Methodologies underpinning polygenic risk scores estimation: a comprehensive overview

Area of Science:

• Genomics
• Personalized Medicine
• Population Genetics

Background:

• Polygenic risk scores (PRS) are developed using genome-wide association studies (GWAS), predominantly from European ancestry populations.
• Existing PRS models show limited validation and potential bias in non-European populations, impacting clinical utility.
• There's a critical need for risk prediction frameworks that are inclusive of diverse genetic backgrounds and consider complex interactions.

Purpose of the Study:

• To review and analyze the strengths and limitations of various polygenic risk score (PRS) construction methods.
• To highlight recent advancements in PRS calculation methodologies.
• To identify future research directions for developing robust PRS across diverse populations.

Main Methods:

• Analysis of traditional weighted PRS construction methods.
• Evaluation of novel Bayesian and Frequentist penalized regression approaches for PRS.
• Review of current literature on PRS development and application across different ancestries.

Main Results:

• Traditional PRS methods have limitations when applied to diverse populations.
• Newer penalized regression methods offer potential improvements in PRS construction.
• The development of PRS is complex, influenced by genetic variants across various ancestral backgrounds.

Conclusions:

• PRS have significant potential for enhancing disease risk prediction and personalized medicine.
• Further research is essential to create PRS models that are accurate and equitable across diverse populations.
• Ethical considerations, including bias and fairness, must guide the implementation of PRS.