WISH – Weighted Interaction SNP Hub R Package

Genome-wide association studies (GWAS) provided wealth of information on the link between genotypes and phenotypes but they do not capture the entire genetic variation that exists in complex traits and diseases. One of the reasons for the lack of power of GWAS is that it does not account for genome-wide epistasis. Biologically, genes do not exist in isolation, but interact with each other in complex biological networks to give rise to complex traits. The quantification of epistatic interaction between genomic variants is however a difficult problem, as calculating pairwise interactions on full genomic data sets requires an intractable amount of calculations even with modern computing facilities. To solve this, we have implemented the Weighted Interactions SNP Hub (WISH) method as an R package, which relates epistatic interactions networks with complex traits. The inputs are standard pedigree and transposed pedigree files and a phenotype. The method first calculates individual pairwise epistatic interactions using efficient parallelized linear modelling based on the genomic data, resulting in epistatic correlations for all genotype pairs. Epistatic correlations are then applied in scale-free weighted genetic interaction networks to identify biologically relevant pathways and markers associated with complex traits, also alleviating multiple testing issues. We reduce data dimension based on variant main effects and genetic variation. In relation to other methods our implementation is efficient, accessible and fast, with focus on larger datasets. The package is freely available for download from GitHub and runs on a regular computer, but is scalable for server use with larger data sets.