Polygenic risk score prediction accuracy convergence.

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Polygenic risk scores (PRSs) derived from genome-wide association studies (GWAS) are crucial for studying multifactorial diseases.
• While promising for clinical applications, current PRS performance is limited, with ongoing debate regarding their strengths and weaknesses.

Purpose of the Study:

• To retrospectively assess the progress of PRS prediction accuracy since the advent of large-scale GWAS.
• To investigate factors influencing maximal prediction accuracy using whole-genome sequencing data and advanced modeling techniques.

Main Methods:

• Conducted a retrospective analysis of PRS prediction accuracy for six common diseases using GWAS data.
• Utilized whole-genome sequencing data from 125,000 UK Biobank participants for advanced polygenic outcome modeling.

Main Results:

• PRS accuracy has increased significantly over time, but recent GWAS show diminishing returns in accuracy improvement.
• Merely expanding GWAS sample sizes may yield only marginal enhancements in risk discrimination.
• Increasing variant coverage through imputation or sequencing data is critical for enhancing PRS prediction.

Conclusions:

• Future improvements in PRS accuracy for disease risk prediction will likely depend more on increasing genetic variant coverage than solely on larger GWAS sample sizes.
• Whole-genome sequencing data holds significant potential for advancing PRS predictive power.