69-OR: Multiancestry Genome-Wide Association Study Reveals Distinct Biological Pathways and Cell Types Driving Type 2 Diabetes Risk with Heterogeneous Effects across Diverse Populations

Abstract

The Type 2 Diabetes Global Genomics Initiative aggregated genome-wide association studies (GWAS) of type 2 diabetes (T2D) in 428,452 affected individuals and 2,107,149 controls from five major ancestries (39.7% non-European ancestry) to enhance power for locus discovery and improve understanding of the pathophysiological basis of the disease. We identified 1,289 independent association signals (P < 5 x 10-8) at 611 loci, of which 145 were previously unreported. We applied K-means clustering to the signals based on their associations with 37 cardiometabolic traits and identified eight clusters with distinct profiles. Three clusters were related to insulin secretion and had a positive, negative, or no effect on proinsulin levels. These clusters were significantly enriched (P < 0.05/222, Bonferroni correction for the number of cell types) for epigenomic annotations from single-cell ATAC-seq data in pancreatic islets but had different enrichment in pancreatic ductal cells and intestinal endocrine cells. The remaining five clusters related to insulin resistance (IR) physiology: increased body fat; unfavourable fat distribution (decreased gluteofemoral fat and increased visceral fat); obesity (increased body mass index); lipodystrophy (increased fasting insulin and decreased body fat); and liver lipid metabolism (increased liver fat and liver enzymes). IR clusters showed more diverse patterns of enrichment: obesity in adrenal cells; fat distribution in endothelial cells and pericytes in blood vessels; and lipodystrophy in adipocytes. T2D signals were more strongly enriched for ancestry-correlated heterogeneity in effect sizes in insulin secretion clusters than IR clusters. These data demonstrate distinct clusters of T2D association signals representing different trajectories to disease with observable differences in effect sizes across ancestry groups. Disclosure K.Suzuki: None. Diamante consortium: n/a. Va million veteran program: n/a. J.B.Meigs: Consultant; Quest Diagnostics. M.I.Mccarthy: Employee; Genentech, Inc., Stock/Shareholder; Roche Pharmaceuticals. A.Mahajan: Employee; Genentech, Inc. J.M.Mercader: None. C.Spracklen: None. M.Boehnke: None. M.Vujkovic: None. J.I.Rotter: None. K.Hatzikotoulas: None. B.F.Voight: None. E.Zeggini: None. A.Morris: None. Type 2 diabetes global genomics initiative: n/a. L.Southam: None. X.Yin: None. K.Lorenz: None. R.Mandla: None. H.J.Taylor: None. A.Huerta: None. N.W.Rayner: None.