Stanford Team Builds 813 Polygenic Risk Scores from UK Biobank Data
Researchers at Stanford University developed 813 polygenic risk score models across more than 1,500 traits using UK Biobank genetic and phenotype data, according to a 2022 study in PLOS Genetics. The sparse models showed significant predictive performance (p < 2.5 x 10-5) when compared against covariate-only baselines that accounted for age, sex, genotyping array type, and principal component loadings. A correlation emerged between the number of genetic variants selected and incremental predictive performance: Spearman's ⍴ = 0.61 (p = 2.2 x 10-59) for quantitative traits and ⍴ = 0.21 (p = 9.6 x 10-4) for binary traits. Models trained on European individuals showed limited transferability when tested on non-European individuals in the UK Biobank. The PRS model weights are available on the Global Biobank Engine at https://biobankengine.stanford.edu/prs.
The work provides public PRS models but exposes the persistent challenge of cross-ancestry transferability in genetic prediction.
Written by the Genomes desk from the primary source linked above and checked against it. Research use only; not medical advice. Corrections: [email protected].