Practical aspects of genome-wide association interaction analysis

Abstract

Large-scale epistasis studies can give new clues to system-level genetic mechanisms and a better understanding of the underlying biology of human complex disease traits. Though many novel methods have been proposed to carry out such studies, so far only a few of them have demonstrated replicable results. Here, we propose a minimal protocol for genome-wide association interaction (GWAI) analysis to identify gene-gene interactions from large-scale genomic data. The different steps of the developed protocol are discussed and motivated, and encompass interaction screening in a hypothesis-free and hypothesis-driven manner. In particular, we examine a wide range of aspects related to epistasis discovery in the context of complex traits in humans, hereby giving practical recommendations for data quality control, variant selection or prioritization strategies and analytic tools, replication and meta-analysis, biological validation of statistical findings and other related aspects. The minimal protocol provides guidelines and attention points for anyone involved in GWAI analysis and aims to enhance the biological relevance of GWAI findings. At the same time, the protocol improves a better assessment of strengths and weaknesses of published GWAI methodologies.

A cautionary note on the impact of protocol changes for genome-wide association SNP × SNP interaction studies: an example on ankylosing spondylitis

“…Any GWAI analysis involves making choices about the input data (e.g., filtering using candidate genes or using prior 393 biological knowledge), about LD-pruning thresholds, about adjusting for lower order effects (and how to encode 394 these), and about the selection of the analytical tool (e.g., non-parametric, semi-parametric or fully parametric), as 395 well as, the corrective method for multiple testing (Gusareva and Van Steen 2014 …”

Performance of Model Based Multifactor Dimensionality Reduction Methods for Epistasis Detection by Controlling Population Structure

“…In addition, we considered the influence of unequal sample sizes. As epistasis analytic tool, we relied on MB-MDR analytics, which should be seen as part of an entire analysis pipeline that involves making marker selection choices and performing post-analysis steps to validate and replicate findings, as well as seeking biological evidence for flagged interacting regions [18].…”

“…However, the growing interest in the importance of detecting gene-gene interactions in the development and progression of complex diseases has led to the development of several tools; to name but a few: generalized linear regression models (GLM), BOOST [8], Model Based Multifactor Dimensionality Reduction (MB-MDR) [9,10], Multifactor Dimensionality Reduction (MDR) [11], Random Forest [12], PLINK [13], BiForce [14], Bayesian Models (e.g., BEAM) [15] and several others. For extensive reviews and appropriate references, please refer to [16][17][18][19][20]. However, the literature on epistasis detection in structured populations is very limited, apart from scenarios using a regression framework for association testing.…”

Giving the Genes a Shuffle: Using Natural Variation to Understand Host Genetic Contributions to Viral Infections

“…Human genetic studies are conducted using either candidate-gene approaches or genomewide association studies (GWASs) [66]. Depending on the research question, both study designs have their own advantages, disadvantages, and limitations.…”