Tissue-specific genetic changes in gene and protein expression drive type 2 diabetes risk
A multi-ancestry Mendelian randomization and colocalization study shows genetic effects on gene and protein levels influence type 2 diabetes risk in a tissue-dependent way, and many causal signals are not visible in blood alone. The results stress the need for diverse ancestry QTL data and disease-relevant tissues to map mechanisms and prioritize targets.
- Using T2DGWAS data from about 2.5 million people and cis-QTLs, blood-based analyses identified causal effects for 335 genes and 46 proteins on T2D risk with replication rates of 16.4% for genes and 50% for proteins.
- Analyses across seven T2D-relevant tissues found causal links for 676 genes and a total of 928 genes when combining tissue and blood results, revealing many signals missed in blood-only studies.
- Causal effects were largely shared across ancestries with low ancestry-related heterogeneity, but hundreds of genes and proteins were detectable only in non-European QTL datasets because of allele frequencies and instrument availability.
- Causal effects showed strong tissue specificity: only 18% of genes identified in a primary tissue were also significant in blood and 75% of genes tested across tissues had evidence of heterogeneity.
- The study underscores that mapping T2D biology and drug targets requires both broader ancestry representation in QTL studies and molecular data from disease-relevant tissues rather than relying on blood alone