Multi-Ancestry Polygenic Risk Score for Coronary Heart Disease Based on an Ancestrally Diverse Genome-Wide...

Area of Science:

• Genetics
• Genomic Medicine
• Population Health

Background:

• Polygenic risk scores (PRS) for coronary heart disease (CHD) exhibit variable predictive performance across different global populations.
• Developing equitable PRS for clinical use necessitates accounting for genetic ancestry variations.

Purpose of the Study:

• To develop and evaluate ancestry-specific and multi-ancestry PRS for coronary heart disease (CHD).
• To compare the predictive performance of different PRS methodologies across diverse genetic ancestry groups.

Main Methods:

• Derived ancestry-specific and multi-ancestry PRS for CHD using pruning and thresholding (PRS_PT) and continuous shrinkage priors (PRS_CSx).
• Utilized summary statistics from a large multi-ancestry genome-wide association study meta-analysis (1.1 million participants).
• Trained and optimized PRS in the Million Veteran Program, then validated in 9 diverse cohorts (176,988 individuals).

Main Results:

• Multi-ancestry PRS_PT and PRS_CSx consistently outperformed ancestry-specific PRS across various tuning parameters.
• The best-performing multi-ancestry PRS (PRS_PTmult and PRS_CSxmult) showed significant CHD associations across South Asian, European, East Asian, Hispanic/Latino, and African ancestries.
• PRS_PTmult demonstrated the strongest CHD association in South Asian and European ancestry individuals.

Conclusions:

• Leveraging multi-ancestry genome-wide meta-analysis summary statistics enhances PRS_CHD performance in most populations compared to single-ancestry methods.
• Predictive performance improvements were limited in individuals of African ancestry, underscoring the need for larger, diverse genomic datasets.
• Increased representation in genome-wide association studies is crucial for improving PRS_CHD equity and accuracy across all populations.