We hypothesized that a SNP aggregation approach that integrates information on whether a SNP regulates the expression of a gene can greatly increase the power to identify trait-associated loci either from a strong functional SNP signal or from a combination of modest signals, the so-called grey area of GWAS. The present study suggests that PrediXcan, a novel method that incorporates information on gene regulation from a set of markers, increases the power to detect associations relative to traditional SNP-based GWAS and known gene-based tests under a broad range of genetic architectures and provides mechanistic insights and more easily interpreted direction of effect into the observed associations.
