Control for population stratification in genetic association studies based on GWAS summary statistics

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) provide valuable genetic insights but are susceptible to population stratification.
• Sharing individual-level genotype and phenotype data for GWAS is challenging, necessitating methods using only summary statistics.
• Population stratification and cryptic relatedness can confound GWAS results, leading to inflated Type I errors.

Purpose of the Study:

• To compare the performance of Genomic Control (GC) and Linkage Disequilibrium Score Regression (LDSC) for controlling population stratification using GWAS summary statistics.
• To evaluate these methods in diverse population structures, including subpopulations, spatial structuring, and cryptic relatedness.
• To determine the efficacy of LDSC's intercept as a correction factor versus GC.

Main Methods:

• Extensive simulation studies were conducted with various population structures and relatedness.
• Real-world data from the Genetic Analysis Workshop 19 and UK Biobank were utilized for evaluation.
• Performance comparison focused on controlling population stratification without individual-level genotype data.

Main Results:

• The intercept derived from LDSC was found to be a more accurate correction factor for population stratification than GC.
• Both methods were evaluated across simulated populations with varying complexities and real datasets.
• LDSC demonstrated superior performance in mitigating confounding effects from population structure.

Conclusions:

• LDSC provides a more reliable approach for correcting population stratification in GWAS using summary statistics.
• Researchers can confidently use LDSC's intercept for accurate association testing in the presence of population structure.
• This study offers crucial guidance for leveraging GWAS summary statistics while controlling for confounding factors.