SRX11616706 1885 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652006 Islet_RNAseq_1885 1885 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616707 1904 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652005 Islet_RNAseq_1904 1904 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616708 1910 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652008 Islet_RNAseq_1910 1910 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616709 1942 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652007 Islet_RNAseq_1942 1942 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616710 1949 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652009 Islet_RNAseq_1949 1949 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616711 1950 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652011 Islet_RNAseq_1950 1950 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616712 1946 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652010 Islet_RNAseq_1946 1946 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616713 TCCAACTGAA-TAGACCGGCT SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652012 Islet_RNAseq_1741 TCCAACTGAA-TAGACCGGCT RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616714 CCTCGTAATA-TCAACAGTTG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652013 Islet_RNAseq_1744 CCTCGTAATA-TCAACAGTTG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616715 CCAATTAGGC-TGTTCCATGG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652014 Islet_RNAseq_1745 CCAATTAGGC-TGTTCCATGG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616716 AATCAGACGA-AGTAAGCGGT SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652015 Islet_RNAseq_1748 AATCAGACGA-AGTAAGCGGT RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616717 CATTCGATTC-ATAAGCGGAG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652017 Islet_RNAseq_1753 CATTCGATTC-ATAAGCGGAG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616718 1908 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652016 Islet_RNAseq_1908 1908 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616719 CGGAGCATCT-ACGTAGGATA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652018 Islet_RNAseq_1756 CGGAGCATCT-ACGTAGGATA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616720 AAGAGTGGTC-CTCGCTGAGT SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652019 Islet_RNAseq_1759 AAGAGTGGTC-CTCGCTGAGT RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616721 ACTTCTGAGC-ATGGTGGTGC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652020 Islet_RNAseq_1763 ACTTCTGAGC-ATGGTGGTGC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616722 AGACATATGG-ATCAACATCG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652023 Islet_RNAseq_1569 AGACATATGG-ATCAACATCG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616723 TCCGAAGTGG-ACACAATGGT SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652021 Islet_RNAseq_1571 TCCGAAGTGG-ACACAATGGT RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616724 GTAACACCGT-ACGCTTGGTA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652022 Islet_RNAseq_1573 GTAACACCGT-ACGCTTGGTA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616725 CGGAAGATGG-ATGACCACTA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652024 Islet_RNAseq_1575 CGGAAGATGG-ATGACCACTA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616726 AGAACTGGTG-TGGATTGTAG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652025 Islet_RNAseq_1592 AGAACTGGTG-TGGATTGTAG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616727 TCTAGGATTG-ATAGCACCTA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652026 Islet_RNAseq_1595 TCTAGGATTG-ATAGCACCTA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616728 AATCCTTCCG-ACATCGGTCA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652028 Islet_RNAseq_1616 AATCCTTCCG-ACATCGGTCA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616729 1905 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652027 Islet_RNAseq_1905 1905 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616730 AGAAGGCCAG-GACGTCGATA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652029 Islet_RNAseq_1627 AGAAGGCCAG-GACGTCGATA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616731 TCCAGAATGT-TGGTCCAATT SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652030 Islet_RNAseq_1657 TCCAGAATGT-TGGTCCAATT RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616732 AGCACCAGTC-ATCATGCTAC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652031 Islet_RNAseq_1663 AGCACCAGTC-ATCATGCTAC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616733 TCATCCGTGA-TTAGGAGGAA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652032 Islet_RNAseq_1743 TCATCCGTGA-TTAGGAGGAA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616734 AACACCAATG-GCACTTACAA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652033 Islet_RNAseq_1746 AACACCAATG-GCACTTACAA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616735 AGGACTGAAC-AGTGAGCAAC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652034 Islet_RNAseq_1747 AGGACTGAAC-AGTGAGCAAC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616736 CGGTGCACTT-ACCAACTGTC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652035 Islet_RNAseq_1761 CGGTGCACTT-ACCAACTGTC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616737 CGACTCTTAC-TGGCATCTGG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652036 Islet_RNAseq_1764 CGACTCTTAC-TGGCATCTGG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616738 GTCGACTCCT-ACCTTAGCCG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652037 