GWASBrewer: An R Package for Simulating Realistic GWAS Summary Statistics.

Area of Science:

• Statistical Genetics
• Computational Biology

Background:

• Statistical genetics methods often rely on Genome-Wide Association Studies (GWAS) summary statistics.
• Evaluating these methods requires realistic simulation experiments, but simulating individual-level data is computationally intensive.
• Existing methods for simulating GWAS summary statistics are limited, hindering accurate performance assessment.

Purpose of the Study:

• To introduce GWASBrewer, an open-source R package for direct simulation of GWAS summary statistics.
• To provide a computationally efficient alternative to simulating individual-level data for GWAS analysis.
• To enable comprehensive evaluation of statistical genetics methods through realistic simulations.

Main Methods:

• Developed GWASBrewer, an R package for direct simulation of GWAS summary statistics from their theoretical distribution.
• Validated GWASBrewer by comparing the distribution of simulated statistics to those from individual-level data.
• Implemented functionality to simulate standard error estimates alongside summary statistics.

Main Results:

• GWASBrewer generates summary statistics with distributions identical to those derived from individual-level data.
• Simulations using GWASBrewer are significantly more computationally efficient than those requiring individual-level data.
• The package supports simulation for multiple traits with complex effect size distributions and causal relationships.

Conclusions:

• GWASBrewer offers a flexible and computationally efficient solution for simulating GWAS summary statistics.
• The package facilitates robust evaluation of statistical genetics methods, including Mendelian randomization, polygenic risk scores, and heritability estimation.
• This tool addresses a critical need for comprehensive GWAS summary statistic simulation in the research community.