POLARIS: Polygenic LD-adjusted risk score approach for set-based analysis of GWAS data

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Polygenic risk scores (PRSs) summarize genetic variance from SNPs.
• PRSs can improve the power of set-based analyses by using public genome-wide association study (GWAS) data.
• Assessing genetic liability for phenotypes often relies on PRSs derived from independent datasets.

Purpose of the Study:

• To introduce POLARIS (POlygenic Ld-Adjusted RIsk Score), a novel set-based method for genetic analysis.
• To enhance PRS application by incorporating linkage disequilibrium (LD) adjustment.
• To extend PRS methods for analyzing biologically relevant SNP sets.

Main Methods:

• POLARIS utilizes spectral decomposition of the SNP correlation matrix to identify LD structure.
• It replaces individual SNP allele counts with LD-adjusted dosages.
• The method uses raw genotype data and effect sizes from an independent dataset for set-based analysis.

Main Results:

• Simulations demonstrated POLARIS's power compared to MAGMA, another LD-aware set-based method.
• POLARIS showed increased power over MAGMA when using only raw genotype data.
• POLARIS achieved comparable or less power than combined analysis of both datasets.

Conclusions:

• POLARIS offers advantages by generating a per-person, per-set risk score using all available SNPs.
• The method leverages discovery set effect sizes within a self-contained test for association in the test dataset.
• POLARIS provides a powerful and flexible tool for set-based genetic association studies.