TraitSimulation.jl

TraitSimulation.jl simulates phenotypes under diverse genetic architectures for unrelated and familial study designs to enable realistic modeling of trait distributions (including non-normal and qualitative traits) for genetic studies.

Key Features:

• Diverse trait distributions: Supports simulation of non-normal continuous traits and qualitative traits beyond Gaussian or transformable-to-normal models.
• Statistical models: Implements generalized linear models (GLMs) and generalized linear mixed models (GLMMs) for phenotype simulation.
• Study designs: Accommodates unrelated individuals, sibships, pedigrees, and combinations of these designs.
• Genetic dependency modeling: Accounts for pedigree structure and cryptic relatedness when simulating genetic dependencies.
• Integration with OpenMendel: Interoperates with the OpenMendel suite for downstream genetic analyses.
• Julia implementation: Built in Julia to leverage language features for computational speed and memory efficiency.
• Parallelization and scalability: Supports parallel execution across multi-CPU and GPU architectures and deployment in cloud environments for large-scale simulations.
• Realistic phenotype modeling: Enables simulation strategies intended to increase realism of power calculations and diagnostic evaluations.

Scientific Applications:

• Power calculations: Generating realistic phenotypes to estimate statistical power under complex trait models.
• Method benchmarking: Benchmarking association and inference methods using non-normal and qualitative trait simulations.
• Family-based analyses: Simulating pedigrees and sibships to evaluate family-based genetic tests and estimators.
• Cryptic relatedness assessment: Assessing effects of cryptic relationships on association results and corrections.
• Statistical genetics research: Modeling complex genetic architectures for development and validation of analysis methods.

Methodology:

Simulations use generalized linear models (GLMs) and generalized linear mixed models (GLMMs), explicitly model pedigree- and cryptic-relationship genetic dependencies, and exploit Julia-based implementations with parallelization on multi-CPU and GPU architectures and cloud environments.