Detecting genetic interactions in pathway-based genome-wide association studies

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Pathway-based genome-wide association studies (GWAS) enhance detection power by analyzing collective effects of causal variants.
• Existing pathway-based GWAS methods often overlook epistatic effects (interactions between genetic variants), which are crucial for complex traits.

Purpose of the Study:

• To develop a novel pathway-based GWAS method incorporating both main and pairwise epistatic effects of genetic variants.
• To improve the detection of genetic associations in complex diseases by accounting for gene-gene interactions.

Main Methods:

• Employed a Bayesian Lasso logistic regression model for pathway-based GWAS.
• Utilized an efficient group empirical Bayesian Lasso (EBLasso) method for model inference.
• Applied Wald statistics for pathway significance testing and stability selection for effect identification.

Main Results:

• The group EBLasso method demonstrated superior performance compared to two competitive methods in extensive computer simulations.
• Application to a Parkinson disease GWAS dataset identified three significant pathways: primary bile acid biosynthesis, neuroactive ligand-receptor interaction, and MAPK signaling.
• A significant portion of the identified effects, particularly in the primary bile acid biosynthesis pathway, were epistatic.

Conclusions:

• The group EBLasso method is a valuable tool for pathway-based GWAS, capable of identifying both main and epistatic genetic effects.
• The findings underscore the importance of considering gene-gene interactions in understanding complex traits and diseases like Parkinson's.