network_epistasis

network_epistasis detects epistatic interactions between single nucleotide polymorphisms (SNPs) along edges of gene-gene co-function networks to enable gene-level discovery of interactions underlying complex disease biology.

Key Features:

• Gene-Gene Co-Function Network Integration: Maps epistatic tests to edges of a gene-gene co-function network to reduce the number of statistical tests and maintain control over type I error rates.
• SNP-Gene Mapping Strategies: Implements positional overlap, expression quantitative trait loci (eQTL), and 3D chromatin proximity mappings to translate gene interactions into SNP-SNP hypotheses.
• Adaptive Truncated Product Method: Aggregates SNP pair association p-values into gene pair association scores using a non-parametric adaptive truncated product method without requiring a predefined null distribution.
• Application to Genome-Wide Association Studies (GWAS): Applied to GWAS datasets—including inflammatory bowel disease (IBD)—to identify diverse mechanisms implicated in complex diseases.
• Incorporation of Biological Knowledge: Integrates prior biological network information into epistasis detection to enhance interpretability and discovery compared with non-network methods.

Scientific Applications:

• Complex Disease Mechanism Elucidation: Identifies gene-gene interactions contributing to multifactorial diseases to inform biological mechanisms.
• Enhanced Interpretability and Discovery Potential: Leverages networks and mapping strategies to reduce statistical complexity and prioritize biologically plausible SNP-SNP interactions.
• GWAS Analysis: Supports systematic interrogation of epistasis in genome-wide datasets such as IBD studies.

Methodology:

Gene interactions are mapped to SNP pairs using positional overlap, eQTL mappings, and 3D chromatin proximity, and gene pair association scores are computed from SNP pair data using the adaptive truncated product method (a non-parametric, computationally efficient approach).