TY - JOUR T1 - Extensive epistasis within the MHC contributes to the genetic architecture of celiac disease JF - bioRxiv DO - 10.1101/002485 SP - 002485 AU - Benjamin Goudey AU - Gad Abraham AU - Eder Kikianty AU - Qiao Wang AU - Dave Rawlinson AU - Fan Shi AU - Izhak Haviv AU - Linda Stern AU - Adam Kowalczyk AU - Michael Inouye Y1 - 2014/01/01 UR - http://biorxiv.org/content/early/2014/02/07/002485.abstract N2 - Epistasis has long been thought to contribute to the genetic aetiology of complex diseases, yet few robust epistatic interactions in humans have been detected. We have conducted exhaustive genome-wide scans for pairwise epistasis in five independent celiac disease (CeD) case-control studies, using a rapid model-free approach to examine over 500 billion SNP pairs in total. We found extensive epistasis within the MHC region with 7,270 statistically significant pairs achieving stringent replication criteria across multiple studies. These robust epistatic pairs partially tagged CeD risk HLA haplotypes, and replicable evidence for epistatic SNPs outside the MHC was not observed. Both within and between European populations, we observed striking consistency of epistatic models and epistatic model distribution, thus providing empirical estimates of their frequencies in a complex disease. Within the UK population, models of CeD comprised of both epistatic and additive single-SNP effects increased explained CeD variance by approximately 1% over those of single SNPs. Further analysis showed that additive SNP effects tag epistatic effects (and vice versa), sometimes involving SNPs separated by a megabase or more. These findings show that the genetic architecture of CeD consists of overlapping additive and epistatic components, indicating that the genetic architecture of CeD, and potentially other common autoimmune diseases, is more complex than previously thought.Author Summary There are few bona fide examples of interactions between genetic variants (epistasis) which affect human disease risk. Here, we assess multiple genome-wide genotyped case-control datasets to investigate the role that epistasis plays in celiac disease, a common immune-mediated illness. We find thousands of replicable, statistically significant pairs of SNPs exhibiting epistasis and, interestingly, all of these fall within the well-known Major Histocompatibility Complex (MHC) region on chromosome 6. We investigate the underlying distribution of epistatic models and further assess the amount of celiac disease variance that can be explained by epistatic pairs, single SNPs and a combination thereof. Our results indicate that there is a substantial amount of shared disease variance between single SNPs and epistatic pairs, but also that a combination of the effects gives a better model of disease. These findings support powerful and routine epistasis scans for the next generation of genome-wide association studies and indicate that the genetic architecture of celiac disease, and potentially other immune-mediated diseases, is more complex than currently appreciated. ER -