmol/l): 0.759667845,log10 ldl cholesterol (mmol/l): 0.591064607,log10 hdl cholesterol (mmol/l): 0.204119983,log10 total triglycerides (mmol/l): 0.250420002,log10 serum apoa1 (g/l): 0.260071388,log10 serum apob (g/l): 0.11058971,log10 urinary albumin excretion rate (ug/min): 0.602059991
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098197
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249096,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978024
GSM1098197
GSE45159
0.340697
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM473
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249096
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978024
1
age: 50,tissue: adipose tissue,log10 body mass index: 1.403499536,log10 basal metabolic rate (kcal): 1650,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.623318869,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.887438141,plasma free fatty acids under the curve ogtt (mmol/l * min): 15.75,fat mass (%): 21.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.63481105,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 135,log10 homair (insulin resistance index based on homa): 0.361308025,log10 homais (insulin secretion index based on homa): 1.973127854,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.70981249,insgenin (insulinogenic index): 2.36361198,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.648320972,log10 matsuda insulin sensitivity index: 0.62415725,muscle mass (%): 45.1,lg10 serum c-reactive protein (mg/l): -0.122628654,lg10 plasma adiponectin (mg/l): 1.1430148,ogtt fasting plasma free fatty acid (mmol/l): 0.25,ogtt 30 min plasma free fatty acid (mmol/l): 0.17,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.5,ogtt 30 min plasma glucose (mmol/l): 7.9,ogtt 120 min plasma glucose (mmol/l): 5.3,log10 il1 receptor antagonist (pg/ml): 2.276714495,log10 il1 beta (pg/ml): -0.698970004,log10 ogtt fasting plasma insulin (mu/l): 0.973127854,ogtt 30 min plasma insulin (mu/l): 2.007747778,ogtt 120 min plasma insulin (mu/l): 1.707570176,log10 ogtt fasting plasma proinsulin (pm/l): 1.113943352,ogtt 30 min plasma proinsulin (pm/l): 1.436162647,ogtt 120 min plasma proinsulin (pm/l): 1.525044807,log10 bioimpedance: Resistance: 2.717670503,log10 bioimpedance (reactance): 1.755874856,waist to hip ratio: 0.96969697,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.62324929,log10 creatinine (umol/l): 1.886490725,log10 total cholesterol (mmol/l): 0.804820679,log10 ldl cholesterol (mmol/l): 0.667452953,log10 hdl cholesterol (mmol/l): 0.164352856,log10 total triglycerides (mmol/l): 0.269512944,log10 serum apoa1 (g/l): 0.204119983,log10 serum apob (g/l): 0.167317335,log10 urinary albumin excretion rate (ug/min): 0.810417027
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098198
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249097,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978025
GSM1098198
GSE45159
0.363363
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM482
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249097
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978025
1
age: 58,tissue: adipose tissue,log10 body mass index: 1.487384452,log10 basal metabolic rate (kcal): 1562,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.376566407,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.749816692,plasma free fatty acids under the curve ogtt (mmol/l * min): 20.4,fat mass (%): 25,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.610102063,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 145.5,log10 homair (insulin resistance index based on homa): 0.104487111,log10 homais (insulin secretion index based on homa): 1.77815125,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.587508921,insgenin (insulinogenic index): 2.150466011,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.541504359,log10 matsuda insulin sensitivity index: 0.834858425,muscle mass (%): 41.3,lg10 serum c-reactive protein (mg/l): -0.097997109,lg10 plasma adiponectin (mg/l): 0.954242509,ogtt fasting plasma free fatty acid (mmol/l): 0.44,ogtt 30 min plasma free fatty acid (mmol/l): 0.2,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.3,ogtt 30 min plasma glucose (mmol/l): 9,ogtt 120 min plasma glucose (mmol/l): 3.6,log10 il1 receptor antagonist (pg/ml): 2.234162819,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.73239376,ogtt 30 min plasma insulin (mu/l): 1.966610987,ogtt 120 min plasma insulin (mu/l): 1.255272505,log10 ogtt fasting plasma proinsulin (pm/l): 1.167317335,ogtt 30 min plasma proinsulin (pm/l): 1.494154594,ogtt 120 min plasma proinsulin (pm/l): 1.588831726,log10 bioimpedance: Resistance: 2.671172843,log10 bioimpedance (reactance): 1.672097858,waist to hip ratio: 1.034825871,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.505149978,log10 creatinine (umol/l): 1.939519253,log10 total cholesterol (mmol/l): 0.69019608,log10 ldl cholesterol (mmol/l): 0.530199698,log10 hdl cholesterol (mmol/l): 0.220108088,log10 total triglycerides (mmol/l): 0.045322979,log10 serum apoa1 (g/l): 0.187520721,log10 serum apob (g/l): 0.021189299,log10 urinary albumin excretion rate (ug/min): 0.949661017
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098199
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249098,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978026
GSM1098199
GSE45159
0.156395
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM490
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249098
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978026
1
age: 53,tissue: adipose tissue,log10 body mass index: 1.389436069,log10 basal metabolic rate (kcal): 1758,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.365129111,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.730864987,plasma free fatty acids under the curve ogtt (mmol/l * min): 16.65,fat mass (%): 20.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.485829309,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 105,log10 homair (insulin resistance index based on homa): 0.045583738,log10 homais (insulin secretion index based on homa): 1.787106093,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.142326883,insgenin (insulinogenic index): 1.719505737,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.01494035,log10 matsuda insulin sensitivity index: 1.069394356,muscle mass (%): 44.9,lg10 serum c-reactive protein (mg/l): 0.173477643,lg10 plasma adiponectin (mg/l): 0.857332496,ogtt fasting plasma free fatty acid (mmol/l): 0.3,ogtt 30 min plasma free fatty acid (mmol/l): 0.18,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.1,ogtt 30 min plasma glucose (mmol/l): 7,ogtt 120 min plasma glucose (mmol/l): 4.9,log10 il1 receptor antagonist (pg/ml): 2.129077324,log10 il1 beta (pg/ml): -0.853871964,log10 ogtt fasting plasma insulin (mu/l): 0.69019608,ogtt 30 min plasma insulin (mu/l): 1.33243846,ogtt 120 min plasma insulin (mu/l): 1.324282455,log10 ogtt fasting plasma proinsulin (pm/l): 1.053078443,ogtt 30 min plasma proinsulin (pm/l): 1.26245109,ogtt 120 min plasma proinsulin (pm/l): 1.428134794,log10 bioimpedance: Resistance: 2.705007959,log10 bioimpedance (reactance): 1.698970004,waist to hip ratio: 0.970588235,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.176091259,log10 creatinine (umol/l): 1.949390007,log10 total cholesterol (mmol/l): 0.767155866,log10 ldl cholesterol (mmol/l): 0.604226053,log10 hdl cholesterol (mmol/l): 0.176091259,log10 total triglycerides (mmol/l): -0.065501549,log10 serum apoa1 (g/l): 0.217483944,log10 serum apob (g/l): 0.064457989,log10 urinary albumin excretion rate (ug/min): 0.687029885
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098200
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249099,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978027
GSM1098200
GSE45159
0.258982
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM492
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249099
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978027
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.373285187,log10 basal metabolic rate (kcal): 1546,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.373093233,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.455462055,plasma free fatty acids under the curve ogtt (mmol/l * min): 40.2,fat mass (%): 16.7,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.16616308,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 453,log10 homair (insulin resistance index based on homa): -0.032452024,log10 homais (insulin secretion index based on homa): 1.49560466,log10 insulin area under the curve (ogtt) (pmol/l * min): 3.807332039,insgenin (insulinogenic index): 0.963483166,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.582631439,log10 matsuda insulin sensitivity index: 1.190338397,muscle mass (%): 45.9,lg10 serum c-reactive protein (mg/l): -0.559090918,lg10 plasma adiponectin (mg/l): 1.053078443,ogtt fasting plasma free fatty acid (mmol/l): 0.67,ogtt 30 min plasma free fatty acid (mmol/l): 0.39,ogtt 120 min plasma free fatty acid (mmol/l): 0.15,ogtt fasting plasma glucose (mmol/l): 5.8,ogtt 30 min plasma glucose (mmol/l): 12,ogtt 120 min plasma glucose (mmol/l): 7.6,log10 il1 receptor antagonist (pg/ml): 2.099231615,log10 il1 beta (pg/ml): -0.958607315,log10 ogtt fasting plasma insulin (mu/l): 0.556302501,ogtt 30 min plasma insulin (mu/l): 1.117271296,ogtt 120 min plasma insulin (mu/l): 0.707570176,log10 ogtt fasting plasma proinsulin (pm/l): 0.86923172,ogtt 30 min plasma proinsulin (pm/l): 1.176091259,ogtt 120 min plasma proinsulin (pm/l): 1.385606274,log10 bioimpedance: Resistance: 2.691081492,log10 bioimpedance (reactance): 1.662757832,waist to hip ratio: 0.984536082,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.255272505,log10 creatinine (umol/l): 1.944482672,log10 total cholesterol (mmol/l): 0.774516966,log10 ldl cholesterol (mmol/l): 0.572871602,log10 hdl cholesterol (mmol/l): 0.359835482,log10 total triglycerides (mmol/l): -0.148741651,log10 serum apoa1 (g/l): 0.264817823,log10 serum apob (g/l): -0.045757491,log10 urinary albumin excretion rate (ug/min): 0.646581935
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098201
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249100,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978028
GSM1098201
GSE45159
0.176817
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM520
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249100
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978028
1
age: 58,tissue: adipose tissue,log10 body mass index: 1.406329326,log10 basal metabolic rate (kcal): 1726,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.230109829,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.46745006,plasma free fatty acids under the curve ogtt (mmol/l * min): 18.15,fat mass (%): 16.5,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.080817528,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): -22.5,log10 homair (insulin resistance index based on homa): -0.217686662,log10 homais (insulin secretion index based on homa): 1.684246748,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.372875197,insgenin (insulinogenic index): NA,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.33264041,log10 matsuda insulin sensitivity index: 1.164045038,muscle mass (%): 47.1,lg10 serum c-reactive protein (mg/l): 0.140508043,lg10 plasma adiponectin (mg/l): 0.826074803,ogtt fasting plasma free fatty acid (mmol/l): 0.28,ogtt 30 min plasma free fatty acid (mmol/l): 0.21,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 4.7,ogtt 30 min plasma glucose (mmol/l): 4.7,ogtt 120 min plasma glucose (mmol/l): 4.2,log10 il1 receptor antagonist (pg/ml): 2.099646117,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.462397998,ogtt 30 min plasma insulin (mu/l): 1.754348336,ogtt 120 min plasma insulin (mu/l): 1.029383778,log10 ogtt fasting plasma proinsulin (pm/l): 1.033423755,ogtt 30 min plasma proinsulin (pm/l): 1.437750563,ogtt 120 min plasma proinsulin (pm/l): 1.46834733,log10 bioimpedance: Resistance: 2.617000341,log10 bioimpedance (reactance): 1.602059991,waist to hip ratio: 0.917525773,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.531478917,log10 creatinine (umol/l): 1.977723605,log10 total cholesterol (mmol/l): 0.795880017,log10 ldl cholesterol (mmol/l): 0.646403726,log10 hdl cholesterol (mmol/l): 0.212187604,log10 total triglycerides (mmol/l): 0.10720997,log10 serum apoa1 (g/l): 0.167317335,log10 serum apob (g/l): 0.075546961,log10 urinary albumin excretion rate (ug/min): 0.572096768
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098202
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249101,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978029
GSM1098202
GSE45159
0.183983
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM522
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249101
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978029
1
age: 60,tissue: adipose tissue,log10 body mass index: 1.478333113,log10 basal metabolic rate (kcal): 1630,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.202747826,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.508748397,plasma free fatty acids under the curve ogtt (mmol/l * min): 20.1,fat mass (%): 22.7,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.769837844,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 201,log10 homair (insulin resistance index based on homa): 0.215549444,log10 homais (insulin secretion index based on homa): 1.798354636,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.341632336,insgenin (insulinogenic index): 1.854539596,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.235730397,log10 matsuda insulin sensitivity index: 0.850912144,muscle mass (%): 44,lg10 serum c-reactive protein (mg/l): 0.213783299,lg10 plasma adiponectin (mg/l): 0.568201724,ogtt fasting plasma free fatty acid (mmol/l): 0.31,ogtt 30 min plasma free fatty acid (mmol/l): 0.22,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 8.2,ogtt 120 min plasma glucose (mmol/l): 6.6,log10 il1 receptor antagonist (pg/ml): 1.935104021,log10 il1 beta (pg/ml): -0.537602002,log10 ogtt fasting plasma insulin (mu/l): 0.819543936,ogtt 30 min plasma insulin (mu/l): 1.575187845,ogtt 120 min plasma insulin (mu/l): 1.462397998,log10 ogtt fasting plasma proinsulin (pm/l): 0.991226076,ogtt 30 min plasma proinsulin (pm/l): 1.243038049,ogtt 120 min plasma proinsulin (pm/l): 1.495544338,log10 bioimpedance: Resistance: 2.63748973,log10 bioimpedance (reactance): 1.681241237,waist to hip ratio: 1.028846154,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.342422681,log10 creatinine (umol/l): 1.982271233,log10 total cholesterol (mmol/l): 0.680335513,log10 ldl cholesterol (mmol/l): 0.509202522,log10 hdl cholesterol (mmol/l): 0.012837225,log10 total triglycerides (mmol/l): 0.240549248,log10 serum apoa1 (g/l): 0.075546961,log10 serum apob (g/l): 0.08278537,log10 urinary albumin excretion rate (ug/min): 1.245070488
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098203
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249102,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978030
GSM1098203
GSE45159
0.211758
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM575
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249102
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978030
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.393928211,log10 basal metabolic rate (kcal): 1730,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.639220552,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.842943832,plasma free fatty acids under the curve ogtt (mmol/l * min): 24.75,fat mass (%): 19,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.701306462,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 196.5,log10 homair (insulin resistance index based on homa): 0.142945363,log10 homais (insulin secretion index based on homa): 1.816609502,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.196037941,insgenin (insulinogenic index): 1.60159773,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.059070913,log10 matsuda insulin sensitivity index: 0.990185811,muscle mass (%): 46.9,lg10 serum c-reactive protein (mg/l): 0.308777774,lg10 plasma adiponectin (mg/l): 1.133538908,ogtt fasting plasma free fatty acid (mmol/l): 0.38,ogtt 30 min plasma free fatty acid (mmol/l): 0.22,ogtt 120 min plasma free fatty acid (mmol/l): 0.13,ogtt fasting plasma glucose (mmol/l): 5.3,ogtt 30 min plasma glucose (mmol/l): 10,ogtt 120 min plasma glucose (mmol/l): 3.4,log10 il1 receptor antagonist (pg/ml): 2.073351702,log10 il1 beta (pg/ml): -1,log10 ogtt fasting plasma insulin (mu/l): 0.770852012,ogtt 30 min plasma insulin (mu/l): 1.57054294,ogtt 120 min plasma insulin (mu/l): 0.819543936,log10 ogtt fasting plasma proinsulin (pm/l): 0.792391689,ogtt 30 min plasma proinsulin (pm/l): 1.204119983,ogtt 120 min plasma proinsulin (pm/l): 1.152288344,log10 bioimpedance: Resistance: 2.684845362,log10 bioimpedance (reactance): 1.72427587,waist to hip ratio: 0.917525773,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.397940009,log10 creatinine (umol/l): 1.875061263,log10 total cholesterol (mmol/l): 0.693726949,log10 ldl cholesterol (mmol/l): 0.477121255,log10 hdl cholesterol (mmol/l): 0.238046103,log10 total triglycerides (mmol/l): -0.080921908,log10 serum apoa1 (g/l): 0.173186268,log10 serum apob (g/l): -0.050609993,log10 urinary albumin excretion rate (ug/min): 1.044812344
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098204
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249103,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978031
GSM1098204
GSE45159
0.066173
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM598
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249103
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978031
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.465206508,log10 basal metabolic rate (kcal): 1713,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.475855676,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.886433506,plasma free fatty acids under the curve ogtt (mmol/l * min): 31.65,fat mass (%): 24.5,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.981567282,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 303,log10 homair (insulin resistance index based on homa): 0.49785074,log10 homais (insulin secretion index based on homa): 2,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.714966482,insgenin (insulinogenic index): 2.158814647,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.635845508,log10 matsuda insulin sensitivity index: 0.50747184,muscle mass (%): 43.8,lg10 serum c-reactive protein (mg/l): 0.29841638,lg10 plasma adiponectin (mg/l): 0.799340549,ogtt fasting plasma free fatty acid (mmol/l): 0.65,ogtt 30 min plasma free fatty acid (mmol/l): 0.29,ogtt 120 min plasma free fatty acid (mmol/l): 0.1,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 9.9,ogtt 120 min plasma glucose (mmol/l): 7.3,log10 il1 receptor antagonist (pg/ml): 2.045714059,log10 il1 beta (pg/ml): -0.853871964,log10 ogtt fasting plasma insulin (mu/l): 1.079181246,ogtt 30 min plasma insulin (mu/l): 2.033825694,ogtt 120 min plasma insulin (mu/l): 1.643452676,log10 ogtt fasting plasma proinsulin (pm/l): 1.340444115,ogtt 30 min plasma proinsulin (pm/l): 1.870988814,ogtt 120 min plasma proinsulin (pm/l): 1.968482949,log10 bioimpedance: Resistance: 2.686636269,log10 bioimpedance (reactance): 1.763427994,waist to hip ratio: 1.014851485,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.612783857,log10 creatinine (umol/l): 1.919078092,log10 total cholesterol (mmol/l): 0.792391689,log10 ldl cholesterol (mmol/l): 0.369215857,log10 hdl cholesterol (mmol/l): 0.033423755,log10 total triglycerides (mmol/l): 0.883093359,log10 serum apoa1 (g/l): 0.120573931,log10 serum apob (g/l): 0.071882007,log10 urinary albumin excretion rate (ug/min): 0.87185875
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098205
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249104,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978032
GSM1098205
GSE45159
0.113722
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM783
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249104
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978032
1
age: 51,tissue: adipose tissue,log10 body mass index: 1.48669827,log10 basal metabolic rate (kcal): 1676,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.069738097,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.600171425,plasma free fatty acids under the curve ogtt (mmol/l * min): 27.75,fat mass (%): 22.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.904634622,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 238.5,log10 homair (insulin resistance index based on homa): 0.543240028,log10 homais (insulin secretion index based on homa): 2.020361283,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.629287245,insgenin (insulinogenic index): 1.980003372,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.520562131,log10 matsuda insulin sensitivity index: 0.524273321,muscle mass (%): 43.3,lg10 serum c-reactive protein (mg/l): -0.152427341,lg10 plasma adiponectin (mg/l): 0.672097858,ogtt fasting plasma free fatty acid (mmol/l): 0.58,ogtt 30 min plasma free fatty acid (mmol/l): 0.25,ogtt 120 min plasma free fatty acid (mmol/l): 0.09,ogtt fasting plasma glucose (mmol/l): 6,ogtt 30 min plasma glucose (mmol/l): 9.6,ogtt 120 min plasma glucose (mmol/l): 6.5,log10 il1 receptor antagonist (pg/ml): 2.261619677,log10 il1 beta (pg/ml): 0.021189299,log10 ogtt fasting plasma insulin (mu/l): 1.117271296,ogtt 30 min plasma insulin (mu/l): 1.847572659,ogtt 120 min plasma insulin (mu/l): 1.774516966,log10 ogtt fasting plasma proinsulin (pm/l): 0.903089987,ogtt 30 min plasma proinsulin (pm/l): 1.305351369,ogtt 120 min plasma proinsulin (pm/l): 1.692846919,log10 bioimpedance: Resistance: 2.605305046,log10 bioimpedance (reactance): 1.602059991,waist to hip ratio: 0.976303318,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.531478917,log10 creatinine (umol/l): 2.029383778,log10 total cholesterol (mmol/l): 0.648360011,log10 ldl cholesterol (mmol/l): 0.456366033,log10 hdl cholesterol (mmol/l): 0.089905111,log10 total triglycerides (mmol/l): 0.152288344,log10 serum apoa1 (g/l): 0.089905111,log10 serum apob (g/l): -0.026872146,log10 urinary albumin excretion rate (ug/min): 1.134546163
