EpiGEN

EpiGEN simulates epistasis in genome-wide association studies (GWAS) to generate synthetic genetic and phenotypic data for evaluation of epistasis detection methods and modeling complex genetic architectures.

Key Features:

• Arbitrary-order SNP interactions: Supports simulation of SNP interactions of arbitrary order to model complex genetic architectures and epistasis.
• Realistic linkage disequilibrium (LD) patterns: Generates realistic LD patterns to preserve correlation structure between SNPs that affects interaction detection.
• Categorical and quantitative phenotypes: Produces both categorical (binary) and quantitative phenotypes for diverse trait simulations.
• Specification of minor allele frequencies (MAFs): Allows users to specify MAFs for both noise and disease-associated SNPs.
• Biased target phenotype distributions: Enables specification of biased target distributions for generated phenotypes to simulate observation bias.
• Implementation: Implemented in Python 3.

Scientific Applications:

• Benchmarking epistasis detection: Provides synthetic GWAS datasets for evaluating performance of epistasis detection algorithms.
• Modeling complex genetic architectures: Enables study of high-order SNP interactions and their effects on traits.
• Assessing LD effects: Allows investigation of how realistic LD patterns influence detection of genetic interactions.
• Evaluating bias and MAF effects: Facilitates simulation of observation bias and controlled MAF scenarios to study their impact on analysis outcomes.

Methodology:

Implemented in Python 3; simulates SNP interactions of arbitrary order, generates realistic linkage disequilibrium patterns, produces categorical and quantitative phenotypes, and permits specification of MAFs for noise and disease-associated SNPs as well as biased target phenotype distributions.