Background: Most of the 215 risk loci associated with inflammatory bowel diseases (IBD) to date were discovered by genotyping arrays and are driven by common variants. Such assays, however, do not adequately capture lower frequency variation. Thus, the role of these variants in IBD pathogenesis is unclear.
Methods: To comprehensively interrogate the role of lower frequency variation, we whole-genome sequenced (WGS) 4280 IBD patients and compared them to 3652 population controls. To increase power to detect association, we imputed into new and existing GWAS cohorts totalling over 35000 individuals, using a reference panel augmented with our sequence data. Finally, we combined our data with publicly available summary statistics and conducted meta-analyses in ∼60,000 individuals.
Results:Rare variants: We identified an excess burden of rare, damaging missense variants in genes previously implicated in IBD, suggesting that rare variants are likely contributors to IBD pathogenesis. However, no such excess burden was confidently detected in any single gene and much larger sample sizes will be required for their identification.
Low frequency variants: Analysis of imputed data identified a missense (Asp439Glu) variant in ADCY7 with 0.6% frequency in the general population, which doubles risk of ulcerative colitis. Despite good power to detect associations of this type, this was the only variant we detected, suggesting a minimal contribution to disease susceptibility by low frequency variants.
Common variants: Our meta-analysis identified 24 novel risk loci, including three which contain integrin genes (ITGA4, ITGAV, ITGB8). At two of these, as well as at the previously associated ITGAL and ICAM1 loci, we found strong evidence that the IBD risk-increasing variant increases expression of the respective integrin gene in activated monocytes. This hints at a mechanism linking non-coding genetic associations to targets of existing therapeutics. We also identified likely causal missense variants in PLCG2, mutations in which are known to cause a primary immunodeficiency, and SLAMF8, a negative regulator of inflammatory responses.
Conclusions: We conducted a large, multi-faceted study to explore the genetic architecture of IBD across the entire allele frequency spectrum. Our results highlight the continued value of GWAS and their potentially pivotal role in understanding aspects of disease biology through the integration of genomic and functional datasets in specific cells and contexts. We found minimal evidence for strong effects from low frequency variants, despite good power, while the effects of rare variants will require larger sample sizes to be more thoroughly investigated.