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098206
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249105,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978033
GSM1098206
GSE45159
0.280541
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM803
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249105
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978033
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.327687058,log10 basal metabolic rate (kcal): 1553,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.584945002,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.437725264,plasma free fatty acids under the curve ogtt (mmol/l * min): 25.95,fat mass (%): 11.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.439830884,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 46.5,log10 homair (insulin resistance index based on homa): -0.101274818,log10 homais (insulin secretion index based on homa): 1.540790335,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.057704132,insgenin (insulinogenic index): 1.586265724,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.956408571,log10 matsuda insulin sensitivity index: 1.170063428,muscle mass (%): 49.7,lg10 serum c-reactive protein (mg/l): -0.583359493,lg10 plasma adiponectin (mg/l): 0.770852012,ogtt fasting plasma free fatty acid (mmol/l): 0.46,ogtt 30 min plasma free fatty acid (mmol/l): 0.28,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 6.1,ogtt 120 min plasma glucose (mmol/l): 5.5,log10 il1 receptor antagonist (pg/ml): 2.045362103,log10 il1 beta (pg/ml): -0.568636236,log10 ogtt fasting plasma insulin (mu/l): 0.51851394,ogtt 30 min plasma insulin (mu/l): 0.892094603,ogtt 120 min plasma insulin (mu/l): 1.488550717,log10 ogtt fasting plasma proinsulin (pm/l): 0.995635195,ogtt 30 min plasma proinsulin (pm/l): 1.017033339,ogtt 120 min plasma proinsulin (pm/l): 1.390935107,log10 bioimpedance: Resistance: 2.679427897,log10 bioimpedance (reactance): 1.662757832,waist to hip ratio: 0.840659341,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 0.903089987,log10 creatinine (umol/l): 1.886490725,log10 total cholesterol (mmol/l): 0.750508395,log10 ldl cholesterol (mmol/l): 0.53529412,log10 hdl cholesterol (mmol/l): 0.307496038,log10 total triglycerides (mmol/l): -0.091514981,log10 serum apoa1 (g/l): 0.204119983,log10 serum apob (g/l): -0.031517051,log10 urinary albumin excretion rate (ug/min): 0.508676069
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098207
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249106,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978034
GSM1098207
GSE45159
0.187657
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM817
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249106
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978034
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.517036236,log10 basal metabolic rate (kcal): 1684,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.611832412,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.11805107,plasma free fatty acids under the curve ogtt (mmol/l * min): 14.7,fat mass (%): 24.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.515699838,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 84,log10 homair (insulin resistance index based on homa): 0.261024834,log10 homais (insulin secretion index based on homa): 1.903089987,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.4954472,insgenin (insulinogenic index): 2.385118576,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.411973058,log10 matsuda insulin sensitivity index: 0.783578563,muscle mass (%): 42.7,lg10 serum c-reactive protein (mg/l): 0.241795431,lg10 plasma adiponectin (mg/l): 0.612783857,ogtt fasting plasma free fatty acid (mmol/l): 0.25,ogtt 30 min plasma free fatty acid (mmol/l): 0.16,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 6.5,ogtt 120 min plasma glucose (mmol/l): 5.8,log10 il1 receptor antagonist (pg/ml): 2.337738717,log10 il1 beta (pg/ml): -0.552841969,log10 ogtt fasting plasma insulin (mu/l): 0.880813592,ogtt 30 min plasma insulin (mu/l): 1.716837723,ogtt 120 min plasma insulin (mu/l): 1.64246452,log10 ogtt fasting plasma proinsulin (pm/l): 1.245512668,ogtt 30 min plasma proinsulin (pm/l): 1.666517981,ogtt 120 min plasma proinsulin (pm/l): 1.909556029,log10 bioimpedance: Resistance: 2.603144373,log10 bioimpedance (reactance): 1.633468456,waist to hip ratio: 0.990909091,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.556302501,log10 creatinine (umol/l): 1.886490725,log10 total cholesterol (mmol/l): 0.693726949,log10 ldl cholesterol (mmol/l): 0.50242712,log10 hdl cholesterol (mmol/l): 0.056904851,log10 total triglycerides (mmol/l): 0.318063335,log10 serum apoa1 (g/l): 0.100370545,log10 serum apob (g/l): 0.025305865,log10 urinary albumin excretion rate (ug/min): 0.922510263
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098208
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249107,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978035
GSM1098208
GSE45159
0.323616
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM833
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249107
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978035
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.396552355,log10 basal metabolic rate (kcal): 1663,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.485385281,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.705854221,plasma free fatty acids under the curve ogtt (mmol/l * min): 29.7,fat mass (%): 21.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.04028972,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 333,log10 homair (insulin resistance index based on homa): 0.610660163,log10 homais (insulin secretion index based on homa): 2.087781418,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.808730734,insgenin (insulinogenic index): 2.151817564,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.72722396,log10 matsuda insulin sensitivity index: 0.385433076,muscle mass (%): 45,lg10 serum c-reactive protein (mg/l): -0.53313238,lg10 plasma adiponectin (mg/l): 0.414973348,ogtt fasting plasma free fatty acid (mmol/l): 0.34,ogtt 30 min plasma free fatty acid (mmol/l): 0.35,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 6,ogtt 30 min plasma glucose (mmol/l): 9.9,ogtt 120 min plasma glucose (mmol/l): 8.2,log10 il1 receptor antagonist (pg/ml): 2.269466243,log10 il1 beta (pg/ml): -0.568636236,log10 ogtt fasting plasma insulin (mu/l): 1.184691431,ogtt 30 min plasma insulin (mu/l): 2.031408464,ogtt 120 min plasma insulin (mu/l): 1.954242509,log10 ogtt fasting plasma proinsulin (pm/l): 1.298853076,ogtt 30 min plasma proinsulin (pm/l): 1.688419822,ogtt 120 min plasma proinsulin (pm/l): 1.849419414,log10 bioimpedance: Resistance: 2.728353782,log10 bioimpedance (reactance): 1.763427994,waist to hip ratio: 1.01025641,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.568201724,log10 creatinine (umol/l): 1.913813852,log10 total cholesterol (mmol/l): 0.580924976,log10 ldl cholesterol (mmol/l): 0.267171728,log10 hdl cholesterol (mmol/l): 0.068185862,log10 total triglycerides (mmol/l): 0.475671188,log10 serum apoa1 (g/l): 0.139879086,log10 serum apob (g/l): -0.022276395,log10 urinary albumin excretion rate (ug/min): 1.007061855
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098209
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249108,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978036
GSM1098209
GSE45159
0.149286
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM844
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249108
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978036
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.472965295,log10 basal metabolic rate (kcal): 1577,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.450543015,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.820917525,plasma free fatty acids under the curve ogtt (mmol/l * min): 18,fat mass (%): 23.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.587777516,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 97.5,log10 homair (insulin resistance index based on homa): 0.068103367,log10 homais (insulin secretion index based on homa): 1.650908559,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.321411998,insgenin (insulinogenic index): 2.007320953,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.244944753,log10 matsuda insulin sensitivity index: 0.964545375,muscle mass (%): 42.1,lg10 serum c-reactive protein (mg/l): 0.574841195,lg10 plasma adiponectin (mg/l): 0.880813592,ogtt fasting plasma free fatty acid (mmol/l): 0.38,ogtt 30 min plasma free fatty acid (mmol/l): 0.19,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 7.6,ogtt 120 min plasma glucose (mmol/l): 5.1,log10 il1 receptor antagonist (pg/ml): 2.070665753,log10 il1 beta (pg/ml): -0.769551079,log10 ogtt fasting plasma insulin (mu/l): 0.672097858,ogtt 30 min plasma insulin (mu/l): 1.586587305,ogtt 120 min plasma insulin (mu/l): 1.390935107,log10 ogtt fasting plasma proinsulin (pm/l): 1.033423755,ogtt 30 min plasma proinsulin (pm/l): 1.338456494,ogtt 120 min plasma proinsulin (pm/l): 1.633468456,log10 bioimpedance: Resistance: 2.652246341,log10 bioimpedance (reactance): 1.633468456,waist to hip ratio: 1.014354067,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.361727836,log10 creatinine (umol/l): 1.924279286,log10 total cholesterol (mmol/l): 0.71432976,log10 ldl cholesterol (mmol/l): 0.494154594,log10 hdl cholesterol (mmol/l): 0.217483944,log10 total triglycerides (mmol/l): 0.053078443,log10 serum apoa1 (g/l): 0.220108088,log10 serum apob (g/l): 0,log10 urinary albumin excretion rate (ug/min): 0.647287531
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098210
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249109,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978037
GSM1098210
GSE45159
0.226469
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM874
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249109
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978037
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.430991056,log10 basal metabolic rate (kcal): 1436,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.20051725,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.26100052,plasma free fatty acids under the curve ogtt (mmol/l * min): 22.95,fat mass (%): 21,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.714245518,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 264,log10 homair (insulin resistance index based on homa): -0.068881289,log10 homais (insulin secretion index based on homa): 1.789146635,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.568717876,insgenin (insulinogenic index): 2.0678017,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.533568713,log10 matsuda insulin sensitivity index: 0.898372406,muscle mass (%): 44.5,lg10 serum c-reactive protein (mg/l): 0.707825568,lg10 plasma adiponectin (mg/l): 1.045322979,ogtt fasting plasma free fatty acid (mmol/l): 0.38,ogtt 30 min plasma free fatty acid (mmol/l): 0.25,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 4.8,ogtt 30 min plasma glucose (mmol/l): 7.7,ogtt 120 min plasma glucose (mmol/l): 6.8,log10 il1 receptor antagonist (pg/ml): 2.219767845,log10 il1 beta (pg/ml): -0.102372909,log10 ogtt fasting plasma insulin (mu/l): 0.602059991,ogtt 30 min plasma insulin (mu/l): 1.781755375,ogtt 120 min plasma insulin (mu/l): 1.741939078,log10 ogtt fasting plasma proinsulin (pm/l): 1,ogtt 30 min plasma proinsulin (pm/l): 1.423245874,ogtt 120 min plasma proinsulin (pm/l): 1.776701184,log10 bioimpedance: Resistance: 2.705007959,log10 bioimpedance (reactance): 1.755874856,waist to hip ratio: 1.010752688,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.544068044,log10 creatinine (umol/l): 1.986771734,log10 total cholesterol (mmol/l): 0.877946952,log10 ldl cholesterol (mmol/l): 0.730782276,log10 hdl cholesterol (mmol/l): 0.240549248,log10 total triglycerides (mmol/l): -0.148741651,log10 serum apoa1 (g/l): 0.247973266,log10 serum apob (g/l): 0.170261715,log10 urinary albumin excretion rate (ug/min): 0.575642049
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098211
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249110,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978038
GSM1098211
GSE45159
0.237078
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM888
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249110
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978038
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.414995599,log10 basal metabolic rate (kcal): 1754,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.389519992,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.635899375,plasma free fatty acids under the curve ogtt (mmol/l * min): 17.4,fat mass (%): 17.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.714245518,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 180,log10 homair (insulin resistance index based on homa): -0.120458135,log10 homais (insulin secretion index based on homa): 1.491361694,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.198024426,insgenin (insulinogenic index): 1.583576586,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.131779009,log10 matsuda insulin sensitivity index: 1.106043024,muscle mass (%): 48.8,lg10 serum c-reactive protein (mg/l): -0.573488739,lg10 plasma adiponectin (mg/l): 1.243038049,ogtt fasting plasma free fatty acid (mmol/l): 0.22,ogtt 30 min plasma free fatty acid (mmol/l): 0.22,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.5,ogtt 30 min plasma glucose (mmol/l): 9.1,ogtt 120 min plasma glucose (mmol/l): 4.7,log10 il1 receptor antagonist (pg/ml): 2.19002338,log10 il1 beta (pg/ml): -0.356547324,log10 ogtt fasting plasma insulin (mu/l): 0.491361694,ogtt 30 min plasma insulin (mu/l): 1.416640507,ogtt 120 min plasma insulin (mu/l): 1.354108439,log10 ogtt fasting plasma proinsulin (pm/l): 0.991226076,ogtt 30 min plasma proinsulin (pm/l): 1.357934847,ogtt 120 min plasma proinsulin (pm/l): 1.672097858,log10 bioimpedance: Resistance: 2.62838893,log10 bioimpedance (reactance): 1.698970004,waist to hip ratio: 0.904761905,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.322219295,log10 creatinine (umol/l): 1.944482672,log10 total cholesterol (mmol/l): 0.797959644,log10 ldl cholesterol (mmol/l): 0.550228353,log10 hdl cholesterol (mmol/l): 0.419955748,log10 total triglycerides (mmol/l): -0.075720714,log10 serum apoa1 (g/l): 0.342422681,log10 serum apob (g/l): -0.013228266,log10 urinary albumin excretion rate (ug/min): 1.078012216
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098212
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249111,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978039
GSM1098212
GSE45159
0.254287
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM906
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249111
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978039
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.423812763,log10 basal metabolic rate (kcal): 1629,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.084411853,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.302689358,plasma free fatty acids under the curve ogtt (mmol/l * min): 12.45,fat mass (%): 20.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.714245518,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 192,log10 homair (insulin resistance index based on homa): 0.504062883,log10 homais (insulin secretion index based on homa): 2.146128036,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.808609287,insgenin (insulinogenic index): 2.158362492,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.738645811,log10 matsuda insulin sensitivity index: 0.474421526,muscle mass (%): 46.8,lg10 serum c-reactive protein (mg/l): -0.067019178,lg10 plasma adiponectin (mg/l): 0.832508913,ogtt fasting plasma free fatty acid (mmol/l): 0.17,ogtt 30 min plasma free fatty acid (mmol/l): 0.15,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 8,ogtt 120 min plasma glucose (mmol/l): 6.2,log10 il1 receptor antagonist (pg/ml): 2.857784651,log10 il1 beta (pg/ml): 0.152288344,log10 ogtt fasting plasma insulin (mu/l): 1.123851641,ogtt 30 min plasma insulin (mu/l): 1.87909588,ogtt 120 min plasma insulin (mu/l): 2.123851641,log10 ogtt fasting plasma proinsulin (pm/l): 1.127104798,ogtt 30 min plasma proinsulin (pm/l): 1.383815366,ogtt 120 min plasma proinsulin (pm/l): 1.81756537,log10 bioimpedance: Resistance: 2.678518379,log10 bioimpedance (reactance): 1.763427994,waist to hip ratio: 1.041884817,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.770852012,log10 creatinine (umol/l): 2.021189299,log10 total cholesterol (mmol/l): 0.706717782,log10 ldl cholesterol (mmol/l): 0.5132176,log10 hdl cholesterol (mmol/l): 0.133538908,log10 total triglycerides (mmol/l): 0.240549248,log10 serum apoa1 (g/l): 0.184691431,log10 serum apob (g/l): 0.064457989,log10 urinary albumin excretion rate (ug/min): 0.794995807
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098213
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249112,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978040
GSM1098213
GSE45159
0.169843
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM908
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249112
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978040
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.429511788,log10 basal metabolic rate (kcal): 1492,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.511090418,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.6875184,plasma free fatty acids under the curve ogtt (mmol/l * min): 19.8,fat mass (%): 20.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.900112063,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 319.5,log10 homair (insulin resistance index based on homa): 0.258584077,log10 homais (insulin secretion index based on homa): 1.932248216,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.586902223,insgenin (insulinogenic index): 1.950781977,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.519618012,log10 matsuda insulin sensitivity index: 0.701131284,muscle mass (%): 43.2,lg10 serum c-reactive protein (mg/l): -0.044793462,lg10 plasma adiponectin (mg/l): 0.968482949,ogtt fasting plasma free fatty acid (mmol/l): 0.4,ogtt 30 min plasma free fatty acid (mmol/l): 0.2,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.3,ogtt 30 min plasma glucose (mmol/l): 9.5,ogtt 120 min plasma glucose (mmol/l): 6.8,log10 il1 receptor antagonist (pg/ml): 1.772321707,log10 il1 beta (pg/ml): -0.356547324,log10 ogtt fasting plasma insulin (mu/l): 0.886490725,ogtt 30 min plasma insulin (mu/l): 1.846337112,ogtt 120 min plasma insulin (mu/l): 1.671172843,log10 ogtt fasting plasma proinsulin (pm/l): 1.029383778,ogtt 30 min plasma proinsulin (pm/l): 1.534026106,ogtt 120 min plasma proinsulin (pm/l): 1.689308859,log10 bioimpedance: Resistance: 2.685741739,log10 bioimpedance (reactance): 1.662757832,waist to hip ratio: 1.026737968,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.491361694,log10 creatinine (umol/l): 1.908485019,log10 total cholesterol (mmol/l): 0.761927838,log10 ldl cholesterol (mmol/l): 0.559906625,log10 hdl cholesterol (mmol/l): 0.209515015,log10 total triglycerides (mmol/l): 0.278753601,log10 serum apoa1 (g/l): 0.227886705,log10 serum apob (g/l): 0.086359831,log10 urinary albumin excretion rate (ug/min): 0.681241237
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098214
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249113,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978041
GSM1098214
GSE45159
0.009933
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM951
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249113
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978041
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.340452376,log10 basal metabolic rate (kcal): 1556,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.723610629,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.694080445,plasma free fatty acids under the curve ogtt (mmol/l * min): 8.4,fat mass (%): 17.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.482807957,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 67.5,log10 homair (insulin resistance index based on homa): -0.172630727,log10 homais (insulin secretion index based on homa): 1.469434426,log10 insulin area under the curve (ogtt) (pmol/l * min): 3.982813762,insgenin (insulinogenic index): 1.640978057,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.880584956,log10 matsuda insulin sensitivity index: 1.274560314,muscle mass (%): 47.4,lg10 serum c-reactive protein (mg/l): -0.404503778,lg10 plasma adiponectin (mg/l): 0.892094603,ogtt fasting plasma free fatty acid (mmol/l): 0.1,ogtt 30 min plasma free fatty acid (mmol/l): 0.1,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 7.8,ogtt 120 min plasma glucose (mmol/l): 3.7,log10 il1 receptor antagonist (pg/ml): 1.736635498,log10 il1 beta (pg/ml): -0.744727495,log10 ogtt fasting plasma insulin (mu/l): 0.447158031,ogtt 30 min plasma insulin (mu/l): 1.307496038,ogtt 120 min plasma insulin (mu/l): 0.880813592,log10 ogtt fasting plasma proinsulin (pm/l): 0.954242509,ogtt 30 min plasma proinsulin (pm/l): 1.336459734,ogtt 120 min plasma proinsulin (pm/l): 1.413299764,log10 bioimpedance: Resistance: 2.764176132,log10 bioimpedance (reactance): 1.812913357,waist to hip ratio: 0.924731183,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.301029996,log10 creatinine (umol/l): 1.857332496,log10 total cholesterol (mmol/l): 0.827369273,log10 ldl cholesterol (mmol/l): 0.648360011,log10 hdl cholesterol (mmol/l): 0.298853076,log10 total triglycerides (mmol/l): 0.041392685,log10 serum apoa1 (g/l): 0.260071388,log10 serum apob (g/l): 0.093421685,log10 urinary albumin excretion rate (ug/min): 0.739313246
