Systematic Detection of Epistatic Interactions Based on Allele Pair Frequencies

Epistatic genetic interactions are key for understanding the genetic contribution to complex traits. Epistasis is always defined with respect to some trait such as growth rate or fitness. Whereas most existing epistasis screens explicitly test for a trait, it is also possible to implicitly test for fitness traits by searching for the over- or under-representation of allele pairs in a given population. Such analysis of imbalanced allele pair frequencies of distant loci has not been exploited yet on a genome-wide scale, mostly due to statistical difficulties such as the multiple testing problem. We propose a new approach called Imbalanced Allele Pair frequencies (ImAP) for inferring epistatic interactions that is exclusively based on DNA sequence information. Our approach is based on genome-wide SNP data sampled from a population with known family structure. We make use of genotype information of parent-child trios and inspect 3×3 contingency tables for detecting pairs of alleles from different genomic positions that are over- or under-represented in the population. We also developed a simulation setup which mimics the pedigree structure by simultaneously assuming independence of the markers. When applied to mouse SNP data, our method detected 168 imbalanced allele pairs, which is substantially more than in simulations assuming no interactions. We could validate a significant number of the interactions with external data, and we found that interacting loci are enriched for genes involved in developmental processes. Author Summary: Elucidating non-additive (epistatic) interactions between genes is crucial for understanding the molecular mechanisms of complex diseases. Even though high-throughput, systematic testing of genetic interactions is possible in simple model organisms, such screens have so far not been successful in mammals. Here, we propose a computational screening method that only requires genotype information of family trios for predicting genetic interactions. We tested our framework on a set of more than 2,000 heterozygous mice and found 168 imbalanced allele pairs, which is substantially more than expected by chance. We confirmed many of these interactions using data from recombinant inbred lines. The number of significant allele pair imbalances that we detected is surprisingly large and was not expected based on the published evidence. Our framework sets the stage for similar work in human trios.