GEPSi

GEPSi simulates complex phenotypes from genotype data to produce biologically realistic datasets for benchmarking and development of Genome-Wide Association Study (GWAS) methods.

Key Features:

• Biological complexity modeling: Incorporates heritability, dominance, population stratification, and epistatic interactions between SNPs into phenotype simulations.
• Non-random causal SNPs: Models non-random occurrence of causal SNPs when assigning genetic effects.
• Epistatic interaction modeling: Accounts for epistatic interactions between SNPs in generating phenotypes.
• Parameter customization: Allows adjustment of parameters such as heritability levels and epistatic effect magnitudes to explore different scenarios.
• Genotype-based simulation: Uses user-supplied genotype data as the input basis for phenotype generation.
• Benchmarking dataset generation: Produces simulated datasets suitable for evaluating GWAS methods and machine learning algorithms.

Scientific Applications:

• GWAS method development and validation: Provides realistic phenotype datasets to validate and refine GWAS analytical approaches.
• Machine learning method comparison: Facilitates comparative evaluation of machine learning and other analytical techniques using consistent biologically plausible data.
• Genetic architecture simulation studies: Enables investigation of how heritability, dominance, population stratification, and epistasis influence trait variation.

Methodology:

GEPSi uses user-supplied genotype data and specified biological parameters (heritability, dominance, population stratification, epistatic interactions) to generate phenotype simulations, with an algorithmic framework that models non-random selection of causal SNPs and epistatic interactions.