XPXP

XPXP improves polygenic risk score (PRS) prediction by integrating cross-population and cross-phenotype GWAS and biobank data to enhance predictive accuracy across diverse populations.

Key Features:

• Cross-Population Analysis: Utilizes data from diverse populations, including European, East Asian, and African groups, to improve PRS performance in under-represented populations.
• Cross-Phenotype Integration: Incorporates genetic correlations across multiple phenotypes to refine PRS estimates.
• Biobank and GWAS Integration: Leverages European biobank-scale datasets alongside multiple GWASs that are genetically correlated with target phenotypes.
• Population- and Phenotype-Specific Effects: Explicitly models and incorporates both population-specific and phenotype-specific genetic effects into PRS construction.
• Enhanced Accuracy and Validation: Demonstrated via simulation studies and real-world analyses, showing a 9% improvement in predicted R² for height PRSs in East Asian populations and an 18% improvement in African populations.
• Disease Risk Stratification: Improves identification of individuals at high genetic risk for conditions such as type 2 diabetes.

Scientific Applications:

• Personalized Medicine: Enables more accurate disease risk prediction and stratification across diverse population groups for applications such as prevention and targeted intervention.
• Genetic Research: Provides a framework to study genetic correlations across phenotypes and to transfer PRS information across populations in genetic epidemiology studies.

Methodology:

XPXP integrates European biobank-scale data with multiple genetically correlated GWASs to construct PRSs that account for population-specific and phenotype-specific effects.