We applied PrediXcan to seven complex disease phenotypes from the WTCCC study22. For this purpose, we utilized the DGN whole blood elastic net prediction models. We correlated the estimated genetically regulated gene expression for close to 8700 genes with disease status for each WTCCC dataset and identified 41 significant associations (Bonferroni corrected p < 0.05) with five diseases (Table 1). Notably, we identified 29 genes associated with type 1 diabetes (T1D) risk (Table 1 and Fig. 6), 8 of which were outside of the extended MHC. Complete results for the remaining 6 diseases are shown in Supplemental Figures 5 and 6. Consistent with the original GWAS of WTCCC diseases, our most significant results were for autoimmune diseases22.
