SRA

Non-coding regions of the genome contain cis-regulatory elements (CREs) that govern transcriptional activity critical for development and physiologic functions of major metabolic tissues, such as islets, adipose, liver, and skeletal muscle. Moreover, they contain the majority of genetic variants associated with risk of type 2 diabetes and quantitative alterations in related metabolic phenotypes. Here, we defined comprehensive, representative CRE maps of these major metabolic tissues in mice of both sexes using ATAC and pcHiC data and incorporating genetic variation of the 8 major Collaborative Cross founders and diet-response. Cross-species comparison of these representative maps with those from corresponding human tissues revealed conservation and functional preservation of ~28% of CREs in one or more major metabolic tissues. Importantly, this included an average of 700 (1.8%) cross-species concordant peaks that harbored genetic variants associated with T2D and related metabolic traits, including variants in the CAMK1D/CDC123, C2CD4A/B, and ADCY5 loci. Germline knock-out of the cross-species concordant CRE in ADCY5 yielded mice with impaired fasting glucose, demonstrating the importance of these functionally preserved CREs to glucose homeostasis and genetic risk of type 2 diabetes and progression. Together, the creation and cross-species comparison of comprehensive CRE maps in major metabolic tissues provides a critical resource for variant-to-function studies of type 2 diabetes and an array of related metabolic and cardiovascular traits and demonstrates the power of these mapping strategies to identify conserved metabolic cis-regulatory networks and enable data-driven creation of new preclinical models of diabetes and related metabolic diseases. Sample preparation: Promotor capture libraries were constructed via the following method. Crosslinked cells were lysed, digested with DpnII, ends filled in using biotin-14-dATP followed by overnight proximity ligation. Ligated samples underwent crosslink reversal, DNA extraction with proteinase K, and biotin removal from unligated fragments. DNA was fragmented to 500-600 bp, column purified, and biotinylated fragments pulled down. Biotinylated fragments were converted into Illumina compatible libraries using the SureSelect XT HS2 DNA System and a custom SureSelect probe set (design ID S3346176) (Agilent Technologies) according to the manufacture's protocol with the following exception: pre-capture PCR was performed as four 50 ul reactions that were recombined during clean-up. The quality and concentration of the libraries were assessed using the High Sensitivity D5000 ScreenTape (Agilent Technologies) and Qubit dsDNA HS Assay (ThermoFisher), respectively, according to the manufacturers' instructions. Libraries were sequenced 150 bp paired end on an Illumina NovaSeq 6000. Overall design: The total target bait design resulted in 125,187 in silico targeted fragments. In initial analysis testing, we found that the full capture array resulted in too little coverage per fragment. Therefore, to increase the coverage per fragment we binned all in silico digest fragments and targeted fragments into 5kb virtual bins with the 'makeBins.R' script from the Chicago package [PMID: 27306882] with the '--include_baits' flag. The binned files were passed the 'makeDesignFiles_py3.py' script from the Chicago package with '6-cutter suggested settings' per suggestions in Freire-Pritchett, et al. 1. Binned fragments and design files were used in all downstream analysis. pcHiC data for three strains (C57BL/6J, NZO/HlLtJ, and CAST/EiJ) from, islet, liver, and adipose tissues were processed in the following way. Reads were quality trimmed with AGeNT Trim (https://www.agilent.com/en/product/next-generation-sequencing/hybridization-based-next-generation-sequencing-ngs/ngs-software/agent-232879). Trimmed reads were processed with the HiCUP pipeline [PMID: 26835000]. Following the HiCUP filter step, mapped, hicup_filter alignments were demultiplex with UMI aware AGeNT Locatit. On target 'captured.bam' files were then generated for demultiplex alignments with `HiCUP_Capture` a tool from the HiCUP pipeline. On target captured BAMs files were converted to Chicago format with 'bam2chicago_V02.sh' using the 5kb binned fragments and target baitmap as inputs 2. The resulting 'chinput' files were process with Chicago via the 'runChicago.R' script with suggested 6-base-cutter settings. Processed resulting output with 'fitDistCurve.R' with the flag '-t -5' to obtain adjusted p-value weight settings relative to our data. The 'chinput' were then reprocessed with 'runChicago.R' with final settings from the re-weighting analysis. Due to difficulty associated with harvesting tissue, Islet and adipose samples yielded low per sample coverage. Therefore, to maintain consistency across all tissues on target alignments were pooled by strain and pooled data were used in the downstream Chicago analysis. 1 Freire-Pritchett, P. et al. Detecting chromosomal interactions in Capture Hi-C data with CHiCAGO and companion tools. Nat Protoc 16, 4144-4176 (2021). https://doi.org:10.1038/s41596-021-00567-5 2 Cairns, J. et al. CHiCAGO: robust detection of DNA looping interactions in Capture Hi-C data. Genome Biol 17, 127 (2016). https://doi.org:10.1186/s13059-016-0992-2