Detecting Genetic Interactions with Visible Neural Networks

Non-linear interactions among single nucleotide polymorphisms (SNPs), genes, and pathways play an important role in human diseases, but identifying these interactions is a challenging task. Neural networks are state-of-the-art predictors in many domains due to their ability to analyze big data and model complex patterns, including non-linear interactions. In genetics, visible neural networks are gaining popularity as they provide insight into the most important SNPs, genes and pathways for prediction. Visible neural networks use prior knowledge (e.g. gene and pathway annotations) to define the connections between nodes in the network, making them sparse and interpretable. Currently, most of these networks provide measures for the importance of SNPs, genes, and pathways but lack details on the nature of the interactions. In this paper, we explore different methods to detect non-linear interactions with visible neural networks. We adapted and sped up existing methods, created a comprehensive benchmark with simulated data from GAMETES and EpiGEN, and demonstrated that these methods can extract multiple types of interactions from trained visible neural networks. Finally, we applied these methods to a genome-wide case-control study of inflammatory bowel disease and found high consistency of the epistasis pairs candidates between the interpretation methods. The follow-up association test on these candidate pairs identified seven significant epistasis pairs.