Islet-specific DNA hypomethylation identifies gene enhancer loci driving Type 2 Diabetes risk

Genome-wide association studies (GWAS) have identified hundreds of genetic loci linked to pancreatic islet dysfunction and type 2 diabetes (T2D). Many of these loci map to islet genes and their enhancers; however, a complete understanding of the gene regulatory contributions to T2D is lacking. We have previously shown that cell-type specific DNA hypomethylation patterns identify gene enhancers that contribute to the heritability of cell-type relevant traits. Here, we determine a reference dataset of islet hypomethylated regions (HMRs) from DNA methylomes of donors without T2D. Comparing islet HMRs to profiles from functionally diverse cell types, we identify ~4,800 islet-specific HMRs and demonstrate their enrichment for key genes and transcription factor motifs essential for beta cell function. We further show that islet-specific HMRs are highly enriched for GWAS variants that contribute specifically to T2D heritability. Leveraging genotype-phenotype data from the Vanderbilt BioVU biobank, which links patient DNA to electronic health records (EHR), we perform phenome- and lab-wide association scans that replicate islet HMR associations with T2D and uncover novel links to related clinical traits. To overcome challenges posed by stringent GWAS significance thresholds, we develop an HMR-WAS, which allows detection of T2D signals in our smaller BioVU cohort. Our comprehensive framework reveals enhancers of genes involved in insulin secretion and glucose metabolism and demonstrates the impact of T2D-associated variants on their transcriptional activity. By integrating HMRs with patient EHR, our work underscores the potential to reveal new enhancer loci associated with disease risk, especially those missed in a conventional GWAS.