CCAFE: Estimating Case and Control Allele Frequencies from GWAS Summary Statistics

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Summary statistics from genetic studies are valuable but often lack case and control allele frequencies (AFs).
• Post-hoc analyses of genome-wide association studies (GWAS) frequently require these specific AFs, limiting data reusability.

Purpose of the Study:

• To present two novel frameworks for deriving case and control AFs from existing GWAS summary statistics.
• To enhance the utility and reusability of publicly available genetic data.

Main Methods:

• Developed two frameworks utilizing odds ratios, sample sizes, and either total AF or standard error (SE) from GWAS summary statistics.
• Validated methods through simulations and real-world genetic data analysis.
• Incorporated an adjustment using gnomAD AFs to mitigate bias when deriving AFs from SE.

Main Results:

• Derivations using total AF demonstrated high accuracy across various genetic settings.
• Derivations using SE showed underestimation of common variant AFs (>0.3) when covariates were present.
• The gnomAD AF adjustment effectively reduced bias in SE-based derivations.

Conclusions:

• Estimating case and control AFs using total AF is highly accurate and preferred.
• Estimating from SE offers broader applicability as SE is derivable from commonly reported p-values and beta estimates.
• The developed methods and the accompanying R package CCAFE expand the utility of GWAS summary statistics.