1722-P: Colocalization of TOPMed Whole Genome Sequencing Analysis and Tissue-Specific eQTL Signals Detects Target Genes for Type 2 Diabetes Risk

Abstract

Type 2 diabetes (T2D) risk is highly heritable, and although many associated loci have been identified, analyses of regulatory regions in most whole-genome sequencing (WGS) studies have not been specific to relevant tissues. The availability of tissue-specific expression quantitative trait (eQTL) information can further elucidate the functional role of non-coding regulatory variants. A WGS (>38x sequencing depth) association study was performed for variants in 9,663 cases and 35,050 controls from the NHLBI’s Trans-Omics for Precision Medicine (TOPMed) program. eQTL data from 48 different tissues was obtained from the Genotype-Tissue Expression (GTEx) Portal. To generate loci of interest, 200-kb windows were considered upstream and downstream of each variant found to be genome-wide significant (p < 5x10^-8) in both the TOPMed and GTEx datasets. A total of 11 regions in different tissues were significantly associated with T2D status, centered around variants such as rs6585201 in the ascending aorta (TCF7L2, TOPMed p = 1.18x10^-9, GTEx p = 1.12x10^-10, but did not localize to the strongest signal in the region), rs76895963 in the cerebellum (CCND2, TOPMed p = 3.51x10^-9, GTEx p = 4.40x10^-17), and rs4898431 in ovarian tissue (DUSP9, TOPMed p = 1.88x10^-8, GTEx p = 4.39x10^-8). We plan to implement eCAVIAR to examine causal variant colocalization, with the advantages of leveraging summary statistics and accounting for the possibility of more than one causal variant per locus. These results confirm previously published studies, but provide validation of methods we will extend to larger samples and diabetes-related outcomes. Disclosure M.D. Szeto: None. H.M. Highland: None. A. Manning: None. Funding National Institute of Diabetes and Digestive and Kidney Diseases (U01DK078616, K01DK107836)