Admixture-informed polygenic risk reporting using the ePRS framework.

Area of Science:

• Genetics
• Population Genetics
• Statistical Genetics

Background:

• Polygenic risk scores (PRS) exhibit variability across different genetic ancestries due to population-specific allele frequencies and linkage disequilibrium.
• Existing PRS methods often struggle with population stratification, leading to biased association estimates.

Purpose of the Study:

• To introduce a novel framework for calibrating polygenic risk scores based on an individual's ancestral makeup.
• To define and validate the expected polygenic risk score (ePRS) and residual polygenic risk score (rPRS) for improved PRS interpretation.

Main Methods:

• Developed the expected polygenic risk score (ePRS) as the expected PRS value based on global or local admixture patterns.
• Defined the residual polygenic risk score (rPRS) as the deviation of PRS from ePRS, representing an ancestry-agnostic genetic liability.
• Validated the framework using simulation studies and real-world datasets (TOPMed and All of Us).

Main Results:

• Simulation studies demonstrated that adjusting for ePRS yields unbiased polygenic risk score-outcome association estimates without needing principal components.
• Effect size estimates for rPRS (adjusted for ePRS) were comparable to traditional PRS methods that adjust for genetic principal components.
• The ePRS framework effectively mitigates population stratification in association analyses.

Conclusions:

• The ePRS framework provides a robust method for calibrating polygenic risk scores across diverse populations.
• This approach enhances the equitable interpretation of genetic risk by separating ancestry-driven components from ancestry-agnostic liability.
• The ePRS framework offers a promising strategy for accurate and equitable genetic risk assessment in population genetics and precision medicine.