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098215
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249114,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978042
GSM1098215
GSE45159
0.276479
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM964
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249114
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978042
1
age: 53,tissue: adipose tissue,log10 body mass index: 1.409596272,log10 basal metabolic rate (kcal): 1779,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.58477133,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.879855404,plasma free fatty acids under the curve ogtt (mmol/l * min): 47.85,fat mass (%): 20.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.794415866,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 264,log10 homair (insulin resistance index based on homa): -0.131029196,log10 homais (insulin secretion index based on homa): 1.575731053,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.154545692,insgenin (insulinogenic index): 1.522650109,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.07809415,log10 matsuda insulin sensitivity index: 1.123009983,muscle mass (%): 46.8,lg10 serum c-reactive protein (mg/l): 0.823082797,lg10 plasma adiponectin (mg/l): 0.763427994,ogtt fasting plasma free fatty acid (mmol/l): 0.82,ogtt 30 min plasma free fatty acid (mmol/l): 0.51,ogtt 120 min plasma free fatty acid (mmol/l): 0.11,ogtt fasting plasma glucose (mmol/l): 5.2,ogtt 30 min plasma glucose (mmol/l): 9,ogtt 120 min plasma glucose (mmol/l): 6,log10 il1 receptor antagonist (pg/ml): 2.219348717,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.505149978,ogtt 30 min plasma insulin (mu/l): 1.385606274,ogtt 120 min plasma insulin (mu/l): 1.28780173,log10 ogtt fasting plasma proinsulin (pm/l): 0.934498451,ogtt 30 min plasma proinsulin (pm/l): 1.247973266,ogtt 120 min plasma proinsulin (pm/l): 1.505149978,log10 bioimpedance: Resistance: 2.682145076,log10 bioimpedance (reactance): 1.748188027,waist to hip ratio: 0.95049505,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.342422681,log10 creatinine (umol/l): 1.892094603,log10 total cholesterol (mmol/l): 0.809559715,log10 ldl cholesterol (mmol/l): 0.617000341,log10 hdl cholesterol (mmol/l): 0.198657087,log10 total triglycerides (mmol/l): -0.346787486,log10 serum apoa1 (g/l): 0.173186268,log10 serum apob (g/l): 0.053078443,log10 urinary albumin excretion rate (ug/min): 1.06694679
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098216
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249115,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978043
GSM1098216
GSE45159
0.307238
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1019
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249115
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978043
1
age: 53,tissue: adipose tissue,log10 body mass index: 1.390438731,log10 basal metabolic rate (kcal): 1641,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.542620469,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.74798165,plasma free fatty acids under the curve ogtt (mmol/l * min): 13.05,fat mass (%): 19.8,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.651051691,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 132,log10 homair (insulin resistance index based on homa): 0.369133362,log10 homais (insulin secretion index based on homa): 1.951938555,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.585990584,insgenin (insulinogenic index): 2.216045882,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.502140227,log10 matsuda insulin sensitivity index: 0.671973688,muscle mass (%): 45.5,lg10 serum c-reactive protein (mg/l): 0.421932813,lg10 plasma adiponectin (mg/l): 0.707570176,ogtt fasting plasma free fatty acid (mmol/l): 0.25,ogtt 30 min plasma free fatty acid (mmol/l): 0.14,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 7.8,ogtt 120 min plasma glucose (mmol/l): 5.6,log10 il1 receptor antagonist (pg/ml): 2.044735697,log10 il1 beta (pg/ml): -0.795880017,log10 ogtt fasting plasma insulin (mu/l): 0.973127854,ogtt 30 min plasma insulin (mu/l): 1.843232778,ogtt 120 min plasma insulin (mu/l): 1.669316881,log10 ogtt fasting plasma proinsulin (pm/l): 1.025305865,ogtt 30 min plasma proinsulin (pm/l): 1.584331224,ogtt 120 min plasma proinsulin (pm/l): 1.796574333,log10 bioimpedance: Resistance: 2.712649702,log10 bioimpedance (reactance): 1.73239376,waist to hip ratio: 0.916256158,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.397940009,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.725911632,log10 ldl cholesterol (mmol/l): 0.519827994,log10 hdl cholesterol (mmol/l): 0.130333768,log10 total triglycerides (mmol/l): 0.004321374,log10 serum apoa1 (g/l): 0.155336037,log10 serum apob (g/l): -0.022276395,log10 urinary albumin excretion rate (ug/min): 1.245923737
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098217
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249116,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978044
GSM1098217
GSE45159
0.023671
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1086
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249116
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978044
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.427737755,log10 basal metabolic rate (kcal): 1605,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.480201657,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.671088803,plasma free fatty acids under the curve ogtt (mmol/l * min): 15.45,fat mass (%): 17.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.45224124,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 64.5,log10 homair (insulin resistance index based on homa): -0.180748617,log10 homais (insulin secretion index based on homa): 1.492915522,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.419922721,insgenin (insulinogenic index): 2.440909082,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.385284455,log10 matsuda insulin sensitivity index: 1.075030415,muscle mass (%): 45.5,lg10 serum c-reactive protein (mg/l): 0.244771761,lg10 plasma adiponectin (mg/l): 0.531478917,ogtt fasting plasma free fatty acid (mmol/l): 0.45,ogtt 30 min plasma free fatty acid (mmol/l): 0.13,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.3,ogtt 30 min plasma glucose (mmol/l): 6.6,ogtt 120 min plasma glucose (mmol/l): 5,log10 il1 receptor antagonist (pg/ml): 2.06669855,log10 il1 beta (pg/ml): -0.853871964,log10 ogtt fasting plasma insulin (mu/l): 0.447158031,ogtt 30 min plasma insulin (mu/l): 1.796574333,ogtt 120 min plasma insulin (mu/l): 1.113943352,log10 ogtt fasting plasma proinsulin (pm/l): 0.72427587,ogtt 30 min plasma proinsulin (pm/l): 1.495544338,ogtt 120 min plasma proinsulin (pm/l): 1.534026106,log10 bioimpedance: Resistance: 2.603144373,log10 bioimpedance (reactance): 1.556302501,waist to hip ratio: 0.897959184,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.342422681,log10 creatinine (umol/l): 1.913813852,log10 total cholesterol (mmol/l): 0.568201724,log10 ldl cholesterol (mmol/l): 0.365487985,log10 hdl cholesterol (mmol/l): -0.013228266,log10 total triglycerides (mmol/l): -0.102372909,log10 serum apoa1 (g/l): 0.017033339,log10 serum apob (g/l): -0.148741651,log10 urinary albumin excretion rate (ug/min): 0.599768195
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098218
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249117,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978045
GSM1098218
GSE45159
0.229011
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1092
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249117
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978045
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.387714902,log10 basal metabolic rate (kcal): 1650,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.496036167,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.622149964,plasma free fatty acids under the curve ogtt (mmol/l * min): 26.4,fat mass (%): 15.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.13378441,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 415.5,log10 homair (insulin resistance index based on homa): 0.293730757,log10 homais (insulin secretion index based on homa): 1.795880017,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.612062965,insgenin (insulinogenic index): 1.810922422,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.550619653,log10 matsuda insulin sensitivity index: 0.619624223,muscle mass (%): 47.9,lg10 serum c-reactive protein (mg/l): -0.510041521,lg10 plasma adiponectin (mg/l): 0.672097858,ogtt fasting plasma free fatty acid (mmol/l): 0.55,ogtt 30 min plasma free fatty acid (mmol/l): 0.28,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 9.6,ogtt 120 min plasma glucose (mmol/l): 10.2,log10 il1 receptor antagonist (pg/ml): 1.777716739,log10 il1 beta (pg/ml): -0.585026652,log10 ogtt fasting plasma insulin (mu/l): 0.875061263,ogtt 30 min plasma insulin (mu/l): 1.675778342,ogtt 120 min plasma insulin (mu/l): 1.933993164,log10 ogtt fasting plasma proinsulin (pm/l): 1.075546961,ogtt 30 min plasma proinsulin (pm/l): 1.654176542,ogtt 120 min plasma proinsulin (pm/l): 1.879669206,log10 bioimpedance: Resistance: 2.636487896,log10 bioimpedance (reactance): 1.633468456,waist to hip ratio: 0.928205128,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.447158031,log10 creatinine (umol/l): 1.913813852,log10 total cholesterol (mmol/l): 0.571708832,log10 ldl cholesterol (mmol/l): 0.298853076,log10 hdl cholesterol (mmol/l): 0.117271296,log10 total triglycerides (mmol/l): 0.245512668,log10 serum apoa1 (g/l): 0.133538908,log10 serum apob (g/l): -0.102372909,log10 urinary albumin excretion rate (ug/min): 0.713910354
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098219
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249118,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978046
GSM1098219
GSE45159
0.251424
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1111
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249118
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978046
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.463304683,log10 basal metabolic rate (kcal): 1468,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.673759129,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.856303676,plasma free fatty acids under the curve ogtt (mmol/l * min): 34.2,fat mass (%): 22.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.03617361,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 222,log10 homair (insulin resistance index based on homa): 0.468937806,log10 homais (insulin secretion index based on homa): 1.751822312,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.525951341,insgenin (insulinogenic index): 1.739233184,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.425827564,log10 matsuda insulin sensitivity index: 0.57716663,muscle mass (%): 42.1,lg10 serum c-reactive protein (mg/l): 0.27392678,lg10 plasma adiponectin (mg/l): 0.380211242,ogtt fasting plasma free fatty acid (mmol/l): 0.65,ogtt 30 min plasma free fatty acid (mmol/l): 0.37,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 6.9,ogtt 30 min plasma glucose (mmol/l): 9.7,ogtt 120 min plasma glucose (mmol/l): 8.1,log10 il1 receptor antagonist (pg/ml): 2.348285388,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.982271233,ogtt 30 min plasma insulin (mu/l): 1.546542663,ogtt 120 min plasma insulin (mu/l): 1.870403905,log10 ogtt fasting plasma proinsulin (pm/l): 1.149219113,ogtt 30 min plasma proinsulin (pm/l): 1.447158031,ogtt 120 min plasma proinsulin (pm/l): 1.84260924,log10 bioimpedance: Resistance: 2.678518379,log10 bioimpedance (reactance): 1.672097858,waist to hip ratio: 0.94,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.591064607,log10 creatinine (umol/l): 1.86332286,log10 total cholesterol (mmol/l): 0.791690649,log10 ldl cholesterol (mmol/l): 0.57634135,log10 hdl cholesterol (mmol/l): 0.250420002,log10 total triglycerides (mmol/l): 0.123851641,log10 serum apoa1 (g/l): 0.220108088,log10 serum apob (g/l): 0.045322979,log10 urinary albumin excretion rate (ug/min): 0.700546393
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098220
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249119,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978047
GSM1098220
GSE45159
0.320832
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1160
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249119
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978047
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.283277711,log10 basal metabolic rate (kcal): 1607,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.367816206,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.251302989,plasma free fatty acids under the curve ogtt (mmol/l * min): 20.25,fat mass (%): 12.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.653740779,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 217.5,log10 homair (insulin resistance index based on homa): -0.281775196,log10 homais (insulin secretion index based on homa): 1.535113202,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.264510876,insgenin (insulinogenic index): 1.70410412,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.221648928,log10 matsuda insulin sensitivity index: 1.171841611,muscle mass (%): 50,lg10 serum c-reactive protein (mg/l): -0.853871964,lg10 plasma adiponectin (mg/l): 0.832508913,ogtt fasting plasma free fatty acid (mmol/l): 0.33,ogtt 30 min plasma free fatty acid (mmol/l): 0.24,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 4.9,ogtt 30 min plasma glucose (mmol/l): 8.6,ogtt 120 min plasma glucose (mmol/l): 4.8,log10 il1 receptor antagonist (pg/ml): 2.226728757,log10 il1 beta (pg/ml): -0.769551079,log10 ogtt fasting plasma insulin (mu/l): 0.380211242,ogtt 30 min plasma insulin (mu/l): 1.526339277,ogtt 120 min plasma insulin (mu/l): 1.352182518,log10 ogtt fasting plasma proinsulin (pm/l): 0.819543936,ogtt 30 min plasma proinsulin (pm/l): 1.1430148,ogtt 120 min plasma proinsulin (pm/l): 1.600972896,log10 bioimpedance: Resistance: 2.754348336,log10 bioimpedance (reactance): 1.755874856,waist to hip ratio: 0.869565217,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 0.954242509,log10 creatinine (umol/l): 1.944482672,log10 total cholesterol (mmol/l): 0.691965103,log10 ldl cholesterol (mmol/l): 0.498310554,log10 hdl cholesterol (mmol/l): 0.222716471,log10 total triglycerides (mmol/l): -0.124938737,log10 serum apoa1 (g/l): 0.068185862,log10 serum apob (g/l): -0.142667504,log10 urinary albumin excretion rate (ug/min): 0.49785074
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098221
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249120,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978048
GSM1098221
GSE45159
0.239579
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1162
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249120
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978048
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.490049526,log10 basal metabolic rate (kcal): 1504,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.977929407,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.289010918,plasma free fatty acids under the curve ogtt (mmol/l * min): 31.05,fat mass (%): 25,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.04028972,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 405,log10 homair (insulin resistance index based on homa): 0.602927713,log10 homais (insulin secretion index based on homa): 2.244992866,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.83943405,insgenin (insulinogenic index): 1.951013539,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.756400263,log10 matsuda insulin sensitivity index: 0.360849371,muscle mass (%): 40.4,lg10 serum c-reactive protein (mg/l): 0.244771761,lg10 plasma adiponectin (mg/l): 0.740362689,ogtt fasting plasma free fatty acid (mmol/l): 0.56,ogtt 30 min plasma free fatty acid (mmol/l): 0.34,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 9,ogtt 120 min plasma glucose (mmol/l): 9.6,log10 il1 receptor antagonist (pg/ml): 2.578536067,log10 il1 beta (pg/ml): -0.283996656,log10 ogtt fasting plasma insulin (mu/l): 1.222716471,ogtt 30 min plasma insulin (mu/l): 1.846955325,ogtt 120 min plasma insulin (mu/l): 2.194791758,log10 ogtt fasting plasma proinsulin (pm/l): 1.103803721,ogtt 30 min plasma proinsulin (pm/l): 1.307496038,ogtt 120 min plasma proinsulin (pm/l): 1.750508395,log10 bioimpedance: Resistance: 2.669316881,log10 bioimpedance (reactance): 1.633468456,waist to hip ratio: 1.014150943,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.431363764,log10 creatinine (umol/l): 1.785329835,log10 total cholesterol (mmol/l): 0.725911632,log10 ldl cholesterol (mmol/l): 0.549003262,log10 hdl cholesterol (mmol/l): 0.093421685,log10 total triglycerides (mmol/l): 0.204119983,log10 serum apoa1 (g/l): 0.093421685,log10 serum apob (g/l): 0.045322979,log10 urinary albumin excretion rate (ug/min): 0.768391413
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098222
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249121,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978049
GSM1098222
GSE45159
0.314603
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1165
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249121
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978049
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.319423937,log10 basal metabolic rate (kcal): 1483,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.898000295,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.710734607,plasma free fatty acids under the curve ogtt (mmol/l * min): 13.5,fat mass (%): 14.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.533329732,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 129,log10 homair (insulin resistance index based on homa): -0.076410618,log10 homais (insulin secretion index based on homa): 1.665111737,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.101918834,insgenin (insulinogenic index): 1.598267002,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.999174056,log10 matsuda insulin sensitivity index: 1.146953948,muscle mass (%): 47.4,lg10 serum c-reactive protein (mg/l): -0.003050752,lg10 plasma adiponectin (mg/l): 0.740362689,ogtt fasting plasma free fatty acid (mmol/l): 0.24,ogtt 30 min plasma free fatty acid (mmol/l): 0.15,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.1,ogtt 30 min plasma glucose (mmol/l): 7.4,ogtt 120 min plasma glucose (mmol/l): 4.9,log10 il1 receptor antagonist (pg/ml): 2.157758886,log10 il1 beta (pg/ml): -0.920818754,log10 ogtt fasting plasma insulin (mu/l): 0.568201724,ogtt 30 min plasma insulin (mu/l): 1.276461804,ogtt 120 min plasma insulin (mu/l): 1.309630167,log10 ogtt fasting plasma proinsulin (pm/l): 1,ogtt 30 min plasma proinsulin (pm/l): 1.11058971,ogtt 120 min plasma proinsulin (pm/l): 1.348304863,log10 bioimpedance: Resistance: 2.748188027,log10 bioimpedance (reactance): 1.73239376,waist to hip ratio: 0.928961749,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.431363764,log10 creatinine (umol/l): 1.806179974,log10 total cholesterol (mmol/l): 0.704150517,log10 ldl cholesterol (mmol/l): 0.511883361,log10 hdl cholesterol (mmol/l): 0.181843588,log10 total triglycerides (mmol/l): -0.180456064,log10 serum apoa1 (g/l): 0.117271296,log10 serum apob (g/l): -0.080921908,log10 urinary albumin excretion rate (ug/min): 0.783903579
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098223
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249122,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978050
GSM1098223
GSE45159
0.257102
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1167
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249122
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978050
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.465941054,log10 basal metabolic rate (kcal): 1624,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.149075129,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.507046139,plasma free fatty acids under the curve ogtt (mmol/l * min): 21.15,fat mass (%): 27.6,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.946906274,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 267,log10 homair (insulin resistance index based on homa): 0.174156759,log10 homais (insulin secretion index based on homa): 1.651278014,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.386730765,insgenin (insulinogenic index): 1.703430064,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.30815874,log10 matsuda insulin sensitivity index: 0.818178344,muscle mass (%): 40.1,lg10 serum c-reactive protein (mg/l): -0.232844134,lg10 plasma adiponectin (mg/l): 0.755874856,ogtt fasting plasma free fatty acid (mmol/l): 0.36,ogtt 30 min plasma free fatty acid (mmol/l): 0.24,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 6,ogtt 30 min plasma glucose (mmol/l): 9.1,ogtt 120 min plasma glucose (mmol/l): 7.8,log10 il1 receptor antagonist (pg/ml): 2.398356731,log10 il1 beta (pg/ml): 0.100370545,log10 ogtt fasting plasma insulin (mu/l): 0.748188027,ogtt 30 min plasma insulin (mu/l): 1.501059262,ogtt 120 min plasma insulin (mu/l): 1.663700925,log10 ogtt fasting plasma proinsulin (pm/l): 1.123851641,ogtt 30 min plasma proinsulin (pm/l): 1.367355921,ogtt 120 min plasma proinsulin (pm/l): 1.674861141,log10 bioimpedance: Resistance: 2.619093331,log10 bioimpedance (reactance): 1.633468456,waist to hip ratio: 0.957746479,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.204119983,log10 creatinine (umol/l): 2.004321374,log10 total cholesterol (mmol/l): 0.592176757,log10 ldl cholesterol (mmol/l): 0.357934847,log10 hdl cholesterol (mmol/l): 0.037426498,log10 total triglycerides (mmol/l): 0,log10 serum apoa1 (g/l): 0.004321374,log10 serum apob (g/l): -0.214670165,log10 urinary albumin excretion rate (ug/min): 0.90045523
