PUMAS

PUMAS fine-tunes polygenic risk score (PRS) models using genome-wide association study (GWAS) summary statistics to optimize predictive accuracy and increase statistical power in downstream association analyses.

Key Features:

• Model-Tuning Procedures: Performs model-tuning procedures using GWAS summary statistics to refine PRS parameters without individual-level genotype data.
• Benchmarking and Optimization: Benchmarks existing PRS models and optimizes them under diverse genetic architectures to tailor scores to trait-specific architectures.
• Improved Statistical Power: Enhances statistical power in downstream association analyses by fine-tuning PRS models.
• Broad Applicability: Validated via extensive simulations and external validations across 65 different traits, demonstrating applicability across diverse human genetics settings.

Scientific Applications:

• PRS model optimization from summary statistics: Enables tuning of PRSs when only GWAS summary statistics are available.
• Association analysis enhancement: Increases power to detect genetic associations in downstream association studies using optimized PRSs.
• Cross-trait and trait-specific evaluation: Provides a framework for evaluating and tailoring PRSs across a wide range of traits, as demonstrated across 65 different traits.

Methodology:

Uses GWAS summary statistics for PRS parameter tuning and evaluates performance via extensive simulations and external validations across 65 different traits.