Having validated the utility of MAGENTA, we next used the method to test whether mitochondria-related gene sets are enriched for multiple genes that lie near common variants with modest effects on T2D susceptibility. We tested three molecular hypotheses based on the observations of reduced OXPHOS activity and expression levels and fewer and smaller mitochondria in diabetic muscle (described in the Introduction). The three hypotheses were: DNA variants that alter the function of different nuclear regulators of the OXPHOS pathway and/or other mitochondrial processes are associated with T2D, variants that cause core defects in OXPHOS activity that may result in compensatory alterations of OXPHOS levels are associated with T2D, and variants that affect other mitochondrial functions in addition to the OXPHOS process are associated with T2D. To test these hypotheses, we tested for enrichment of T2D associations in the following three gene sets: a set of known nuclear regulators of mitochondrial genes, the OXPHOS genes, and all known nuclear-encoded mitochondrial genes. In parallel to testing the relevance of these three sets to T2D, we explored their possible associations (in non-diabetic individuals) with seven specific glycemic traits that are risk factors for T2D (listed below).
