Large-scale analyses provide no evidence for gene-gene interactions influencing type 2 diabetes risk

A growing number of genetic loci have been shown to influence individual predisposition to type 2 diabetes (T2D). Despite longstanding interest in understanding whether non-linear interactions between these risk-variants additionally influence T2D-risk, the ability to detect significant gene-gene interaction (GGI) effects has to date been limited. To increase power to detect GGI effects, we combined recent advances in the fine-mapping of causal T2D-risk variants with the increased sample size available within UK Biobank (375,736 unrelated European participants, including 16,430 T2D cases). In addition to conventional single variant-based analysis, we employed a complementary polygenic score-based approach which included partitioned T2D-risk scores that capture biological processes relevant to T2D pathophysiology. Nevertheless, we found no evidence in support of GGI effects influencing T2D-risk. The present study was powered to detect interactions between common variants with odds ratios >1.2, so these findings place limits on the contribution of GGIs to the overall heritability of T2D.