Abstract
BACKGROUND: Genome-wide association studies of inflammatory bowel disease have identified hundreds of common genetic variants that are associated with inflammatory bowel disease, but few promising therapeutic targets. The "omnigenic" sparse effector hypothesis postulates that the polygenic effects of common SNPs on a typical complex trait are mediated by trans-effects that coalesce on the expression of a sparse set of core genes. The objective of this study was to identify core genes for inflammatory bowel disease. METHODS: Using summary statistics from studies of transcript levels in whole blood or proteins in plasma, we constructed genome-wide aggregated trans-effects (GATE) scores for predicted gene expression in the UK Biobank cohort and tested these scores for association with inflammatory bowel disease (7949 cases, 452 790 noncases). RESULTS: Inflammatory bowel disease was inversely associated with GATE scores for 5 interferon-stimulated genes-IFIT1, IFI44, HERC5, MX1, IFI44L-regulated by the same trans-expression quantitative trait locus, and with the GATE score for IFNL1. For 6 other genes, GATE score associations with inflammatory bowel disease were supported by other criteria: reported associations with nearby genetic variants, perturbation in experimental models, association with measured protein levels, or drug effects. CONCLUSIONS: These results implicate down-regulation of Type III interferon signaling as a core pathway in the etiology of inflammatory bowel disease, supported by reports of monogenic inflammatory bowel disease caused by rare loss-of-function variants and by perturbation in experimental models of colitis. Deficient Type III interferon signaling may be amenable to therapeutic intervention.