Islet_RNAseq_1814 GTCGACTCCT-ACCTTAGCCG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616739 TGCATATAGG-GAACTGTCGC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652038 Islet_RNAseq_1553 TGCATATAGG-GAACTGTCGC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616740 1906 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652039 Islet_RNAseq_1906 1906 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616741 AACACCGGTT-ACGAGACGTC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652040 Islet_RNAseq_1594 AACACCGGTT-ACGAGACGTC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616742 AAGCATCCTG-GACCGATACA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652041 Islet_RNAseq_1681 AAGCATCCTG-GACCGATACA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616743 CGATAGCAGG-TAAGTGTCGA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652042 Islet_RNAseq_1682 CGATAGCAGG-TAAGTGTCGA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616744 AGACGAGGTC-TAGGCCAACA SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652043 Islet_RNAseq_1683 AGACGAGGTC-TAGGCCAACA RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616745 CCTAACGACT-TAGTGTTACG SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652044 Islet_RNAseq_1686 CCTAACGACT-TAGTGTTACG RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616746 CCATAGATCT-TGTGCTTGAC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652046 Islet_RNAseq_1732 CCATAGATCT-TGTGCTTGAC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616747 GTCACGACTG-TAACGAGCGC SRP330685 bp0 Islets from 32 mice were sequenced: 4 males and 4 females from each diet (high-fat and low-fat) and each age (20wk and 30wk), n=32. Islet RNA was extracted using the RNeasy MinElute Cleanup kit (Qiagen), RNA concentration was measured via Nanodrop and RNA quality/integrity was assessed with a BioAnalyzer (Agilent). Libraries were prepped using the SMARTer cDNA synthesis kit (Takara Bio) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. FASTQ files were trimmed and filtered to remove low quality reads and aligned against a SM/J custom genome using STAR. Read counts for -cell-specific genes were normalized via TMM normalization and pairwise differential expression between cohorts was performed using edgeR31. Differential expression analysis for all 316 -cell-specific genes across select cohort comparisons reported in SRS9652045 Islet_RNAseq_1733 GTCACGACTG-TAACGAGCGC RNA-Seq TRANSCRIPTOMIC RT-PCR Illumina NovaSeq 6000 SRX11616748 1838 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652047 Islet_RNAseq_1838 1838 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616749 1841 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652048 Islet_RNAseq_1841 1841 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616750 1842 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652049 Islet_RNAseq_1842 1842 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616751 1875 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652050 Islet_RNAseq_1875 1875 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000 SRX11616752 1877 SRP330685 bp0 Single cell RNA sequencing (scRNA-seq) was performed on islets isolated from 15 SM/J mice representing 6 cohorts: 20wk high-fat females (n=3), 20wk high-fat males (n=3), 30wk high-fat females (n=3), 30wk high-fat males (n=2), 20wk low-fat females (n=2), and 20wk low-fat males (n=2). Isolated islets were dissociated into single cell suspensions using Accumax cell/tissue dissociation solution (Innovative Cell Technologies). Libraries were prepped using the Chromium Single Cell 3 GEM, Library & Gel Bead Kit v3 (10xGenomics) and sequenced at 2x150 paired end reads using a NovaSeq S4. After sequencing, reads were de-multiplexed and assigned to individual samples. Reads were aligned using 10x Genomics CellRanger (3.1.0) against our custom SM/J reference22. Samples that were prepped together were aggregated into batches using CellRanger aggregate. In the R environment (4.0.0), each aggregated batch was run through SoupX (1.5.0)23 to estimate and correct for ambient RNA contamination. A contamination fraction of 0.05 was chosen. Removal of Ins2 ambient RNA shown in Supplemental Figure 1D-E. Adjusted counts were imported into Seurat (3.2.2)24, where cells were filtered for number of features detected (500-3000), total counts detected (1000-30000), percent mitochondrial genes (0-30), visualized in Supplemental Figure 1A. For additional quality control, we excluded cells where nCount was not predictive of nFeature; the predictive error (residual) of a cell had to be within 3 standard deviations of the mean predictive error (~0). Cell counts for samples from one batch shown before and after filtering step shown in Supplemental Figure 1B-C. Expression was then normalized in Seurat (normalization.method = LogNormalize), batches were integrated, and clustered using a shared nearest neighbor approach. Using the top 10 principal components of the filtered expression data and a resolution of 0.14, we identified 9 clusters of cells using Clustree (0.4.3)25. Cell types were assigned by identifying top over-expressed genes for each cluster relative to all other clusters using a Wilcoxon rank sum test, with an average log-fold-change threshold of >=0.25 and requiring at least 25% of cells express the gene. Identities were assigned by comparing top over-expressed genes for each cluster with known cell-type specific markers for islet cells. SRS9652051 Islet_RNAseq_1877 1877 RNA-Seq TRANSCRIPTOMIC SINGLE CELL RT-PCR Illumina NovaSeq 6000