Integrating tissue-specific transcriptome-wide association studies (TWAS) with ancestry-aware gene expression models, this analysis of the largest European T2D GWAS (242,283 cases; 1,569,734 controls) identified 684–750 significant gene-T2D associations per tissue across six metabolically relevant sites: subcutaneous adipose, visceral adipose, hypothalamus, liver, skeletal muscle, and pancreas. Conditional analyses narrowed these to 289–322 independent signals per tissue. JAZF1 and IDE showed consistent associations across all six tissues, while TCF7L2 and WFS1 displayed tissue-restricted effects. Cross-ancestry replication in African American individuals (50,251 cases) confirmed enrichment in adipose, muscle, and pancreas.
This work matters because pinpointing causal genes—not just associated loci—is the critical bottleneck between GWAS discovery and therapeutic development. By revealing that canonical T2D genes like TCF7L2 act in tissue-specific windows, it reframes how drug targets should be prioritized and validated. The IDE finding is particularly intriguing: IDE (insulin-degrading enzyme) has long been discussed as a therapeutic target, and cross-tissue consistency strengthens its candidacy. Practically, these regulatory maps could eventually guide precision medicine approaches stratifying T2D by tissue-driven mechanisms rather than treating it as a monolithic disease.
Key limitations deserve weight: the AFA gene expression models were trained on only 111 individuals, making cross-ancestry inference statistically fragile. The study is observational and correlative—TWAS infers, but does not prove, causal gene action. As a preprint not yet peer-reviewed, findings may be revised. Overall, this is solid incremental science advancing functional genomics, not a paradigm shift.