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098224
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249123,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978051
GSM1098224
GSE45159
0.166311
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1180
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249123
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978051
1
age: 51,tissue: adipose tissue,log10 body mass index: 1.378213836,log10 basal metabolic rate (kcal): 1510,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.339395984,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.423859499,plasma free fatty acids under the curve ogtt (mmol/l * min): 20.25,fat mass (%): 18.8,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.632086413,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 145.5,log10 homair (insulin resistance index based on homa): 0.57333584,log10 homais (insulin secretion index based on homa): 2.215400993,log10 insulin area under the curve (ogtt) (pmol/l * min): 5.206804296,insgenin (insulinogenic index): 2.870571102,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 5.175395831,log10 matsuda insulin sensitivity index: 0.282185572,muscle mass (%): 45.9,lg10 serum c-reactive protein (mg/l): 0.529173603,lg10 plasma adiponectin (mg/l): 0.662757832,ogtt fasting plasma free fatty acid (mmol/l): 0.35,ogtt 30 min plasma free fatty acid (mmol/l): 0.22,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 8.2,ogtt 120 min plasma glucose (mmol/l): 4.9,log10 il1 receptor antagonist (pg/ml): 2.484641445,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 1.193124598,ogtt 30 min plasma insulin (mu/l): 2.558708571,ogtt 120 min plasma insulin (mu/l): 2.035029282,log10 ogtt fasting plasma proinsulin (pm/l): 1.103803721,ogtt 30 min plasma proinsulin (pm/l): 1.759667845,ogtt 120 min plasma proinsulin (pm/l): 1.808885867,log10 bioimpedance: Resistance: 2.73479983,log10 bioimpedance (reactance): 1.770852012,waist to hip ratio: 1.016042781,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.477121255,log10 creatinine (umol/l): 1.959041392,log10 total cholesterol (mmol/l): 0.744292983,log10 ldl cholesterol (mmol/l): 0.394451681,log10 hdl cholesterol (mmol/l): 0.029383778,log10 total triglycerides (mmol/l): 0.758911892,log10 serum apoa1 (g/l): 0.220108088,log10 serum apob (g/l): 0.235528447,log10 urinary albumin excretion rate (ug/min): 0.95248423
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098225
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249124,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978052
GSM1098225
GSE45159
0.200507
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1223
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249124
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978052
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.475868926,log10 basal metabolic rate (kcal): 1738,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 5.971589741,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.51279523,plasma free fatty acids under the curve ogtt (mmol/l * min): 18.6,fat mass (%): 22.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.35645197,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 67.5,log10 homair (insulin resistance index based on homa): -0.360956442,log10 homais (insulin secretion index based on homa): 1.455931956,log10 insulin area under the curve (ogtt) (pmol/l * min): 3.990072335,insgenin (insulinogenic index): 1.922206277,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.920853496,log10 matsuda insulin sensitivity index: 1.390450291,muscle mass (%): 43.3,lg10 serum c-reactive protein (mg/l): 0.682325619,lg10 plasma adiponectin (mg/l): 1.187520721,ogtt fasting plasma free fatty acid (mmol/l): 0.2,ogtt 30 min plasma free fatty acid (mmol/l): 0.2,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 4.9,ogtt 30 min plasma glucose (mmol/l): 6.4,ogtt 120 min plasma glucose (mmol/l): 4.4,log10 il1 receptor antagonist (pg/ml): 2.091209565,log10 il1 beta (pg/ml): -0.602059991,log10 ogtt fasting plasma insulin (mu/l): 0.301029996,ogtt 30 min plasma insulin (mu/l): 1.359835482,ogtt 120 min plasma insulin (mu/l): 0.698970004,log10 ogtt fasting plasma proinsulin (pm/l): 0.924279286,ogtt 30 min plasma proinsulin (pm/l): 1.100370545,ogtt 120 min plasma proinsulin (pm/l): 1.149219113,log10 bioimpedance: Resistance: 2.62324929,log10 bioimpedance (reactance): 1.62324929,waist to hip ratio: 0.930555556,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.041392685,log10 creatinine (umol/l): 2.049218023,log10 total cholesterol (mmol/l): 0.692846919,log10 ldl cholesterol (mmol/l): 0.450249108,log10 hdl cholesterol (mmol/l): 0.247973266,log10 total triglycerides (mmol/l): -0.119186408,log10 serum apoa1 (g/l): 0.152288344,log10 serum apob (g/l): -0.102372909,log10 urinary albumin excretion rate (ug/min): 0.752552832
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098226
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249125,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978053
GSM1098226
GSE45159
0.26757
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1228
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249125
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978053
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.446063944,log10 basal metabolic rate (kcal): 1828,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.68421978,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.031315588,plasma free fatty acids under the curve ogtt (mmol/l * min): 11.25,fat mass (%): 19,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.401946124,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 16.5,log10 homair (insulin resistance index based on homa): 0.187520721,log10 homais (insulin secretion index based on homa): 1.799340549,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.571032996,insgenin (insulinogenic index): 2.51243325,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.514627432,log10 matsuda insulin sensitivity index: 0.826593247,muscle mass (%): 47,lg10 serum c-reactive protein (mg/l): 0.35545152,lg10 plasma adiponectin (mg/l): 0.819543936,ogtt fasting plasma free fatty acid (mmol/l): 0.16,ogtt 30 min plasma free fatty acid (mmol/l): 0.08,ogtt 120 min plasma free fatty acid (mmol/l): 0.09,ogtt fasting plasma glucose (mmol/l): 5.5,ogtt 30 min plasma glucose (mmol/l): 7.2,ogtt 120 min plasma glucose (mmol/l): 3.6,log10 il1 receptor antagonist (pg/ml): 2.047664195,log10 il1 beta (pg/ml): -0.522878745,log10 ogtt fasting plasma insulin (mu/l): 0.799340549,ogtt 30 min plasma insulin (mu/l): 1.99343623,ogtt 120 min plasma insulin (mu/l): 0.653212514,log10 ogtt fasting plasma proinsulin (pm/l): 1.08278537,ogtt 30 min plasma proinsulin (pm/l): 1.357934847,ogtt 120 min plasma proinsulin (pm/l): 1.133538908,log10 bioimpedance: Resistance: 2.592176757,log10 bioimpedance (reactance): 1.62324929,waist to hip ratio: 0.888888889,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.322219295,log10 creatinine (umol/l): 1.86332286,log10 total cholesterol (mmol/l): 0.665580991,log10 ldl cholesterol (mmol/l): 0.450249108,log10 hdl cholesterol (mmol/l): 0.204119983,log10 total triglycerides (mmol/l): -0.086186148,log10 serum apoa1 (g/l): 0.158362492,log10 serum apob (g/l): -0.086186148,log10 urinary albumin excretion rate (ug/min): 1.164264052
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098227
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249126,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978054
GSM1098227
GSE45159
0.221683
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1247
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249126
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978054
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.517639221,log10 basal metabolic rate (kcal): 1710,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.526495272,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.167338599,plasma free fatty acids under the curve ogtt (mmol/l * min): 29.55,fat mass (%): 26.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.949097156,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 280.5,log10 homair (insulin resistance index based on homa): 0.683487317,log10 homais (insulin secretion index based on homa): 2.185636577,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.937282125,insgenin (insulinogenic index): 2.171726454,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.865133598,log10 matsuda insulin sensitivity index: 0.281999031,muscle mass (%): 39.9,lg10 serum c-reactive protein (mg/l): -0.209714836,lg10 plasma adiponectin (mg/l): 0.86923172,ogtt fasting plasma free fatty acid (mmol/l): 0.46,ogtt 30 min plasma free fatty acid (mmol/l): 0.34,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 8.7,ogtt 120 min plasma glucose (mmol/l): 8.4,log10 il1 receptor antagonist (pg/ml): 2.211200662,log10 il1 beta (pg/ml): -0.619788758,log10 ogtt fasting plasma insulin (mu/l): 1.264817823,ogtt 30 min plasma insulin (mu/l): 1.942999593,ogtt 120 min plasma insulin (mu/l): 2.2955671,log10 ogtt fasting plasma proinsulin (pm/l): 1.46834733,ogtt 30 min plasma proinsulin (pm/l): 1.706717782,ogtt 120 min plasma proinsulin (pm/l): 1.980457892,log10 bioimpedance: Resistance: 2.62324929,log10 bioimpedance (reactance): 1.568201724,waist to hip ratio: 1.018181818,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.681241237,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.630427875,log10 ldl cholesterol (mmol/l): 0.382017043,log10 hdl cholesterol (mmol/l): 0.190331698,log10 total triglycerides (mmol/l): 0.08278537,log10 serum apoa1 (g/l): 0.184691431,log10 serum apob (g/l): -0.113509275,log10 urinary albumin excretion rate (ug/min): 0.798602876
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098228
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249127,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978055
GSM1098228
GSE45159
0.467875
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1263
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249127
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978055
1
age: 53,tissue: adipose tissue,log10 body mass index: 1.441288134,log10 basal metabolic rate (kcal): 1833,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.171115448,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.602931318,plasma free fatty acids under the curve ogtt (mmol/l * min): 17.4,fat mass (%): 21.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.430452552,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 42,log10 homair (insulin resistance index based on homa): 0.104487111,log10 homais (insulin secretion index based on homa): 1.746552264,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.779213971,insgenin (insulinogenic index): 2.629965318,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.750747461,log10 matsuda insulin sensitivity index: 0.76782338,muscle mass (%): 46.6,lg10 serum c-reactive protein (mg/l): 0.124830149,lg10 plasma adiponectin (mg/l): 0.832508913,ogtt fasting plasma free fatty acid (mmol/l): 0.34,ogtt 30 min plasma free fatty acid (mmol/l): 0.16,ogtt 120 min plasma free fatty acid (mmol/l): 0.06,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 7.6,ogtt 120 min plasma glucose (mmol/l): 3.4,log10 il1 receptor antagonist (pg/ml): 2.067554377,log10 il1 beta (pg/ml): -0.657577319,log10 ogtt fasting plasma insulin (mu/l): 0.72427587,ogtt 30 min plasma insulin (mu/l): 2.20871002,ogtt 120 min plasma insulin (mu/l): 0.73239376,log10 ogtt fasting plasma proinsulin (pm/l): 0.924279286,ogtt 30 min plasma proinsulin (pm/l): 1.589949601,ogtt 120 min plasma proinsulin (pm/l): 1.298853076,log10 bioimpedance: Resistance: 2.662757832,log10 bioimpedance (reactance): 1.763427994,waist to hip ratio: 0.954545455,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.322219295,log10 creatinine (umol/l): 2,log10 total cholesterol (mmol/l): 0.814913181,log10 ldl cholesterol (mmol/l): 0.652246341,log10 hdl cholesterol (mmol/l): 0.149219113,log10 total triglycerides (mmol/l): 0.26245109,log10 serum apoa1 (g/l): 0.245512668,log10 serum apob (g/l): 0.146128036,log10 urinary albumin excretion rate (ug/min): 0.304490528
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098229
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249128,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978056
GSM1098229
GSE45159
0.25875
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1264
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249128
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978056
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.41701389,log10 basal metabolic rate (kcal): 1670,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.656291109,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.800841523,plasma free fatty acids under the curve ogtt (mmol/l * min): 12.6,fat mass (%): 18.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.811374694,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 202.5,log10 homair (insulin resistance index based on homa): 0.151608165,log10 homais (insulin secretion index based on homa): 1.679664849,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.23482088,insgenin (insulinogenic index): 1.577236408,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.120968565,log10 matsuda insulin sensitivity index: 0.951756747,muscle mass (%): 47.3,lg10 serum c-reactive protein (mg/l): -0.54515514,lg10 plasma adiponectin (mg/l): 0.832508913,ogtt fasting plasma free fatty acid (mmol/l): 0.31,ogtt 30 min plasma free fatty acid (mmol/l): 0.11,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.8,ogtt 30 min plasma glucose (mmol/l): 11.2,ogtt 120 min plasma glucose (mmol/l): 3.1,log10 il1 receptor antagonist (pg/ml): 2.313761796,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.740362689,ogtt 30 min plasma insulin (mu/l): 1.596597096,ogtt 120 min plasma insulin (mu/l): 0.959041392,log10 ogtt fasting plasma proinsulin (pm/l): 0.880813592,ogtt 30 min plasma proinsulin (pm/l): 1.170261715,ogtt 120 min plasma proinsulin (pm/l): 1.257678575,log10 bioimpedance: Resistance: 2.645422269,log10 bioimpedance (reactance): 1.69019608,waist to hip ratio: 1,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.230448921,log10 creatinine (umol/l): 1.86923172,log10 total cholesterol (mmol/l): 0.691965103,log10 ldl cholesterol (mmol/l): 0.471291711,log10 hdl cholesterol (mmol/l): 0.222716471,log10 total triglycerides (mmol/l): -0.107905397,log10 serum apoa1 (g/l): 0.195899652,log10 serum apob (g/l): -0.060480747,log10 urinary albumin excretion rate (ug/min): 0.551754873
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098230
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249129,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978057
GSM1098230
GSE45159
0.07069
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1303
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249129
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978057
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.373741634,log10 basal metabolic rate (kcal): 1582,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.359654801,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.416505697,plasma free fatty acids under the curve ogtt (mmol/l * min): 29.1,fat mass (%): 14.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.742309436,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 256.5,log10 homair (insulin resistance index based on homa): -0.08501079,log10 homais (insulin secretion index based on homa): 1.693140461,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.295281141,insgenin (insulinogenic index): 1.782950133,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.23230984,log10 matsuda insulin sensitivity index: 1.046350387,muscle mass (%): 48,lg10 serum c-reactive protein (mg/l): -0.554395797,lg10 plasma adiponectin (mg/l): 0.72427587,ogtt fasting plasma free fatty acid (mmol/l): 0.66,ogtt 30 min plasma free fatty acid (mmol/l): 0.29,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5,ogtt 30 min plasma glucose (mmol/l): 8.6,ogtt 120 min plasma glucose (mmol/l): 5.9,log10 il1 receptor antagonist (pg/ml): 2.167346878,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.568201724,ogtt 30 min plasma insulin (mu/l): 1.603144373,ogtt 120 min plasma insulin (mu/l): 1.264817823,log10 ogtt fasting plasma proinsulin (pm/l): 0.880813592,ogtt 30 min plasma proinsulin (pm/l): 1.307496038,ogtt 120 min plasma proinsulin (pm/l): 1.426511261,log10 bioimpedance: Resistance: 2.661812686,log10 bioimpedance (reactance): 1.662757832,waist to hip ratio: 0.955555556,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.662757832,log10 creatinine (umol/l): 1.949390007,log10 total cholesterol (mmol/l): 0.6599162,log10 ldl cholesterol (mmol/l): 0.354108439,log10 hdl cholesterol (mmol/l): 0.334453751,log10 total triglycerides (mmol/l): -0.214670165,log10 serum apoa1 (g/l): 0.227886705,log10 serum apob (g/l): -0.187086643,log10 urinary albumin excretion rate (ug/min): 0.454106891
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098231
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249130,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978058
GSM1098231
GSE45159
0.2848
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1343
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249130
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978058
1
age: 51,tissue: adipose tissue,log10 body mass index: 1.396434654,log10 basal metabolic rate (kcal): 1503,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.553885977,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.619779709,plasma free fatty acids under the curve ogtt (mmol/l * min): 34.05,fat mass (%): 18.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.842350343,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 186,log10 homair (insulin resistance index based on homa): 0.114388554,log10 homais (insulin secretion index based on homa): 1.567297885,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.374363358,insgenin (insulinogenic index): 1.813128834,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.305845582,log10 matsuda insulin sensitivity index: 0.87703081,muscle mass (%): 46.1,lg10 serum c-reactive protein (mg/l): 0.550350672,lg10 plasma adiponectin (mg/l): 0.556302501,ogtt fasting plasma free fatty acid (mmol/l): 0.39,ogtt 30 min plasma free fatty acid (mmol/l): 0.41,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 6.1,ogtt 30 min plasma glucose (mmol/l): 9.2,ogtt 120 min plasma glucose (mmol/l): 6.1,log10 il1 receptor antagonist (pg/ml): 2.317917154,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.681241237,ogtt 30 min plasma insulin (mu/l): 1.584331224,ogtt 120 min plasma insulin (mu/l): 1.542825427,log10 ogtt fasting plasma proinsulin (pm/l): 1.053078443,ogtt 30 min plasma proinsulin (pm/l): 1.257678575,ogtt 120 min plasma proinsulin (pm/l): 1.615950052,log10 bioimpedance: Resistance: 2.696356389,log10 bioimpedance (reactance): 1.72427587,waist to hip ratio: 1.027322404,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.51851394,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.726727209,log10 ldl cholesterol (mmol/l): 0.561101384,log10 hdl cholesterol (mmol/l): 0.086359831,log10 total triglycerides (mmol/l): -0.036212173,log10 serum apoa1 (g/l): 0.11058971,log10 serum apob (g/l): 0.004321374,log10 urinary albumin excretion rate (ug/min): 0.86923172
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098232
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249131,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978059
GSM1098232
GSE45159
0.226614
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1353
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249131
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978059
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.466297045,log10 basal metabolic rate (kcal): 1691,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.392120527,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.738230244,plasma free fatty acids under the curve ogtt (mmol/l * min): 23.7,fat mass (%): 22.6,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.940313597,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 226.5,log10 homair (insulin resistance index based on homa): 0.605520523,log10 homais (insulin secretion index based on homa): 2.012234456,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.529545843,insgenin (insulinogenic index): 1.779235632,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.370716589,log10 matsuda insulin sensitivity index: 0.532221588,muscle mass (%): 45,lg10 serum c-reactive protein (mg/l): 0.301029996,lg10 plasma adiponectin (mg/l): 0.612783857,ogtt fasting plasma free fatty acid (mmol/l): 0.42,ogtt 30 min plasma free fatty acid (mmol/l): 0.23,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 6.3,ogtt 30 min plasma glucose (mmol/l): 10.3,ogtt 120 min plasma glucose (mmol/l): 6,log10 il1 receptor antagonist (pg/ml): 2.60148417,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 1.158362492,ogtt 30 min plasma insulin (mu/l): 1.736396502,ogtt 120 min plasma insulin (mu/l): 1.680335513,log10 ogtt fasting plasma proinsulin (pm/l): 1.152288344,ogtt 30 min plasma proinsulin (pm/l): 1.318063335,ogtt 120 min plasma proinsulin (pm/l): 1.51851394,log10 bioimpedance: Resistance: 2.651278014,log10 bioimpedance (reactance): 1.72427587,waist to hip ratio: 1.02,log10 serum bilirubin (umol/l): NA,log10 serum alanine aminotransfrase (u/l): 1.579783597,log10 creatinine (umol/l): 1.939519253,log10 total cholesterol (mmol/l): 0.749736316,log10 ldl cholesterol (mmol/l): 0.640481437,log10 hdl cholesterol (mmol/l): -0.036212173,log10 total triglycerides (mmol/l): 0.309630167,log10 serum apoa1 (g/l): 0.071882007,log10 serum apob (g/l): 0.130333768,log10 urinary albumin excretion rate (ug/min): 0.712884519
