Control of population stratification by correlation-selected principal components.

Area of Science:

• Genetics and Bioinformatics
• Statistical Genomics

Background:

• Population stratification is a major confounder in genome-wide association studies (GWAS), leading to inflated Type I error rates.
• Principal component analysis (PCA) of genotype data is a common method to detect and correct for population structure.
• Existing PCA-based methods often rely on eigenvalue significance for component selection, which may not be optimal.

Purpose of the Study:

• To investigate the relationship between genotype principal components and association test statistic inflation.
• To connect PCA-based stratification control with genomic control methods.
• To propose a novel, more effective method for selecting principal components for stratification control.

Main Methods:

• Exploration of the precise relationship between genotype principal components and association test statistic inflation.
• Development and application of the EigenCorr method, which selects principal components based on both eigenvalues and phenotype correlation.
• Analysis of simulated and real genotype data to evaluate EigenCorr's performance.

Main Results:

• The study provides theoretical justification for using principal components in GWAS.
• EigenCorr selects fewer principal components compared to methods relying solely on eigenvalue significance.
• EigenCorr demonstrates improved statistical power and significant computational savings.

Conclusions:

• The proposed EigenCorr method offers a more principled approach to selecting principal components for controlling population stratification in GWAS.
• EigenCorr enhances analytical power and efficiency by optimizing the selection of relevant principal components.
• This approach provides a valuable alternative for robust genetic association studies.