Leveraging Genetic Correlations to Prioritize Drug Groups for Repurposing in Type 2 Diabetes

Type 2 diabetes (T2D) is a complex, polygenic disease with substantial health impact. Despite extensive genome-wide association studies (GWAS) identifying risk loci, therapeutic translation remains limited. We applied a Bayesian Linear Regression (BLR) multi-trait gene set model to prioritize druggable gene sets, integrating GWAS summary statistics with drug-gene interaction data from the Drug Gene Interaction Database (DGIdb). For each drug group, defined at the ATC 4th level, we calculated posterior inclusion probabilities (PIP) to assess relevance. Known antidiabetic agents showed strong associations with T2D, validating the model. Additionally, carboxamide derivatives, fibrates, uric acid inhibitors, and various immunomodulatory and antineoplastic agents demonstrated significant genetic relevance. Gene-level analyses highlighted key T2D-associated genes, including PPARG , KCNQ1 , TNF , and GCK . Notably, bezafibrate, a PPAR pan-agonist, demonstrated substantial genetic overlap with T2D loci, supporting its potential in metabolic disease. This study introduces a genetically informed pipeline for drug repurposing based on multi-trait gene set analysis.