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098233
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249132,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978060
GSM1098233
GSE45159
0.184536
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1366
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249132
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978060
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.4863247,log10 basal metabolic rate (kcal): 1636,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.748807692,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.144046278,plasma free fatty acids under the curve ogtt (mmol/l * min): 28.35,fat mass (%): 22.7,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.703903573,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 150,log10 homair (insulin resistance index based on homa): 0.385963571,log10 homais (insulin secretion index based on homa): 1.940878548,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.608804064,insgenin (insulinogenic index): 2.191768133,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.527810282,log10 matsuda insulin sensitivity index: 0.668479802,muscle mass (%): 42.2,lg10 serum c-reactive protein (mg/l): -0.160521953,lg10 plasma adiponectin (mg/l): 0.698970004,ogtt fasting plasma free fatty acid (mmol/l): 0.45,ogtt 30 min plasma free fatty acid (mmol/l): 0.3,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 5.7,ogtt 30 min plasma glucose (mmol/l): 9.4,ogtt 120 min plasma glucose (mmol/l): 4.1,log10 il1 receptor antagonist (pg/ml): 2.395169075,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.982271233,ogtt 30 min plasma insulin (mu/l): 2.02325246,ogtt 120 min plasma insulin (mu/l): 0.819543936,log10 ogtt fasting plasma proinsulin (pm/l): 1.086359831,ogtt 30 min plasma proinsulin (pm/l): 1.645422269,ogtt 120 min plasma proinsulin (pm/l): 1.403120521,log10 bioimpedance: Resistance: 2.609594409,log10 bioimpedance (reactance): 1.568201724,waist to hip ratio: 1.028571429,log10 serum bilirubin (umol/l): 1.544068044,log10 serum alanine aminotransfrase (u/l): 1.681241237,log10 creatinine (umol/l): 1.84509804,log10 total cholesterol (mmol/l): 0.655138435,log10 ldl cholesterol (mmol/l): 0.480006943,log10 hdl cholesterol (mmol/l): 0.056904851,log10 total triglycerides (mmol/l): 0.164352856,log10 serum apoa1 (g/l): 0.068185862,log10 serum apob (g/l): 0.012837225,log10 urinary albumin excretion rate (ug/min): 0.934498451
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098234
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249133,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978061
GSM1098234
GSE45159
0.358195
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1452
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249133
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978061
1
age: 59,tissue: adipose tissue,log10 body mass index: 1.38474035,log10 basal metabolic rate (kcal): 1576,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.278526307,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.251888771,plasma free fatty acids under the curve ogtt (mmol/l * min): 15.45,fat mass (%): 12.7,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.582141982,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 58.5,log10 homair (insulin resistance index based on homa): 0.218010043,log10 homais (insulin secretion index based on homa): 1.720159303,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.389449613,insgenin (insulinogenic index): 2.086359831,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.300595484,log10 matsuda insulin sensitivity index: 0.871543768,muscle mass (%): 49.2,lg10 serum c-reactive protein (mg/l): -0.368556231,lg10 plasma adiponectin (mg/l): 0.763427994,ogtt fasting plasma free fatty acid (mmol/l): 0.41,ogtt 30 min plasma free fatty acid (mmol/l): 0.14,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 8.6,ogtt 120 min plasma glucose (mmol/l): 3.6,log10 il1 receptor antagonist (pg/ml): 2.20199762,log10 il1 beta (pg/ml): 0.799340549,log10 ogtt fasting plasma insulin (mu/l): 0.799340549,ogtt 30 min plasma insulin (mu/l): 1.786751422,ogtt 120 min plasma insulin (mu/l): 0.851258349,log10 ogtt fasting plasma proinsulin (pm/l): 0.929418926,ogtt 30 min plasma proinsulin (pm/l): 1.392696953,ogtt 120 min plasma proinsulin (pm/l): 1.488550717,log10 bioimpedance: Resistance: 2.609594409,log10 bioimpedance (reactance): 1.602059991,waist to hip ratio: 0.896907216,log10 serum bilirubin (umol/l): 1.079181246,log10 serum alanine aminotransfrase (u/l): 1.176091259,log10 creatinine (umol/l): 1.963787827,log10 total cholesterol (mmol/l): 0.750508395,log10 ldl cholesterol (mmol/l): 0.563481085,log10 hdl cholesterol (mmol/l): 0.173186268,log10 total triglycerides (mmol/l): 0.037426498,log10 serum apoa1 (g/l): 0.136720567,log10 serum apob (g/l): 0.053078443,log10 urinary albumin excretion rate (ug/min): 1.152118399
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098235
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249134,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978062
GSM1098235
GSE45159
0.015028
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1483
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249134
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978062
1
age: 60,tissue: adipose tissue,log10 body mass index: 1.428052757,log10 basal metabolic rate (kcal): 1622,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.425060246,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.600699475,plasma free fatty acids under the curve ogtt (mmol/l * min): 15.3,fat mass (%): 21.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.839991071,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 256.5,log10 homair (insulin resistance index based on homa): -0.067751784,log10 homais (insulin secretion index based on homa): 1.544068044,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.570402822,insgenin (insulinogenic index): 1.939519253,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.539928788,log10 matsuda insulin sensitivity index: 0.876235596,muscle mass (%): 44.8,lg10 serum c-reactive protein (mg/l): 0.60075515,lg10 plasma adiponectin (mg/l): 0.380211242,ogtt fasting plasma free fatty acid (mmol/l): 0.17,ogtt 30 min plasma free fatty acid (mmol/l): 0.19,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.5,ogtt 30 min plasma glucose (mmol/l): 9.1,ogtt 120 min plasma glucose (mmol/l): 6.4,log10 il1 receptor antagonist (pg/ml): 2.250810204,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.544068044,ogtt 30 min plasma insulin (mu/l): 1.745855195,ogtt 120 min plasma insulin (mu/l): 1.794488047,log10 ogtt fasting plasma proinsulin (pm/l): 1.017033339,ogtt 30 min plasma proinsulin (pm/l): 1.264817823,ogtt 120 min plasma proinsulin (pm/l): 1.648360011,log10 bioimpedance: Resistance: 2.69019608,log10 bioimpedance (reactance): 1.73239376,waist to hip ratio: 1.030612245,log10 serum bilirubin (umol/l): 1.041392685,log10 serum alanine aminotransfrase (u/l): 1.414973348,log10 creatinine (umol/l): 1.924279286,log10 total cholesterol (mmol/l): 0.631443769,log10 ldl cholesterol (mmol/l): 0.440909082,log10 hdl cholesterol (mmol/l): -0.050609993,log10 total triglycerides (mmol/l): -0.060480747,log10 serum apoa1 (g/l): 0.029383778,log10 serum apob (g/l): -0.040958608,log10 urinary albumin excretion rate (ug/min): 0.756156957
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098236
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249135,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978063
GSM1098236
GSE45159
0.189036
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1504
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249135
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978063
1
age: 62,tissue: adipose tissue,log10 body mass index: 1.407947934,log10 basal metabolic rate (kcal): 1420,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.496686471,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.459195841,plasma free fatty acids under the curve ogtt (mmol/l * min): 11.1,fat mass (%): 35,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.491853096,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 84,log10 homair (insulin resistance index based on homa): 0.025305865,log10 homais (insulin secretion index based on homa): 1.698970004,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.288696261,insgenin (insulinogenic index): 1.827070891,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.209515015,log10 matsuda insulin sensitivity index: 1.016750866,muscle mass (%): 39.7,lg10 serum c-reactive protein (mg/l): 0.301029996,lg10 plasma adiponectin (mg/l): 1.004321374,ogtt fasting plasma free fatty acid (mmol/l): 0.2,ogtt 30 min plasma free fatty acid (mmol/l): 0.12,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.3,ogtt 30 min plasma glucose (mmol/l): 7.9,ogtt 120 min plasma glucose (mmol/l): 3.7,log10 il1 receptor antagonist (pg/ml): 2.247899651,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.653212514,ogtt 30 min plasma insulin (mu/l): 1.526339277,ogtt 120 min plasma insulin (mu/l): 1.409933123,log10 ogtt fasting plasma proinsulin (pm/l): 0.919078092,ogtt 30 min plasma proinsulin (pm/l): 1.130333768,ogtt 120 min plasma proinsulin (pm/l): 1.46686762,log10 bioimpedance: Resistance: 2.758154622,log10 bioimpedance (reactance): 1.785329835,waist to hip ratio: 0.989473684,log10 serum bilirubin (umol/l): 1.146128036,log10 serum alanine aminotransfrase (u/l): 1.447158031,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.791690649,log10 ldl cholesterol (mmol/l): 0.574031268,log10 hdl cholesterol (mmol/l): 0.243038049,log10 total triglycerides (mmol/l): 0.049218023,log10 serum apoa1 (g/l): 0.214843848,log10 serum apob (g/l): 0.053078443,log10 urinary albumin excretion rate (ug/min): 0.562041415
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098237
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249136,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978064
GSM1098237
GSE45159
0.177973
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1516
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249136
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978064
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.465206508,log10 basal metabolic rate (kcal): 1845,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.655647312,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.065821111,plasma free fatty acids under the curve ogtt (mmol/l * min): 12.3,fat mass (%): 24.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.747353829,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 103.5,log10 homair (insulin resistance index based on homa): 0.232487866,log10 homais (insulin secretion index based on homa): 1.639201799,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.391675953,insgenin (insulinogenic index): 2.064457989,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.306425028,log10 matsuda insulin sensitivity index: 0.816468919,muscle mass (%): 46.4,lg10 serum c-reactive protein (mg/l): 0.112939976,lg10 plasma adiponectin (mg/l): 0.505149978,ogtt fasting plasma free fatty acid (mmol/l): 0.18,ogtt 30 min plasma free fatty acid (mmol/l): 0.13,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 6.3,ogtt 30 min plasma glucose (mmol/l): 8.1,ogtt 120 min plasma glucose (mmol/l): 6.2,log10 il1 receptor antagonist (pg/ml): 2.138302698,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.785329835,ogtt 30 min plasma insulin (mu/l): 1.611723308,ogtt 120 min plasma insulin (mu/l): 1.540329475,log10 ogtt fasting plasma proinsulin (pm/l): 0.929418926,ogtt 30 min plasma proinsulin (pm/l): 1.201397124,ogtt 120 min plasma proinsulin (pm/l): 1.515873844,log10 bioimpedance: Resistance: 2.695481676,log10 bioimpedance (reactance): 1.875061263,waist to hip ratio: 1.01,log10 serum bilirubin (umol/l): 1.204119983,log10 serum alanine aminotransfrase (u/l): 1.491361694,log10 creatinine (umol/l): 1.875061263,log10 total cholesterol (mmol/l): 0.6599162,log10 ldl cholesterol (mmol/l): 0.488550717,log10 hdl cholesterol (mmol/l): 0.053078443,log10 total triglycerides (mmol/l): 0.025305865,log10 serum apoa1 (g/l): 0.037426498,log10 serum apob (g/l): -0.045757491,log10 urinary albumin excretion rate (ug/min): 0.77003336
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098238
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249137,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978065
GSM1098238
GSE45159
0.150794
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1518
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249137
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978065
1
age: 65,tissue: adipose tissue,log10 body mass index: 1.326961696,log10 basal metabolic rate (kcal): 1199,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.421557173,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 5.877749085,plasma free fatty acids under the curve ogtt (mmol/l * min): 16.95,fat mass (%): 29.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.08081753,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 363,log10 homair (insulin resistance index based on homa): 0.158362492,log10 homais (insulin secretion index based on homa): 1.635483747,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.713305698,insgenin (insulinogenic index): 2.022062253,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.679337031,log10 matsuda insulin sensitivity index: 0.688709281,muscle mass (%): 47.8,lg10 serum c-reactive protein (mg/l): 0.112939976,lg10 plasma adiponectin (mg/l): 1.113943352,ogtt fasting plasma free fatty acid (mmol/l): 0.36,ogtt 30 min plasma free fatty acid (mmol/l): 0.17,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 6,ogtt 30 min plasma glucose (mmol/l): 13.1,ogtt 120 min plasma glucose (mmol/l): 4.6,log10 il1 receptor antagonist (pg/ml): 2.183696809,log10 il1 beta (pg/ml): -0.13667714,log10 ogtt fasting plasma insulin (mu/l): 0.73239376,ogtt 30 min plasma insulin (mu/l): 2.113609151,ogtt 120 min plasma insulin (mu/l): 1.214843848,log10 ogtt fasting plasma proinsulin (pm/l): 0.995635195,ogtt 30 min plasma proinsulin (pm/l): 1.46686762,ogtt 120 min plasma proinsulin (pm/l): 1.411619706,log10 bioimpedance: Resistance: 2.797959644,log10 bioimpedance (reactance): 1.806179974,waist to hip ratio: 0.903225806,log10 serum bilirubin (umol/l): 1.079181246,log10 serum alanine aminotransfrase (u/l): 1.431363764,log10 creatinine (umol/l): 1.919078092,log10 total cholesterol (mmol/l): 0.767155866,log10 ldl cholesterol (mmol/l): 0.507855872,log10 hdl cholesterol (mmol/l): 0.29666519,log10 total triglycerides (mmol/l): 0.267171728,log10 serum apoa1 (g/l): 0.152288344,log10 serum apob (g/l): -0.070581074,log10 urinary albumin excretion rate (ug/min): 0.773742131
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098239
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249138,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978066
GSM1098239
GSE45159
0.219178
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1574
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249138
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978066
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.367376327,log10 basal metabolic rate (kcal): 1545,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.617168971,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.647337353,plasma free fatty acids under the curve ogtt (mmol/l * min): 15.6,fat mass (%): 19.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.849405101,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 214.5,log10 homair (insulin resistance index based on homa): 0.12623967,log10 homais (insulin secretion index based on homa): 1.62838893,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.369549802,insgenin (insulinogenic index): 1.800631173,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.295479133,log10 matsuda insulin sensitivity index: 0.873622119,muscle mass (%): 45.8,lg10 serum c-reactive protein (mg/l): 0.725012725,lg10 plasma adiponectin (mg/l): 1.049218023,ogtt fasting plasma free fatty acid (mmol/l): 0.38,ogtt 30 min plasma free fatty acid (mmol/l): 0.15,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 9.1,ogtt 120 min plasma glucose (mmol/l): 6.4,log10 il1 receptor antagonist (pg/ml): 2.21285319,log10 il1 beta (pg/ml): -0.431798276,log10 ogtt fasting plasma insulin (mu/l): 0.707570176,ogtt 30 min plasma insulin (mu/l): 1.588831726,ogtt 120 min plasma insulin (mu/l): 1.522444234,log10 ogtt fasting plasma proinsulin (pm/l): 1.127104798,ogtt 30 min plasma proinsulin (pm/l): 1.359835482,ogtt 120 min plasma proinsulin (pm/l): 1.646403726,log10 bioimpedance: Resistance: 2.752048448,log10 bioimpedance (reactance): 1.785329835,waist to hip ratio: 0.909090909,log10 serum bilirubin (umol/l): 1.113943352,log10 serum alanine aminotransfrase (u/l): 1.255272505,log10 creatinine (umol/l): 1.880813592,log10 total cholesterol (mmol/l): 0.671172843,log10 ldl cholesterol (mmol/l): 0.338456494,log10 hdl cholesterol (mmol/l): 0.390935107,log10 total triglycerides (mmol/l): -0.207608311,log10 serum apoa1 (g/l): 0.222716471,log10 serum apob (g/l): -0.207608311,log10 urinary albumin excretion rate (ug/min): 0.879860146
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098240
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249139,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978067
GSM1098240
GSE45159
0.300427
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1599
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249139
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978067
1
age: 60,tissue: adipose tissue,log10 body mass index: 1.375963588,log10 basal metabolic rate (kcal): 1791,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.291799537,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.502500249,plasma free fatty acids under the curve ogtt (mmol/l * min): 17.4,fat mass (%): 19.6,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.796850795,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 181.5,log10 homair (insulin resistance index based on homa): 0.108865574,log10 homais (insulin secretion index based on homa): 1.611014834,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.432375938,insgenin (insulinogenic index): 1.899573413,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.371714203,log10 matsuda insulin sensitivity index: 0.860605238,muscle mass (%): 47,lg10 serum c-reactive protein (mg/l): 0.116939647,lg10 plasma adiponectin (mg/l): 0.963787827,ogtt fasting plasma free fatty acid (mmol/l): 0.38,ogtt 30 min plasma free fatty acid (mmol/l): 0.18,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 9,ogtt 120 min plasma glucose (mmol/l): 5.8,log10 il1 receptor antagonist (pg/ml): 2.498834149,log10 il1 beta (pg/ml): -0.397940009,log10 ogtt fasting plasma insulin (mu/l): 0.69019608,ogtt 30 min plasma insulin (mu/l): 1.661812686,ogtt 120 min plasma insulin (mu/l): 1.572871602,log10 ogtt fasting plasma proinsulin (pm/l): 1.033423755,ogtt 30 min plasma proinsulin (pm/l): 1.376576957,ogtt 120 min plasma proinsulin (pm/l): 1.655138435,log10 bioimpedance: Resistance: 2.717670503,log10 bioimpedance (reactance): 1.770852012,waist to hip ratio: 0.912621359,log10 serum bilirubin (umol/l): 1.591064607,log10 serum alanine aminotransfrase (u/l): 1.322219295,log10 creatinine (umol/l): 1.959041392,log10 total cholesterol (mmol/l): 0.692846919,log10 ldl cholesterol (mmol/l): 0.411619706,log10 hdl cholesterol (mmol/l): 0.209515015,log10 total triglycerides (mmol/l): 0.096910013,log10 serum apoa1 (g/l): 0.161368002,log10 serum apob (g/l): -0.086186148,log10 urinary albumin excretion rate (ug/min): 0.61196828
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098241
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249140,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978068
GSM1098241
GSE45159
0.16865
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1619
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249140
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978068
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.467642944,log10 basal metabolic rate (kcal): 1592,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.526495272,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.905597511,plasma free fatty acids under the curve ogtt (mmol/l * min): 26.85,fat mass (%): 23.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.31458324,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 493.5,log10 homair (insulin resistance index based on homa): 0.159700843,log10 homais (insulin secretion index based on homa): 1.522878745,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.267054336,insgenin (insulinogenic index): 1.365147495,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.173040508,log10 matsuda insulin sensitivity index: 0.829928213,muscle mass (%): 41.7,lg10 serum c-reactive protein (mg/l): 0.815112658,lg10 plasma adiponectin (mg/l): 0.838849091,ogtt fasting plasma free fatty acid (mmol/l): 0.46,ogtt 30 min plasma free fatty acid (mmol/l): 0.31,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 6.5,ogtt 30 min plasma glucose (mmol/l): 10.9,ogtt 120 min plasma glucose (mmol/l): 11.6,log10 il1 receptor antagonist (pg/ml): 2.649305645,log10 il1 beta (pg/ml): -0.602059991,log10 ogtt fasting plasma insulin (mu/l): 0.698970004,ogtt 30 min plasma insulin (mu/l): 1.342422681,ogtt 120 min plasma insulin (mu/l): 1.574031268,log10 ogtt fasting plasma proinsulin (pm/l): 1.037426498,ogtt 30 min plasma proinsulin (pm/l): 1.176091259,ogtt 120 min plasma proinsulin (pm/l): 1.511883361,log10 bioimpedance: Resistance: 2.654176542,log10 bioimpedance (reactance): 1.612783857,waist to hip ratio: 0.944444444,log10 serum bilirubin (umol/l): 1.322219295,log10 serum alanine aminotransfrase (u/l): 1.785329835,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.656098202,log10 ldl cholesterol (mmol/l): 0.376576957,log10 hdl cholesterol (mmol/l): 0.318063335,log10 total triglycerides (mmol/l): -0.167491087,log10 serum apoa1 (g/l): 0.139879086,log10 serum apob (g/l): -0.187086643,log10 urinary albumin excretion rate (ug/min): 0.919731658
