Epistasis Detection in Genome-Wide Screening for Complex Human Diseases in Structured Populations

Citations Over TimeTop 16% of 2019 papers

Abstract

Over the years, a more prominent role has been given to gene–gene interaction (epistasis) detection, in view of precision medicine and the hunt for novel drug targets and biomarkers for complex diseases. Acknowledging data complexity as embodied by epistasis potentially increases the power of genome-wide association studies (GWAS) and may reveal relevant biological and biochemical pathways previously undetected. Although confounding of GWAS due to shared genetic ancestry has been well recognized, the extent and impact of such confounding in gene–gene interaction association epistasis studies is much less understood. In the same spirit, the role of population substructure in epistasis detection is largely under-investigated, especially outside a regression framework. This is surprising as inadequate handling of such confounding is likely to lead to spurious epistatic associations, hampering replication and the identification of causal loci in synergy. To improve interpretability and replicability of epistasis results, we introduce “MB-MDR for Structured Populations” (Model-Based Multifactor Dimensionality Reduction-Principal Component [MBMDR-PC]). It extends classic MBMDR that was developed for samples sharing the same genetic ancestry, by replacing the original phenotypes with new phenotypes that have been adjusted for population structure as captured by PCs. The method is applied on International Inflammatory Bowel Disease Genetics Consortium Crohn's disease (CD) data from 15 countries. Significant interacting single nucleotide polymorphisms are found within NOD2 and CYLD genes among others that suggest a potential synergetic effect of NOD2 and CYLD on CD. The study highlights the value of examining epistasis effects and indicates the need for further studies to understand the epistasis effects on CD.

Related Papers

• → Epistasis dominates the genetic architecture of Drosophila quantitative traits(2012)383 cited
• → Detection of QTL for flowering time in multiple families of elite maize(2012)45 cited
• → Problems and Pit-Falls in Testing for G × E and Epistasis in Candidate Gene Studies of Human Behavior(2014)14 cited
• → Assessment of parametric and non-parametric methods for prediction of quantitative traits with non-additive genetic architecture(2020)4 cited
• → Capacitating Epistasis—Detection and Role in the Genetic Architecture of Complex Traits(2014)2 cited