Transcriptome-wide gene-gene interaction associations elucidate pathways and functional enrichment of complex traits

It remains unknown to what extent gene-gene interactions contribute to complex traits. Here, we introduce a new approach using predicted gene expression to perform exhaustive transcriptome-wide interaction studies (TWISs) for multiple traits across all pairs of genes expressed in several tissue types. Using imputed transcriptomes, we simultaneously reduce the computational challenge and improve interpretability and statistical power. We discover (in the UK Biobank) and replicate (in independent cohorts) several interaction associations, and find several hub genes with numerous interactions. We also demonstrate that TWIS can identify novel associated genes because genes with many or strong interactions have smaller single-locus model effect sizes. Finally, we develop a method to test gene set enrichment of TWIS associations (E-TWIS), finding numerous pathways and networks enriched in interaction associations. Epistasis is may be widespread, and our procedure represents a tractable framework for beginning to explore gene interactions and identify novel genomic targets.Author summary: We developed a new method to comprehensively test associations of all pairwise gene-gene interactions with complex traits using imputed expression. We applied the method to 12 complex traits in humans across four tissues or cross-tissue expression measures. We found widespread evidence that gene-gene interactions influence traits, and that accounting for interactions identifies loci not previously identified in traditional single-locus association tests, because the interactions mask the main effects when tested in isolation. We next introduced a gene set analysis to test enrichment of interaction associations in pathways and cell types and identify several gene sets within which gene interactions are enriched in the associations with complex traits. Our analyses identify core hub genes that appear to integrate signals across multiple pathways, providing new biological insight into the genetic influences on these traits. Our findings also confirm the role of gene interactions in complex traits, which has long been hypothesized but never before comprehensively tested due to the computational burden required, but which our new approach can efficiently and effectively deal with.