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098242
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249141,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978069
GSM1098242
GSE45159
0.060876
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1653
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249141
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978069
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.449085526,log10 basal metabolic rate (kcal): 1501,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.537146159,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.676510436,plasma free fatty acids under the curve ogtt (mmol/l * min): 19.95,fat mass (%): 21.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.909143047,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 253.5,log10 homair (insulin resistance index based on homa): 0.281992354,log10 homais (insulin secretion index based on homa): 1.784141614,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.513310843,insgenin (insulinogenic index): 1.959810464,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.43697321,log10 matsuda insulin sensitivity index: 0.71490465,muscle mass (%): 44.7,lg10 serum c-reactive protein (mg/l): -0.351639989,lg10 plasma adiponectin (mg/l): 0.633468456,ogtt fasting plasma free fatty acid (mmol/l): 0.34,ogtt 30 min plasma free fatty acid (mmol/l): 0.21,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 9,ogtt 120 min plasma glucose (mmol/l): 7.4,log10 il1 receptor antagonist (pg/ml): 2.380807986,log10 il1 beta (pg/ml): -0.585026652,log10 ogtt fasting plasma insulin (mu/l): 0.86332286,ogtt 30 min plasma insulin (mu/l): 1.7355989,ogtt 120 min plasma insulin (mu/l): 1.660865478,log10 ogtt fasting plasma proinsulin (pm/l): 1.240549248,ogtt 30 min plasma proinsulin (pm/l): 1.671172843,ogtt 120 min plasma proinsulin (pm/l): 1.826074803,log10 bioimpedance: Resistance: 2.675778342,log10 bioimpedance (reactance): 1.73239376,waist to hip ratio: 0.92,log10 serum bilirubin (umol/l): 1.041392685,log10 serum alanine aminotransfrase (u/l): 1.491361694,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.688419822,log10 ldl cholesterol (mmol/l): 0.45484486,log10 hdl cholesterol (mmol/l): 0.193124598,log10 total triglycerides (mmol/l): 0.167317335,log10 serum apoa1 (g/l): 0.176091259,log10 serum apob (g/l): -0.004364805,log10 urinary albumin excretion rate (ug/min): 1.032184683
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098243
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249142,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978070
GSM1098243
GSE45159
0.187188
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1666
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249142
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978070
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.307654584,log10 basal metabolic rate (kcal): 1606,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.22740466,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.013549143,plasma free fatty acids under the curve ogtt (mmol/l * min): 12.9,fat mass (%): 13.9,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.346513733,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 39,log10 homair (insulin resistance index based on homa): -0.167491087,log10 homais (insulin secretion index based on homa): 1.574031268,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.332276832,insgenin (insulinogenic index): 2.242307528,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.286276786,log10 matsuda insulin sensitivity index: 1.12951172,muscle mass (%): 69.9,lg10 serum c-reactive protein (mg/l): -0.853871964,lg10 plasma adiponectin (mg/l): 0.681241237,ogtt fasting plasma free fatty acid (mmol/l): 0.21,ogtt 30 min plasma free fatty acid (mmol/l): 0.14,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.1,ogtt 30 min plasma glucose (mmol/l): 6.8,ogtt 120 min plasma glucose (mmol/l): 3.7,log10 il1 receptor antagonist (pg/ml): 2.313360951,log10 il1 beta (pg/ml): 0,log10 ogtt fasting plasma insulin (mu/l): 0.477121255,ogtt 30 min plasma insulin (mu/l): 1.720159303,ogtt 120 min plasma insulin (mu/l): 0.934498451,log10 ogtt fasting plasma proinsulin (pm/l): 1.012837225,ogtt 30 min plasma proinsulin (pm/l): 1.383815366,ogtt 120 min plasma proinsulin (pm/l): 1.356025857,log10 bioimpedance: Resistance: 2.668385917,log10 bioimpedance (reactance): 1.763427994,waist to hip ratio: 0.844444444,log10 serum bilirubin (umol/l): 1,log10 serum alanine aminotransfrase (u/l): 1.380211242,log10 creatinine (umol/l): 1.986771734,log10 total cholesterol (mmol/l): 0.764176132,log10 ldl cholesterol (mmol/l): 0.564666064,log10 hdl cholesterol (mmol/l): 0.23299611,log10 total triglycerides (mmol/l): -0.13667714,log10 serum apoa1 (g/l): 0.089905111,log10 serum apob (g/l): -0.091514981,log10 urinary albumin excretion rate (ug/min): 0.693567076
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098244
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249143,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978071
GSM1098244
GSE45159
0.17386
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1674
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249143
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978071
1
age: 58,tissue: adipose tissue,log10 body mass index: 1.346038272,log10 basal metabolic rate (kcal): 1587,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.475108188,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.414990444,plasma free fatty acids under the curve ogtt (mmol/l * min): 10.2,fat mass (%): 13.5,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.900112063,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 175.5,log10 homair (insulin resistance index based on homa): 0.419772231,log10 homais (insulin secretion index based on homa): 1.782950133,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.505692507,insgenin (insulinogenic index): 1.938365747,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.406335758,log10 matsuda insulin sensitivity index: 0.657467551,muscle mass (%): 48.8,lg10 serum c-reactive protein (mg/l): -0.503070352,lg10 plasma adiponectin (mg/l): 0.832508913,ogtt fasting plasma free fatty acid (mmol/l): 0.23,ogtt 30 min plasma free fatty acid (mmol/l): 0.09,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 6.5,ogtt 30 min plasma glucose (mmol/l): 10.4,ogtt 120 min plasma glucose (mmol/l): 5.2,log10 il1 receptor antagonist (pg/ml): 2.158995377,log10 il1 beta (pg/ml): -0.537602002,log10 ogtt fasting plasma insulin (mu/l): 0.959041392,ogtt 30 min plasma insulin (mu/l): 1.8162413,ogtt 120 min plasma insulin (mu/l): 1.451786436,log10 ogtt fasting plasma proinsulin (pm/l): 1.451786436,ogtt 30 min plasma proinsulin (pm/l): 1.596597096,ogtt 120 min plasma proinsulin (pm/l): 1.638489257,log10 bioimpedance: Resistance: 2.679427897,log10 bioimpedance (reactance): 1.672097858,waist to hip ratio: 0.894736842,log10 serum bilirubin (umol/l): 0.84509804,log10 serum alanine aminotransfrase (u/l): 1.230448921,log10 creatinine (umol/l): 1.913813852,log10 total cholesterol (mmol/l): 0.835690571,log10 ldl cholesterol (mmol/l): 0.669316881,log10 hdl cholesterol (mmol/l): 0.198657087,log10 total triglycerides (mmol/l): 0.093421685,log10 serum apoa1 (g/l): 0.187520721,log10 serum apob (g/l): 0.10720997,log10 urinary albumin excretion rate (ug/min): 1.180890142
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098245
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249144,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978072
GSM1098245
GSE45159
0.13427
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1678
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249144
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978072
1
age: 58,tissue: adipose tissue,log10 body mass index: 1.321147221,log10 basal metabolic rate (kcal): 1595,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.66866566,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.508170937,plasma free fatty acids under the curve ogtt (mmol/l * min): 21.75,fat mass (%): 16.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.512740463,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 58.5,log10 homair (insulin resistance index based on homa): -0.189021143,log10 homais (insulin secretion index based on homa): 1.393784049,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.068185862,insgenin (insulinogenic index): 1.892094603,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.992465148,log10 matsuda insulin sensitivity index: 1.229341053,muscle mass (%): 49.6,lg10 serum c-reactive protein (mg/l): -0.853871964,lg10 plasma adiponectin (mg/l): 1.133538908,ogtt fasting plasma free fatty acid (mmol/l): 0.52,ogtt 30 min plasma free fatty acid (mmol/l): 0.21,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 7.1,ogtt 120 min plasma glucose (mmol/l): 4.9,log10 il1 receptor antagonist (pg/ml): 2.056790548,log10 il1 beta (pg/ml): -0.013228266,log10 ogtt fasting plasma insulin (mu/l): 0.414973348,ogtt 30 min plasma insulin (mu/l): 1.344392274,ogtt 120 min plasma insulin (mu/l): 1.113943352,log10 ogtt fasting plasma proinsulin (pm/l): 0.69019608,ogtt 30 min plasma proinsulin (pm/l): 1.130333768,ogtt 120 min plasma proinsulin (pm/l): 1.414973348,log10 bioimpedance: Resistance: 2.779596491,log10 bioimpedance (reactance): 1.875061263,waist to hip ratio: 0.867346939,log10 serum bilirubin (umol/l): 1.146128036,log10 serum alanine aminotransfrase (u/l): 1.176091259,log10 creatinine (umol/l): 1.86332286,log10 total cholesterol (mmol/l): 0.762678564,log10 ldl cholesterol (mmol/l): 0.445604203,log10 hdl cholesterol (mmol/l): 0.369215857,log10 total triglycerides (mmol/l): -0.15490196,log10 serum apoa1 (g/l): 0.235528447,log10 serum apob (g/l): -0.096910013,log10 urinary albumin excretion rate (ug/min): 1.248323645
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098246
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249145,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978073
GSM1098246
GSE45159
0.176752
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1702
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249145
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978073
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.512530195,log10 basal metabolic rate (kcal): 1749,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.723610629,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.392910911,plasma free fatty acids under the curve ogtt (mmol/l * min): 23.7,fat mass (%): 27.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.719388821,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 171,log10 homair (insulin resistance index based on homa): 0.299095496,log10 homais (insulin secretion index based on homa): 1.881900688,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.556736578,insgenin (insulinogenic index): 2.096107498,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.481098497,log10 matsuda insulin sensitivity index: 0.734518713,muscle mass (%): 40.2,lg10 serum c-reactive protein (mg/l): 0.435685138,lg10 plasma adiponectin (mg/l): 0.531478917,ogtt fasting plasma free fatty acid (mmol/l): 0.39,ogtt 30 min plasma free fatty acid (mmol/l): 0.26,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 9.5,ogtt 120 min plasma glucose (mmol/l): 4.2,log10 il1 receptor antagonist (pg/ml): 2.627550188,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.903089987,ogtt 30 min plasma insulin (mu/l): 1.949877704,ogtt 120 min plasma insulin (mu/l): 1.079181246,log10 ogtt fasting plasma proinsulin (pm/l): 1.096910013,ogtt 30 min plasma proinsulin (pm/l): 1.606381365,ogtt 120 min plasma proinsulin (pm/l): 1.612783857,log10 bioimpedance: Resistance: 2.650307523,log10 bioimpedance (reactance): 1.633468456,waist to hip ratio: 1.018181818,log10 serum bilirubin (umol/l): 1.301029996,log10 serum alanine aminotransfrase (u/l): 1.681241237,log10 creatinine (umol/l): 1.857332496,log10 total cholesterol (mmol/l): 0.741939078,log10 ldl cholesterol (mmol/l): 0.51851394,log10 hdl cholesterol (mmol/l): 0.004321374,log10 total triglycerides (mmol/l): 0.338456494,log10 serum apoa1 (g/l): 0.096910013,log10 serum apob (g/l): 0.08278537,log10 urinary albumin excretion rate (ug/min): 0.857162551
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098247
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249146,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978074
GSM1098247
GSE45159
0.175511
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1722
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249146
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978074
1
age: 59,tissue: adipose tissue,log10 body mass index: 1.428578921,log10 basal metabolic rate (kcal): 1734,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.260527536,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.5061315,plasma free fatty acids under the curve ogtt (mmol/l * min): 10.8,fat mass (%): 18.3,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.791976821,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 190.5,log10 homair (insulin resistance index based on homa): 0.230789411,log10 homais (insulin secretion index based on homa): 1.758846095,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.498186452,insgenin (insulinogenic index): 2.051308997,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.427145161,log10 matsuda insulin sensitivity index: 0.785091732,muscle mass (%): 47,lg10 serum c-reactive protein (mg/l): 0.380211242,lg10 plasma adiponectin (mg/l): 0.770852012,ogtt fasting plasma free fatty acid (mmol/l): 0.15,ogtt 30 min plasma free fatty acid (mmol/l): 0.12,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.8,ogtt 30 min plasma glucose (mmol/l): 9.5,ogtt 120 min plasma glucose (mmol/l): 5.1,log10 il1 receptor antagonist (pg/ml): 2.429929836,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.819543936,ogtt 30 min plasma insulin (mu/l): 1.880813592,ogtt 120 min plasma insulin (mu/l): 1.117271296,log10 ogtt fasting plasma proinsulin (pm/l): 0.857332496,ogtt 30 min plasma proinsulin (pm/l): 1.338456494,ogtt 120 min plasma proinsulin (pm/l): 1.260071388,log10 bioimpedance: Resistance: 2.617000341,log10 bioimpedance (reactance): 1.643452676,waist to hip ratio: 0.931372549,log10 serum bilirubin (umol/l): 1.146128036,log10 serum alanine aminotransfrase (u/l): 1.505149978,log10 creatinine (umol/l): 1.968482949,log10 total cholesterol (mmol/l): 0.758911892,log10 ldl cholesterol (mmol/l): 0.56937391,log10 hdl cholesterol (mmol/l): 0.217483944,log10 total triglycerides (mmol/l): 0.164352856,log10 serum apoa1 (g/l): 0.10720997,log10 serum apob (g/l): -0.026872146,log10 urinary albumin excretion rate (ug/min): 1.05270635
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098248
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249147,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978075
GSM1098248
GSE45159
0.204179
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1725
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249147
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978075
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.520671054,log10 basal metabolic rate (kcal): 1381,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 7.120312714,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.3156401,plasma free fatty acids under the curve ogtt (mmol/l * min): 24,fat mass (%): 26,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.25915477,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 397.5,log10 homair (insulin resistance index based on homa): 0.808885867,log10 homais (insulin secretion index based on homa): 2.091770373,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.988630293,insgenin (insulinogenic index): 1.875728895,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.915378727,log10 matsuda insulin sensitivity index: 0.149691359,muscle mass (%): 40,lg10 serum c-reactive protein (mg/l): -0.076238039,lg10 plasma adiponectin (mg/l): 0.77815125,ogtt fasting plasma free fatty acid (mmol/l): 0.49,ogtt 30 min plasma free fatty acid (mmol/l): 0.24,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 6.9,ogtt 30 min plasma glucose (mmol/l): 12.1,ogtt 120 min plasma glucose (mmol/l): 8.8,log10 il1 receptor antagonist (pg/ml): 2.501251022,log10 il1 beta (pg/ml): 0.697229343,log10 ogtt fasting plasma insulin (mu/l): 1.322219295,ogtt 30 min plasma insulin (mu/l): 1.935003151,ogtt 120 min plasma insulin (mu/l): 2.378397901,log10 ogtt fasting plasma proinsulin (pm/l): 1.503790683,ogtt 30 min plasma proinsulin (pm/l): 1.720159303,ogtt 120 min plasma proinsulin (pm/l): 1.984077034,log10 bioimpedance: Resistance: 2.668385917,log10 bioimpedance (reactance): 1.672097858,waist to hip ratio: 1.113861386,log10 serum bilirubin (umol/l): 1,log10 serum alanine aminotransfrase (u/l): 1.832508913,log10 creatinine (umol/l): 1.748188027,log10 total cholesterol (mmol/l): 0.906873535,log10 ldl cholesterol (mmol/l): 0.72916479,log10 hdl cholesterol (mmol/l): 0.204119983,log10 total triglycerides (mmol/l): 0.269512944,log10 serum apoa1 (g/l): 0.209515015,log10 serum apob (g/l): 0.193124598,log10 urinary albumin excretion rate (ug/min): 0.798521898
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098249
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249148,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978076
GSM1098249
GSE45159
0.098109
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1757
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249148
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978076
1
age: 59,tissue: adipose tissue,log10 body mass index: 1.412912526,log10 basal metabolic rate (kcal): 1613,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.686623353,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.732760914,plasma free fatty acids under the curve ogtt (mmol/l * min): 11.85,fat mass (%): 16.7,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.677719642,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 75,log10 homair (insulin resistance index based on homa): 0.147779347,log10 homais (insulin secretion index based on homa): 1.577236408,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.312938609,insgenin (insulinogenic index): 1.894177554,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.227475343,log10 matsuda insulin sensitivity index: 0.927466773,muscle mass (%): 47.3,lg10 serum c-reactive protein (mg/l): -0.450996738,lg10 plasma adiponectin (mg/l): 0.740362689,ogtt fasting plasma free fatty acid (mmol/l): 0.22,ogtt 30 min plasma free fatty acid (mmol/l): 0.12,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 6.2,ogtt 30 min plasma glucose (mmol/l): 9.4,ogtt 120 min plasma glucose (mmol/l): 3.6,log10 il1 receptor antagonist (pg/ml): 2.17912073,log10 il1 beta (pg/ml): -0.795880017,log10 ogtt fasting plasma insulin (mu/l): 0.707570176,ogtt 30 min plasma insulin (mu/l): 1.671172843,ogtt 120 min plasma insulin (mu/l): 1.075546961,log10 ogtt fasting plasma proinsulin (pm/l): 1.103803721,ogtt 30 min plasma proinsulin (pm/l): 1.537819095,ogtt 120 min plasma proinsulin (pm/l): 1.574031268,log10 bioimpedance: Resistance: 2.62838893,log10 bioimpedance (reactance): 1.643452676,waist to hip ratio: 0.865,log10 serum bilirubin (umol/l): 1,log10 serum alanine aminotransfrase (u/l): 1.301029996,log10 creatinine (umol/l): 1.857332496,log10 total cholesterol (mmol/l): 0.564666064,log10 ldl cholesterol (mmol/l): 0.326335861,log10 hdl cholesterol (mmol/l): -0.004364805,log10 total triglycerides (mmol/l): 0.214843848,log10 serum apoa1 (g/l): 0.089905111,log10 serum apob (g/l): -0.065501549,log10 urinary albumin excretion rate (ug/min): 0.810180579
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098250
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249149,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978077
GSM1098250
GSE45159
0.218901
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1764
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249149
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978077
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.44406498,log10 basal metabolic rate (kcal): 1451,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.68421978,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.801017968,plasma free fatty acids under the curve ogtt (mmol/l * min): 42.15,fat mass (%): 22.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.868050854,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 262.5,log10 homair (insulin resistance index based on homa): 0.094975513,log10 homais (insulin secretion index based on homa): 1.677780705,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.189518386,insgenin (insulinogenic index): 1.596920352,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.074487305,log10 matsuda insulin sensitivity index: 0.976146889,muscle mass (%): 43.1,lg10 serum c-reactive protein (mg/l): -0.003488328,lg10 plasma adiponectin (mg/l): 0.826074803,ogtt fasting plasma free fatty acid (mmol/l): 0.57,ogtt 30 min plasma free fatty acid (mmol/l): 0.5,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 9,ogtt 120 min plasma glucose (mmol/l): 6.9,log10 il1 receptor antagonist (pg/ml): 2.48207283,log10 il1 beta (pg/ml): -0.420216403,log10 ogtt fasting plasma insulin (mu/l): 0.698970004,ogtt 30 min plasma insulin (mu/l): 1.437750563,ogtt 120 min plasma insulin (mu/l): 1.281033367,log10 ogtt fasting plasma proinsulin (pm/l): 1.149219113,ogtt 30 min plasma proinsulin (pm/l): 1.440909082,ogtt 120 min plasma proinsulin (pm/l): 1.650307523,log10 bioimpedance: Resistance: 2.698100546,log10 bioimpedance (reactance): 1.716003344,waist to hip ratio: 0.95959596,log10 serum bilirubin (umol/l): 1.301029996,log10 serum alanine aminotransfrase (u/l): 1.579783597,log10 creatinine (umol/l): 1.86332286,log10 total cholesterol (mmol/l): 0.887054378,log10 ldl cholesterol (mmol/l): 0.7355989,log10 hdl cholesterol (mmol/l): 0.326335861,log10 total triglycerides (mmol/l): 0.064457989,log10 serum apoa1 (g/l): 0.285557309,log10 serum apob (g/l): 0.164352856,log10 urinary albumin excretion rate (ug/min): 1.313650938
