ZMIX: estimating ancestry proportions using GWAS association Z-scores

Area of Science:

• Genetics
• Bioinformatics
• Population Genetics

Background:

• Accurate estimation of ancestral proportions is crucial for genome-wide association studies (GWAS).
• Existing methods often require reference allele frequencies (AFs), which are frequently inaccessible due to privacy concerns in large-scale GWAS.
• This limitation hinders downstream analyses like imputation and variant prioritization.

Purpose of the Study:

• To introduce ZMIX, a novel method for estimating ethnic mixing proportions in GWAS.
• To evaluate ZMIX's performance against established reference AF-based methods.
• To demonstrate ZMIX's utility in summary-statistics imputation.

Main Methods:

• Developed ZMIX, a method utilizing only association Z-scores for estimating ethnic mixing proportions.
• Compared ZMIX with reference AF-based methods using simulated and real-world GWAS data.
• Assessed the performance of ZMIX and other methods in summary-statistics imputation.

Main Results:

• ZMIX demonstrated comparable performance to reference AF-based methods in both simulated and real-world GWAS.
• All tested methods, including ZMIX, yielded high-quality imputations with nearly identical results when applied to summary statistics.
• The study validates ZMIX as a viable alternative when reference AFs are unavailable.

Conclusions:

• ZMIX offers a robust and accessible method for estimating ethnic mixing proportions in GWAS.
• The method overcomes the limitations of reference AF dependency, enhancing the applicability of GWAS.
• ZMIX facilitates accurate downstream analyses, including imputation and variant prioritization, in diverse genomic studies.