273-OR: Initial eQTL Discovery in Individual Pancreatic Cell Types from the Human Pancreas Analysis Program

The pancreas is composed of multiple, distinct cell types with etiologically relevant functions for diabetes. Pancreatic studies have collected data largely from bulk tissue, limiting the elucidation of how each pancreatic cell type contributes to disease. Single-cell technologies coupled to tissue collection efforts have expanded for the pancreas, facilitating novel characterization of genetics, expression, and regulation of these cell types at a population scale and across disease states. A leading effort generating this data, the Human Pancreas Analysis Program, aims to comprehensively study the human pancreas in diabetes and integrate data with the clinical history from de-identified donors. Data from this program provides an opportunity to investigate how each of the pancreas’ cell types uniquely contributes to metabolic traits. Here, we report an initial discovery effort to identify genetic variation associated with gene transcript abundance - i.e., expression quantitative trait loci (eQTL) - using 394,151 cells collected from single-cell RNA sequencing of 80 pancreatic donors. Clustering analysis of this data identified pancreatic islets, pancreatic exocrine cells and immune cell types. We initially focused on populations of alpha, beta and delta cells, which had sufficient cell numbers across donors. Using standard analysis pipelines, we observed 1,696, 1,744 and 946 significant eGenes in alpha, beta and delta cells (permutation p-value <0.05 and q-value <0.05), respectively. Formal colocalization with signals identified from a recent type 2 diabetes GWAS (Vujkovic et al, 2020) and our eQTL data identified 19 SNP-to-gene connections, all of which were unique to one of our identified cell types. One eGene, NCK1, has been previously investigated in beta cells, but our colocalization has implicated it in an alpha cell context. Future colocalizations and meta-analyses should improve nomination of putative effector transcripts underlying pancreatic disease.