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098251
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249150,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978078
GSM1098251
GSE45159
0.039329
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1768
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249150
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978078
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.388478806,log10 basal metabolic rate (kcal): 1431,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.600779512,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.521616776,plasma free fatty acids under the curve ogtt (mmol/l * min): 25.35,fat mass (%): 18.5,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.811374694,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 238.5,log10 homair (insulin resistance index based on homa): 0.112456041,log10 homais (insulin secretion index based on homa): 1.72427587,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.335898992,insgenin (insulinogenic index): 1.453755036,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.251784177,log10 matsuda insulin sensitivity index: 0.896028713,muscle mass (%): 46.3,lg10 serum c-reactive protein (mg/l): 0.602059991,lg10 plasma adiponectin (mg/l): 0.986771734,ogtt fasting plasma free fatty acid (mmol/l): 0.45,ogtt 30 min plasma free fatty acid (mmol/l): 0.28,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.5,ogtt 30 min plasma glucose (mmol/l): 9.7,ogtt 120 min plasma glucose (mmol/l): 5.2,log10 il1 receptor antagonist (pg/ml): 2.282576801,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.72427587,ogtt 30 min plasma insulin (mu/l): 1.401400541,ogtt 120 min plasma insulin (mu/l): 1.652246341,log10 ogtt fasting plasma proinsulin (pm/l): 1.127104798,ogtt 30 min plasma proinsulin (pm/l): 1.328379603,ogtt 120 min plasma proinsulin (pm/l): 1.789580712,log10 bioimpedance: Resistance: 2.728353782,log10 bioimpedance (reactance): 1.785329835,waist to hip ratio: 0.947368421,log10 serum bilirubin (umol/l): 1.176091259,log10 serum alanine aminotransfrase (u/l): 1.414973348,log10 creatinine (umol/l): 1.886490725,log10 total cholesterol (mmol/l): 0.728353782,log10 ldl cholesterol (mmol/l): 0.445604203,log10 hdl cholesterol (mmol/l): 0.359835482,log10 total triglycerides (mmol/l): -0.017728767,log10 serum apoa1 (g/l): 0.267171728,log10 serum apob (g/l): -0.075720714,log10 urinary albumin excretion rate (ug/min): 0.949694882
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098252
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249151,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978079
GSM1098252
GSE45159
0.276419
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1777
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249151
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978079
1
age: 54,tissue: adipose tissue,log10 body mass index: 1.422408922,log10 basal metabolic rate (kcal): 1566,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.666941996,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.852710764,plasma free fatty acids under the curve ogtt (mmol/l * min): 21,fat mass (%): 22.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.73978061,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 147,log10 homair (insulin resistance index based on homa): 0.174544349,log10 homais (insulin secretion index based on homa): 1.67669361,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.627632088,insgenin (insulinogenic index): 1.882443217,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.583448167,log10 matsuda insulin sensitivity index: 0.730670091,muscle mass (%): 44.1,lg10 serum c-reactive protein (mg/l): -0.175223538,lg10 plasma adiponectin (mg/l): 0.633468456,ogtt fasting plasma free fatty acid (mmol/l): 0.3,ogtt 30 min plasma free fatty acid (mmol/l): 0.26,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 9.4,ogtt 120 min plasma glucose (mmol/l): 4.5,log10 il1 receptor antagonist (pg/ml): 2.199837422,log10 il1 beta (pg/ml): -0.408935393,log10 ogtt fasting plasma insulin (mu/l): 0.755874856,ogtt 30 min plasma insulin (mu/l): 1.700703717,ogtt 120 min plasma insulin (mu/l): 1.945960704,log10 ogtt fasting plasma proinsulin (pm/l): 1.1430148,ogtt 30 min plasma proinsulin (pm/l): 1.387389826,ogtt 120 min plasma proinsulin (pm/l): 1.790988475,log10 bioimpedance: Resistance: 2.720985744,log10 bioimpedance (reactance): 1.763427994,waist to hip ratio: 0.953846154,log10 serum bilirubin (umol/l): 1.113943352,log10 serum alanine aminotransfrase (u/l): 1.477121255,log10 creatinine (umol/l): 1.86923172,log10 total cholesterol (mmol/l): 0.802773725,log10 ldl cholesterol (mmol/l): 0.6599162,log10 hdl cholesterol (mmol/l): 0.139879086,log10 total triglycerides (mmol/l): 0.093421685,log10 serum apoa1 (g/l): 0.181843588,log10 serum apob (g/l): 0.120573931,log10 urinary albumin excretion rate (ug/min): 1.111365929
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098253
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249152,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978080
GSM1098253
GSE45159
0.294805
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1784
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249152
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978080
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.489355711,log10 basal metabolic rate (kcal): 1738,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.542253687,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.929550479,plasma free fatty acids under the curve ogtt (mmol/l * min): 36.9,fat mass (%): 22.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.737247343,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 157.5,log10 homair (insulin resistance index based on homa): 0.355728148,log10 homais (insulin secretion index based on homa): 1.883784832,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.835969445,insgenin (insulinogenic index): 2.507855872,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.793846239,log10 matsuda insulin sensitivity index: 0.557825389,muscle mass (%): 44.8,lg10 serum c-reactive protein (mg/l): 0.331427297,lg10 plasma adiponectin (mg/l): 0.69019608,ogtt fasting plasma free fatty acid (mmol/l): 0.65,ogtt 30 min plasma free fatty acid (mmol/l): 0.4,ogtt 120 min plasma free fatty acid (mmol/l): 0.07,ogtt fasting plasma glucose (mmol/l): 5.8,ogtt 30 min plasma glucose (mmol/l): 8.5,ogtt 120 min plasma glucose (mmol/l): 5.7,log10 il1 receptor antagonist (pg/ml): 2.421274791,log10 il1 beta (pg/ml): -0.602059991,log10 ogtt fasting plasma insulin (mu/l): 0.944482672,ogtt 30 min plasma insulin (mu/l): 2.186673867,ogtt 120 min plasma insulin (mu/l): 1.662757832,log10 ogtt fasting plasma proinsulin (pm/l): 1.361727836,ogtt 30 min plasma proinsulin (pm/l): 1.851258349,ogtt 120 min plasma proinsulin (pm/l): 1.973127854,log10 bioimpedance: Resistance: 2.597695186,log10 bioimpedance (reactance): 1.643452676,waist to hip ratio: 0.981308411,log10 serum bilirubin (umol/l): 1.602059991,log10 serum alanine aminotransfrase (u/l): 1.763427994,log10 creatinine (umol/l): 1.897627091,log10 total cholesterol (mmol/l): 0.686636269,log10 ldl cholesterol (mmol/l): 0.509202522,log10 hdl cholesterol (mmol/l): -0.036212173,log10 total triglycerides (mmol/l): 0.352182518,log10 serum apoa1 (g/l): 0.056904851,log10 serum apob (g/l): 0.071882007,log10 urinary albumin excretion rate (ug/min): 1.021189299
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098254
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249153,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978081
GSM1098254
GSE45159
0.121957
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1840
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249153
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978081
1
age: 60,tissue: adipose tissue,log10 body mass index: 1.542755717,log10 basal metabolic rate (kcal): 1465,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.385885025,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.76156832,plasma free fatty acids under the curve ogtt (mmol/l * min): 55.05,fat mass (%): 24.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.44604941,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 531,log10 homair (insulin resistance index based on homa): 0.660486016,log10 homais (insulin secretion index based on homa): 1.888164309,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.60260252,insgenin (insulinogenic index): 1.646263654,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.473545359,log10 matsuda insulin sensitivity index: 0.398433097,muscle mass (%): 40.2,lg10 serum c-reactive protein (mg/l): 0.08350262,lg10 plasma adiponectin (mg/l): 1.041392685,ogtt fasting plasma free fatty acid (mmol/l): 0.89,ogtt 30 min plasma free fatty acid (mmol/l): 0.62,ogtt 120 min plasma free fatty acid (mmol/l): 0.1,ogtt fasting plasma glucose (mmol/l): 7.2,ogtt 30 min plasma glucose (mmol/l): 13.5,ogtt 120 min plasma glucose (mmol/l): 10.6,log10 il1 receptor antagonist (pg/ml): 2.280601132,log10 il1 beta (pg/ml): 0.322219295,log10 ogtt fasting plasma insulin (mu/l): 1.155336037,ogtt 30 min plasma insulin (mu/l): 1.783903579,ogtt 120 min plasma insulin (mu/l): 1.795880017,log10 ogtt fasting plasma proinsulin (pm/l): 1.409933123,ogtt 30 min plasma proinsulin (pm/l): 1.627365857,ogtt 120 min plasma proinsulin (pm/l): 1.879669206,log10 bioimpedance: Resistance: 2.571708832,log10 bioimpedance (reactance): 1.505149978,waist to hip ratio: 1.072727273,log10 serum bilirubin (umol/l): 1.230448921,log10 serum alanine aminotransfrase (u/l): 1.322219295,log10 creatinine (umol/l): 1.934498451,log10 total cholesterol (mmol/l): 0.853698212,log10 ldl cholesterol (mmol/l): 0.667452953,log10 hdl cholesterol (mmol/l): 0.100370545,log10 total triglycerides (mmol/l): 0.72916479,log10 serum apoa1 (g/l): 0.240549248,log10 serum apob (g/l): 0.301029996,log10 urinary albumin excretion rate (ug/min): 0.605614946
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098255
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249154,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978082
GSM1098255
GSE45159
0.054484
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1841
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249154
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978082
1
age: 59,tissue: adipose tissue,log10 body mass index: 1.319337614,log10 basal metabolic rate (kcal): 1470,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.511211783,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.244708617,plasma free fatty acids under the curve ogtt (mmol/l * min): 10.2,fat mass (%): 12.1,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.579315938,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 57,log10 homair (insulin resistance index based on homa): -0.201119265,log10 homais (insulin secretion index based on homa): 1.301029996,log10 insulin area under the curve (ogtt) (pmol/l * min): 3.976670881,insgenin (insulinogenic index): 1.93815569,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.889245661,log10 matsuda insulin sensitivity index: 1.268699938,muscle mass (%): 49.5,lg10 serum c-reactive protein (mg/l): -0.410050399,lg10 plasma adiponectin (mg/l): 0.875061263,ogtt fasting plasma free fatty acid (mmol/l): 0.27,ogtt 30 min plasma free fatty acid (mmol/l): 0.08,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 7,ogtt 120 min plasma glucose (mmol/l): 5.7,log10 il1 receptor antagonist (pg/ml): 2.240748878,log10 il1 beta (pg/ml): -0.48148606,log10 ogtt fasting plasma insulin (mu/l): 0.380211242,ogtt 30 min plasma insulin (mu/l): 1.26245109,ogtt 120 min plasma insulin (mu/l): 0.995635195,log10 ogtt fasting plasma proinsulin (pm/l): 0.944482672,ogtt 30 min plasma proinsulin (pm/l): 1.206825876,ogtt 120 min plasma proinsulin (pm/l): 1.264817823,log10 bioimpedance: Resistance: 2.720159303,log10 bioimpedance (reactance): 1.73239376,waist to hip ratio: 0.827956989,log10 serum bilirubin (umol/l): 1.230448921,log10 serum alanine aminotransfrase (u/l): 1,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.725911632,log10 ldl cholesterol (mmol/l): 0.481442629,log10 hdl cholesterol (mmol/l): 0.278753601,log10 total triglycerides (mmol/l): 0.004321374,log10 serum apoa1 (g/l): 0.252853031,log10 serum apob (g/l): -0.031517051,log10 urinary albumin excretion rate (ug/min): 0.801914447
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098256
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249155,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978083
GSM1098256
GSE45159
0.217262
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1885
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249155
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978083
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.409849348,log10 basal metabolic rate (kcal): 1657,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.521311649,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.680077586,plasma free fatty acids under the curve ogtt (mmol/l * min): 30.45,fat mass (%): 20.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.00281502,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 342,log10 homair (insulin resistance index based on homa): 0.29578694,log10 homais (insulin secretion index based on homa): 1.850701918,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.776410589,insgenin (insulinogenic index): 2.137451498,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.733550337,log10 matsuda insulin sensitivity index: 0.569131479,muscle mass (%): 46.3,lg10 serum c-reactive protein (mg/l): -0.244125144,lg10 plasma adiponectin (mg/l): 0.51851394,ogtt fasting plasma free fatty acid (mmol/l): 0.51,ogtt 30 min plasma free fatty acid (mmol/l): 0.32,ogtt 120 min plasma free fatty acid (mmol/l): 0.08,ogtt fasting plasma glucose (mmol/l): 5.7,ogtt 30 min plasma glucose (mmol/l): 9.6,ogtt 120 min plasma glucose (mmol/l): 8.1,log10 il1 receptor antagonist (pg/ml): 2.287465805,log10 il1 beta (pg/ml): 0.029383778,log10 ogtt fasting plasma insulin (mu/l): 0.892094603,ogtt 30 min plasma insulin (mu/l): 1.986771734,ogtt 120 min plasma insulin (mu/l): 1.951337519,log10 ogtt fasting plasma proinsulin (pm/l): 1.113943352,ogtt 30 min plasma proinsulin (pm/l): 1.739572344,ogtt 120 min plasma proinsulin (pm/l): 1.960946196,log10 bioimpedance: Resistance: 2.694605199,log10 bioimpedance (reactance): 1.755874856,waist to hip ratio: 0.95049505,log10 serum bilirubin (umol/l): 1.397940009,log10 serum alanine aminotransfrase (u/l): 1.51851394,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.775974331,log10 ldl cholesterol (mmol/l): 0.562292864,log10 hdl cholesterol (mmol/l): 0.133538908,log10 total triglycerides (mmol/l): 0.586587305,log10 serum apoa1 (g/l): 0.240549248,log10 serum apob (g/l): 0.139879086,log10 urinary albumin excretion rate (ug/min): 0.615978253
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098257
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249156,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978084
GSM1098257
GSE45159
0.218739
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1891
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249156
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978084
1
age: 52,tissue: adipose tissue,log10 body mass index: 1.566823019,log10 basal metabolic rate (kcal): 1768,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.548199059,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 7.358365399,plasma free fatty acids under the curve ogtt (mmol/l * min): 30.9,fat mass (%): 28.6,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.931476234,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 184.5,log10 homair (insulin resistance index based on homa): 0.480198642,log10 homais (insulin secretion index based on homa): 1.822505527,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.325495139,insgenin (insulinogenic index): 1.752048448,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.138081546,log10 matsuda insulin sensitivity index: 0.703550682,muscle mass (%): 39.4,lg10 serum c-reactive protein (mg/l): 0.849910558,lg10 plasma adiponectin (mg/l): 0.591064607,ogtt fasting plasma free fatty acid (mmol/l): 0.42,ogtt 30 min plasma free fatty acid (mmol/l): 0.38,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 6.6,ogtt 30 min plasma glucose (mmol/l): 10.2,ogtt 120 min plasma glucose (mmol/l): 5.9,log10 il1 receptor antagonist (pg/ml): 2.433609843,log10 il1 beta (pg/ml): -0.30980392,log10 ogtt fasting plasma insulin (mu/l): 1.012837225,ogtt 30 min plasma insulin (mu/l): 1.645422269,ogtt 120 min plasma insulin (mu/l): 1.204119983,log10 ogtt fasting plasma proinsulin (pm/l): 1,ogtt 30 min plasma proinsulin (pm/l): 1.33243846,ogtt 120 min plasma proinsulin (pm/l): 1.53529412,log10 bioimpedance: Resistance: 2.599883072,log10 bioimpedance (reactance): 1.591064607,waist to hip ratio: 1.043103448,log10 serum bilirubin (umol/l): 1.505149978,log10 serum alanine aminotransfrase (u/l): 1.531478917,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.733999287,log10 ldl cholesterol (mmol/l): 0.597695186,log10 hdl cholesterol (mmol/l): -0.050609993,log10 total triglycerides (mmol/l): 0.28780173,log10 serum apoa1 (g/l): 0.079181246,log10 serum apob (g/l): 0.10720997,log10 urinary albumin excretion rate (ug/min): 1.168082612
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098258
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249157,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978085
GSM1098258
GSE45159
0.156619
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1892
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249157
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978085
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.406315415,log10 basal metabolic rate (kcal): 1643,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.425563154,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.573596607,plasma free fatty acids under the curve ogtt (mmol/l * min): 18,fat mass (%): 18.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.957827561,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 322.5,log10 homair (insulin resistance index based on homa): 0.492915522,log10 homais (insulin secretion index based on homa): 2.075720714,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.628184508,insgenin (insulinogenic index): 2.004216484,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.524785449,log10 matsuda insulin sensitivity index: 0.551520623,muscle mass (%): 47,lg10 serum c-reactive protein (mg/l): -0.467245621,lg10 plasma adiponectin (mg/l): 0.146128036,ogtt fasting plasma free fatty acid (mmol/l): 0.32,ogtt 30 min plasma free fatty acid (mmol/l): 0.19,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.6,ogtt 30 min plasma glucose (mmol/l): 9.7,ogtt 120 min plasma glucose (mmol/l): 7.3,log10 il1 receptor antagonist (pg/ml): 2.205204364,log10 il1 beta (pg/ml): -0.244125144,log10 ogtt fasting plasma insulin (mu/l): 1.096910013,ogtt 30 min plasma insulin (mu/l): 1.911157609,ogtt 120 min plasma insulin (mu/l): 1.648360011,log10 ogtt fasting plasma proinsulin (pm/l): 0.919078092,ogtt 30 min plasma proinsulin (pm/l): 1.133538908,ogtt 120 min plasma proinsulin (pm/l): 1.392696953,log10 bioimpedance: Resistance: 2.662757832,log10 bioimpedance (reactance): 1.698970004,waist to hip ratio: 0.927835052,log10 serum bilirubin (umol/l): 1.113943352,log10 serum alanine aminotransfrase (u/l): 1.322219295,log10 creatinine (umol/l): 1.929418926,log10 total cholesterol (mmol/l): 0.654176542,log10 ldl cholesterol (mmol/l): 0.409933123,log10 hdl cholesterol (mmol/l): 0.056904851,log10 total triglycerides (mmol/l): 0.198657087,log10 serum apoa1 (g/l): 0.100370545,log10 serum apob (g/l): -0.050609993,log10 urinary albumin excretion rate (ug/min): 0.931691785
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098259
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249158,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978086
GSM1098259
GSE45159
0.641122
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1896
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249158
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978086
1
