PROJECT SUMMARY Type 2 diabetes (T2D) is a highly complex and heterogenous disease, affecting more than 400 million individuals worldwide. While genome-wide association studies (GWAS) have identified more than 500 genomic loci that are associated with T2D, most of these associations fail to delineate the causal genes and explain the molecular mechanisms underlying disease pathology. To identify the causal genes, it is possible to exploit expression quantitative trait locus (eQTL) analyses that have assessed the associations between genetic variation and gene expression in the pancreas. Previous eQTL analyses of pancreatic islets have provided valuable insight into the probable causal genes that contribute to T2D risk and disease, however they have not examined whether these genes and their variant effectors are cell type-associated. Furthermore, the extent to which genetic variation affects isoform expression is largely unknown despite an increasing importance of alternative splicing in regulating beta cell function and survival. Advanced computational methods have been developed to accurately resolve cell type populations within bulk heterogenous samples, allowing us to the map the associations between genotype, cell type proportions, and gene expression. In Aim 1 of this proposal, I will perform cellular deconvolution to estimate the proportions of pancreatic cell types in bulk RNA-seq of 305 whole pancreas and 420 pancreatic islets (725 samples in total) and use these estimations to generate a focused cell-type associated eQTL map. In Aim 2, I will perform Bayesian colocalization between cell type-associated eQTLs and GWAS loci to infer cellular mechanisms affecting T2D traits and disease. To identify putative causal regulatory variants within the associated GWAS loci, I will conduct a fine-mapping study that integrates GWAS summary statistics with functional annotations from epigenomic data and transcription factor (TF) binding activities from SELEX- seq. In Aim 3, I will functionally validate the fine-mapped variants by testing whether or not they alter transcription factor footprints in distinct pancreatic cell types using snATAC-seq. Next, using computational predictions from this analysis and from Aims 1 and 2, I will prioritize 10 putative causal regulatory variants to interrogate their direct effects on nuclear factor binding using electrophoretic mobility shift assay. Altogether, this proposal will employ advanced computational methods to identify T2D risk variants that alter gene or isoform expression and TF binding and determine the specific pancreatic cell types in which they are functional. These results will provide novel insights into the genetic basis and cellular origins of T2D risk and disease.