Epistasis network centrality analysis yields pathway replication across two GWAS cohorts for bipolar disorder

Area of Science:

• Genetics
• Bioinformatics
• Neuroscience

Background:

• Traditional pathway enrichment methods in genome-wide association studies (GWAS) often overlook gene-gene interactions, potentially missing crucial genetic variability.
• A significant portion of heritability for complex diseases like bipolar disorder (BD) may be attributed to these unexamined interactions.

Purpose of the Study:

• To develop and validate a novel method for prioritizing genes in GWAS of BD that integrates gene-gene interaction information with main effect associations.
• To identify novel pathways and genes contributing to BD susceptibility by considering network effects.

Main Methods:

• Utilized machine learning (evaporative cooling) for feature selection and epistasis network centrality analysis to aggregate gene-gene interaction data with main effect associations.
• Employed a two-stage GWAS approach with discovery (Wellcome Trust Case Control Consortium) and replication (National Institute of Mental Health) cohorts of European Ancestry individuals.

Main Results:

• The epistasis network centrality analysis successfully identified replicated enrichment of the Cadherin signaling pathway, previously unhighlighted in BD GWAS.
• Identified other enriched pathways including Wnt signaling, circadian rhythm, axon guidance, and neuroactive ligand-receptor interaction.
• Highlighted the importance of genes ANK3, DGKH, and ODZ4 for BD susceptibility, despite their weak individual effects, underscoring the role of gene networks.

Conclusions:

• The findings suggest that numerous small interactions among common genetic variants contribute to the diathesis for bipolar disorder.
• This study demonstrates the critical importance of incorporating gene-gene interaction network information alongside main genetic effects for robust pathway analysis in GWAS.