age: 56,tissue: adipose tissue,log10 body mass index: 1.43964122,log10 basal metabolic rate (kcal): 1592,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.46520479,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.635255458,plasma free fatty acids under the curve ogtt (mmol/l * min): 16.5,fat mass (%): 18.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.09671515,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 207,log10 homair (insulin resistance index based on homa): 0.336593137,log10 homais (insulin secretion index based on homa): 1.529509324,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.119783586,insgenin (insulinogenic index): 1.017911589,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.925518358,log10 matsuda insulin sensitivity index: 0.840102059,muscle mass (%): 46.1,lg10 serum c-reactive protein (mg/l): 0.079904468,lg10 plasma adiponectin (mg/l): 1.012837225,ogtt fasting plasma free fatty acid (mmol/l): 0.29,ogtt 30 min plasma free fatty acid (mmol/l): 0.18,ogtt 120 min plasma free fatty acid (mmol/l): 0.03,ogtt fasting plasma glucose (mmol/l): 7.4,ogtt 30 min plasma glucose (mmol/l): 13.1,ogtt 120 min plasma glucose (mmol/l): 4.4,log10 il1 receptor antagonist (pg/ml): 2.210425542,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.819543936,ogtt 30 min plasma insulin (mu/l): 1.217483944,ogtt 120 min plasma insulin (mu/l): 1.390935107,log10 ogtt fasting plasma proinsulin (pm/l): 1.245512668,ogtt 30 min plasma proinsulin (pm/l): 1.397940009,ogtt 120 min plasma proinsulin (pm/l): 1.813580989,log10 bioimpedance: Resistance: 2.613841822,log10 bioimpedance (reactance): 1.62324929,waist to hip ratio: 0.96,log10 serum bilirubin (umol/l): 1.176091259,log10 serum alanine aminotransfrase (u/l): 1.230448921,log10 creatinine (umol/l): 1.919078092,log10 total cholesterol (mmol/l): 0.762678564,log10 ldl cholesterol (mmol/l): 0.557507202,log10 hdl cholesterol (mmol/l): 0.176091259,log10 total triglycerides (mmol/l): 0.301029996,log10 serum apoa1 (g/l): 0.217483944,log10 serum apob (g/l): 0.086359831,log10 urinary albumin excretion rate (ug/min): 0.800105826
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098260
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249159,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978087
GSM1098260
GSE45159
0.268976
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1915
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249159
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978087
1
age: 55,tissue: adipose tissue,log10 body mass index: 1.350977817,log10 basal metabolic rate (kcal): 1753,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.75408472,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.916007293,plasma free fatty acids under the curve ogtt (mmol/l * min): 17.4,fat mass (%): 17.4,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.927037169,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 205.5,log10 homair (insulin resistance index based on homa): -0.027488604,log10 homais (insulin secretion index based on homa): 1.357145938,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.010002974,insgenin (insulinogenic index): 1.474997561,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.895256753,log10 matsuda insulin sensitivity index: 1.130449848,muscle mass (%): 47,lg10 serum c-reactive protein (mg/l): -0.077793723,lg10 plasma adiponectin (mg/l): 1.056904851,ogtt fasting plasma free fatty acid (mmol/l): 0.29,ogtt 30 min plasma free fatty acid (mmol/l): 0.18,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 6.4,ogtt 30 min plasma glucose (mmol/l): 10.5,ogtt 120 min plasma glucose (mmol/l): 5.5,log10 il1 receptor antagonist (pg/ml): 1.932270776,log10 il1 beta (pg/ml): -0.853871964,log10 ogtt fasting plasma insulin (mu/l): 0.51851394,ogtt 30 min plasma insulin (mu/l): 1.374748346,ogtt 120 min plasma insulin (mu/l): 0.716003344,log10 ogtt fasting plasma proinsulin (pm/l): 0.826074803,ogtt 30 min plasma proinsulin (pm/l): 1.017033339,ogtt 120 min plasma proinsulin (pm/l): 1.021189299,log10 bioimpedance: Resistance: 2.715167358,log10 bioimpedance (reactance): 1.716003344,waist to hip ratio: 0.94,log10 serum bilirubin (umol/l): 1.113943352,log10 serum alanine aminotransfrase (u/l): 1.602059991,log10 creatinine (umol/l): 1.84509804,log10 total cholesterol (mmol/l): 0.688419822,log10 ldl cholesterol (mmol/l): 0.392696953,log10 hdl cholesterol (mmol/l): 0.367355921,log10 total triglycerides (mmol/l): -0.26760624,log10 serum apoa1 (g/l): 0.252853031,log10 serum apob (g/l): -0.13667714,log10 urinary albumin excretion rate (ug/min): 1.187385025
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098261
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249160,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978088
GSM1098261
GSE45159
0.177057
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1921
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249160
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978088
1
age: 61,tissue: adipose tissue,log10 body mass index: 1.408707506,log10 basal metabolic rate (kcal): 1411,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.74771644,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.710474183,plasma free fatty acids under the curve ogtt (mmol/l * min): 29.55,fat mass (%): 32.6,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.00702727,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 369,log10 homair (insulin resistance index based on homa): 0.452638161,log10 homais (insulin secretion index based on homa): 2.064457989,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.781353197,insgenin (insulinogenic index): 1.945140196,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.716771032,log10 matsuda insulin sensitivity index: 0.480077272,muscle mass (%): 38,lg10 serum c-reactive protein (mg/l): 0.479719235,lg10 plasma adiponectin (mg/l): 0.939519253,ogtt fasting plasma free fatty acid (mmol/l): 0.62,ogtt 30 min plasma free fatty acid (mmol/l): 0.3,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5.5,ogtt 30 min plasma glucose (mmol/l): 10,ogtt 120 min plasma glucose (mmol/l): 7.7,log10 il1 receptor antagonist (pg/ml): 2.230448921,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 1.064457989,ogtt 30 min plasma insulin (mu/l): 1.890421019,ogtt 120 min plasma insulin (mu/l): 2.06595298,log10 ogtt fasting plasma proinsulin (pm/l): 1.245512668,ogtt 30 min plasma proinsulin (pm/l): 1.752048448,ogtt 120 min plasma proinsulin (pm/l): 1.990338855,log10 bioimpedance: Resistance: 2.729974286,log10 bioimpedance (reactance): 1.707570176,waist to hip ratio: 0.949494949,log10 serum bilirubin (umol/l): 1.204119983,log10 serum alanine aminotransfrase (u/l): 1.397940009,log10 creatinine (umol/l): 1.838849091,log10 total cholesterol (mmol/l): 0.872156273,log10 ldl cholesterol (mmol/l): 0.717670503,log10 hdl cholesterol (mmol/l): 0.164352856,log10 total triglycerides (mmol/l): 0.322219295,log10 serum apoa1 (g/l): 0.184691431,log10 serum apob (g/l): 0.195899652,log10 urinary albumin excretion rate (ug/min): 1.061017763
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098262
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249161,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978089
GSM1098262
GSE45159
0.077092
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1924
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249161
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978089
1
age: 59,tissue: adipose tissue,log10 body mass index: 1.437267632,log10 basal metabolic rate (kcal): 1532,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.553362644,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.681414034,plasma free fatty acids under the curve ogtt (mmol/l * min): 28.2,fat mass (%): 21.2,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.21249639,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 478.5,log10 homair (insulin resistance index based on homa): 0.529259204,log10 homais (insulin secretion index based on homa): 2.031408464,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.914427786,insgenin (insulinogenic index): 2.115438343,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.862298384,log10 matsuda insulin sensitivity index: 0.357558815,muscle mass (%): 44,lg10 serum c-reactive protein (mg/l): -0.107348966,lg10 plasma adiponectin (mg/l): 0.740362689,ogtt fasting plasma free fatty acid (mmol/l): 0.44,ogtt 30 min plasma free fatty acid (mmol/l): 0.33,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.9,ogtt 30 min plasma glucose (mmol/l): 11.7,ogtt 120 min plasma glucose (mmol/l): 8.8,log10 il1 receptor antagonist (pg/ml): 2.205826659,log10 il1 beta (pg/ml): -0.769551079,log10 ogtt fasting plasma insulin (mu/l): 1.11058971,ogtt 30 min plasma insulin (mu/l): 2.1430148,ogtt 120 min plasma insulin (mu/l): 2.058805487,log10 ogtt fasting plasma proinsulin (pm/l): 1.382017043,ogtt 30 min plasma proinsulin (pm/l): 1.809559715,ogtt 120 min plasma proinsulin (pm/l): 2.174931594,log10 bioimpedance: Resistance: 2.681241237,log10 bioimpedance (reactance): 1.698970004,waist to hip ratio: 1.005208333,log10 serum bilirubin (umol/l): 1.204119983,log10 serum alanine aminotransfrase (u/l): 1.414973348,log10 creatinine (umol/l): 1.892094603,log10 total cholesterol (mmol/l): 0.840733235,log10 ldl cholesterol (mmol/l): 0.679427897,log10 hdl cholesterol (mmol/l): 0.120573931,log10 total triglycerides (mmol/l): 0.480006943,log10 serum apoa1 (g/l): 0.173186268,log10 serum apob (g/l): 0.195899652,log10 urinary albumin excretion rate (ug/min): 0.600721641
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098263
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249162,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978090
GSM1098263
GSE45159
0.287799
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1932
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249162
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978090
1
age: 58,tissue: adipose tissue,log10 body mass index: 1.424048892,log10 basal metabolic rate (kcal): 1620,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.51621818,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.63545403,plasma free fatty acids under the curve ogtt (mmol/l * min): 5.25,fat mass (%): 19.5,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.292321633,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): -21,log10 homair (insulin resistance index based on homa): 0.128399269,log10 homais (insulin secretion index based on homa): 1.770464422,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.417239121,insgenin (insulinogenic index): 2.864511081,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.344470872,log10 matsuda insulin sensitivity index: 0.927031951,muscle mass (%): 46.5,lg10 serum c-reactive protein (mg/l): 0.216957207,lg10 plasma adiponectin (mg/l): 0.934498451,ogtt fasting plasma free fatty acid (mmol/l): 0.13,ogtt 30 min plasma free fatty acid (mmol/l): 0.04,ogtt 120 min plasma free fatty acid (mmol/l): 0.02,ogtt fasting plasma glucose (mmol/l): 5.4,ogtt 30 min plasma glucose (mmol/l): 5.8,ogtt 120 min plasma glucose (mmol/l): 4.4,log10 il1 receptor antagonist (pg/ml): 2.138618434,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.748188027,ogtt 30 min plasma insulin (mu/l): 1.7355989,ogtt 120 min plasma insulin (mu/l): 1.350248018,log10 ogtt fasting plasma proinsulin (pm/l): 1.086359831,ogtt 30 min plasma proinsulin (pm/l): 1.418301291,ogtt 120 min plasma proinsulin (pm/l): 1.5774918,log10 bioimpedance: Resistance: 2.665580991,log10 bioimpedance (reactance): 1.72427587,waist to hip ratio: 0.92,log10 serum bilirubin (umol/l): 1.230448921,log10 serum alanine aminotransfrase (u/l): 1.041392685,log10 creatinine (umol/l): 1.903089987,log10 total cholesterol (mmol/l): 0.790988475,log10 ldl cholesterol (mmol/l): 0.580924976,log10 hdl cholesterol (mmol/l): 0.303196057,log10 total triglycerides (mmol/l): -0.180456064,log10 serum apoa1 (g/l): 0.198657087,log10 serum apob (g/l): -0.008773924,log10 urinary albumin excretion rate (ug/min): 0.655218496
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098264
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249163,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978091
GSM1098264
GSE45159
0.387732
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1937
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249163
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978091
1
age: 53,tissue: adipose tissue,log10 body mass index: 1.457369054,log10 basal metabolic rate (kcal): 1843,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.328130476,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.816160727,plasma free fatty acids under the curve ogtt (mmol/l * min): 22.65,fat mass (%): 20.5,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.476746204,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 112.5,log10 homair (insulin resistance index based on homa): 0.139179176,log10 homais (insulin secretion index based on homa): 1.917330426,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.678108474,insgenin (insulinogenic index): 2.55590019,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.635393259,log10 matsuda insulin sensitivity index: 0.785164788,muscle mass (%): 45.8,lg10 serum c-reactive protein (mg/l): -0.024568191,lg10 plasma adiponectin (mg/l): 0.908485019,ogtt fasting plasma free fatty acid (mmol/l): 0.32,ogtt 30 min plasma free fatty acid (mmol/l): 0.26,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 5,ogtt 30 min plasma glucose (mmol/l): 6.8,ogtt 120 min plasma glucose (mmol/l): 5.1,log10 il1 receptor antagonist (pg/ml): 2.191730393,log10 il1 beta (pg/ml): -1.045757491,log10 ogtt fasting plasma insulin (mu/l): 0.792391689,ogtt 30 min plasma insulin (mu/l): 2.057285644,ogtt 120 min plasma insulin (mu/l): 1.348304863,log10 ogtt fasting plasma proinsulin (pm/l): 1.146128036,ogtt 30 min plasma proinsulin (pm/l): 1.657055853,ogtt 120 min plasma proinsulin (pm/l): 1.657055853,log10 bioimpedance: Resistance: 2.598790507,log10 bioimpedance (reactance): 1.62324929,waist to hip ratio: 0.961538462,log10 serum bilirubin (umol/l): 1.477121255,log10 serum alanine aminotransfrase (u/l): 1.462397998,log10 creatinine (umol/l): 1.959041392,log10 total cholesterol (mmol/l): 0.835056102,log10 ldl cholesterol (mmol/l): 0.695481676,log10 hdl cholesterol (mmol/l): 0.029383778,log10 total triglycerides (mmol/l): 0.394451681,log10 serum apoa1 (g/l): 0.158362492,log10 serum apob (g/l): 0.184691431,log10 urinary albumin excretion rate (ug/min): 0.839703008
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098265
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249164,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978092
GSM1098265
GSE45159
0.245199
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1942
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249164
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978092
1
age: 57,tissue: adipose tissue,log10 body mass index: 1.415417138,log10 basal metabolic rate (kcal): 1674,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.673759129,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.838816248,plasma free fatty acids under the curve ogtt (mmol/l * min): 27.75,fat mass (%): 20,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 9.8587581,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 232.5,log10 homair (insulin resistance index based on homa): -0.057275608,log10 homais (insulin secretion index based on homa): 1.470781077,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.004235366,insgenin (insulinogenic index): 1.198367654,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 3.883661435,log10 matsuda insulin sensitivity index: 1.136934161,muscle mass (%): 46.3,lg10 serum c-reactive protein (mg/l): -0.441291429,lg10 plasma adiponectin (mg/l): 0.770852012,ogtt fasting plasma free fatty acid (mmol/l): 0.41,ogtt 30 min plasma free fatty acid (mmol/l): 0.33,ogtt 120 min plasma free fatty acid (mmol/l): 0.04,ogtt fasting plasma glucose (mmol/l): 5.8,ogtt 30 min plasma glucose (mmol/l): 9.6,ogtt 120 min plasma glucose (mmol/l): 5.9,log10 il1 receptor antagonist (pg/ml): 2.232614983,log10 il1 beta (pg/ml): -0.744727495,log10 ogtt fasting plasma insulin (mu/l): 0.531478917,ogtt 30 min plasma insulin (mu/l): 1.127104798,ogtt 120 min plasma insulin (mu/l): 1.264817823,log10 ogtt fasting plasma proinsulin (pm/l): 1.012837225,ogtt 30 min plasma proinsulin (pm/l): 1.294466226,ogtt 120 min plasma proinsulin (pm/l): 1.635483747,log10 bioimpedance: Resistance: 2.679427897,log10 bioimpedance (reactance): 1.73239376,waist to hip ratio: 0.954773869,log10 serum bilirubin (umol/l): 1.230448921,log10 serum alanine aminotransfrase (u/l): 1.146128036,log10 creatinine (umol/l): 1.86332286,log10 total cholesterol (mmol/l): 0.728353782,log10 ldl cholesterol (mmol/l): 0.498310554,log10 hdl cholesterol (mmol/l): 0.303196057,log10 total triglycerides (mmol/l): -0.259637311,log10 serum apoa1 (g/l): 0.238046103,log10 serum apob (g/l): -0.055517328,log10 urinary albumin excretion rate (ug/min): 0.62324929
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098266
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249165,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978093
GSM1098266
GSE45159
0.200733
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM1984
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249165
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978093
1
age: 66,tissue: adipose tissue,log10 body mass index: 1.460261511,log10 basal metabolic rate (kcal): 1512,log2 estimated glomerular filtration rate using modification of diet in renal disease (egfr_mdrd): 6.397147031,log10 estimated creatinine clearance rate using cockfcrot-gault formulaa (egcr): 6.435113409,plasma free fatty acids under the curve ogtt (mmol/l * min): 27,fat mass (%): 31.8,log10 plasma glucose area under the curve (ogtt) (mmol/l * min): 10.19229281,plasma glucose area under the curve above basal (ogtt) (mmol/l * min): 390,log10 homair (insulin resistance index based on homa): 0.246060674,log10 homais (insulin secretion index based on homa): 1.609238576,log10 insulin area under the curve (ogtt) (pmol/l * min): 4.514268758,insgenin (insulinogenic index): 1.671257016,log10 insulin area under the curve above basal (ogtt) (pmol/l * min): 4.45158689,log10 matsuda insulin sensitivity index: 0.685639426,muscle mass (%): 46.8,lg10 serum c-reactive protein (mg/l): 0.533772058,lg10 plasma adiponectin (mg/l): 0.819543936,ogtt fasting plasma free fatty acid (mmol/l): 0.41,ogtt 30 min plasma free fatty acid (mmol/l): 0.31,ogtt 120 min plasma free fatty acid (mmol/l): 0.05,ogtt fasting plasma glucose (mmol/l): 6.5,ogtt 30 min plasma glucose (mmol/l): 10.9,ogtt 120 min plasma glucose (mmol/l): 9.3,log10 il1 receptor antagonist (pg/ml): 2.157093949,log10 il1 beta (pg/ml): -0.698970004,log10 ogtt fasting plasma insulin (mu/l): 0.785329835,ogtt 30 min plasma insulin (mu/l): 1.607455023,ogtt 120 min plasma insulin (mu/l): 1.812913357,log10 ogtt fasting plasma proinsulin (pm/l): 1.025305865,ogtt 30 min plasma proinsulin (pm/l): 1.348304863,ogtt 120 min plasma proinsulin (pm/l): 1.701567985,log10 bioimpedance: Resistance: 2.666517981,log10 bioimpedance (reactance): 1.716003344,waist to hip ratio: 0.99047619,log10 serum bilirubin (umol/l): 1.176091259,log10 serum alanine aminotransfrase (u/l): 1.342422681,log10 creatinine (umol/l): 1.924279286,log10 total cholesterol (mmol/l): 0.805500858,log10 ldl cholesterol (mmol/l): 0.633468456,log10 hdl cholesterol (mmol/l): 0.120573931,log10 total triglycerides (mmol/l): 0.209515015,log10 serum apoa1 (g/l): 0.170261715,log10 serum apob (g/l): 0.120573931,log10 urinary albumin excretion rate (ug/min): 0.542262243
675 Charles E. Young Dr. S. MRL 3220
Los Angeles
USA
University of California Los Angeles
Mete,,Civelek
Illumina Casava1.7 software used for basecalling.,Sequencing files were converted to FASTQ format using a custom Perl script. In order to assign the indexed reads to each METSIM subject, a Python script was used to partition the FASTQ file for each lane into multiple FASTQ files, each of which corresponded to one individual. To be included in the partitioned file, a read had to have an exact match to one of the 48 possible index sequences, unambiguously identifying the individual from whom the read originated. Reads which did not exactly match an index sequence were discarded.,The reads were then aligned to the hg19 version of the genome using the Novoalign tool with the following settings: -l16 -t30 -h90 -rA -R 1 -m -g 200 –k. Novoalign software trims the adapter sequences while aligning.,We used the Bioconductor package GenomicRanges for R (v.2.14.0) to count the number of reads with alignment coordinates that overlap the coordinates of known mature miRNAs. Whenever a read mapped to 'x' genomic loci, the read would contribute a count of 1/x to those regions. The genomic coordinates for known mature miRNAs were downloaded from miRBase version 18.,To enable comparison of counts between samples, we normalized the expression values by dividing the counts for a given mature miRNA by the sum of all the miRNA counts for the corresponding individual. For subsequent analysis, we considered the expression levels of 356 miRNAs that had at least 5 reads in half of the study participants.,Genome_build: hg19,Supplementary_files_format_and_content: tab-delimited text files include unnormalized weighted counts for each miRNA
Abdominal subcutaneous adipose tissue was obtained with needle biopsies. Total RNA was isolated from the adipose tissue using Qiagen miRNeasy kit according to manufacturer’s instructions. RNA Integrity Number (RIN) values were assessed with the Agilent Bioanalyzer 2100 instrument. Samples with RIN values greater than 7.0 were used for transcriptional profiling.,Small RNA libraries were prepared using the Illumina TruSeq Small RNA protocol utilizing up to 48 unique index sequences (Illumina Catalog Number FC-102-1009). For two samples (METSIM490 and METSIM6589) we also prepared small RNA libraries using Illumina Small RNA v1.5 sample preparation protocol (Catalog # FC-930-1501)
GSM1098267
Illumina HiSeq 2000
May 15 2019
size fractionation
transcriptomic
RNA-Seq
total RNA
Homo sapiens
GPL11154
SRA: https://www.ncbi.nlm.nih.gov/sra?term=SRX249166,BioSample: https://www.ncbi.nlm.nih.gov/biosample/SAMN01978094
GSM1098267
GSE45159
0.208397
adipose tissue
Public on Apr 01 2013
Mar 14 2013
9606
METSIM2019
SRA
https://www.ncbi.nlm.nih.gov/sra?term=SRX249166
https://www.ncbi.nlm.nih.gov/biosample/SAMN01978094