New framework squeezes more accuracy from existing genetic risk models
Researchers built Adaptive Boosting of pre-trained Polygenic Risk Scores, a framework that refines existing polygenic risk score models to pull out predictive signals the original models miss. The team developed the method using UK Biobank data. They tested it on binary diseases and continuous traits across three independent datasets: All of Us, eMERGE, and Penn Medicine Biobank. Simulations showed the framework can identify signals orthogonal to pre-trained polygenic risk scores while controlling false discovery rates. In real-world data, the approach produced statistically significant improvements in several scenarios and remained competitive in others, the team reported in Nature Communications.
The framework could improve genetic risk prediction without building new models from scratch. It works with pre-trained models already in use.
Written by the Genomes desk from the primary source cited and linked above and checked against it. Research use only; not medical advice. Corrections: [email protected].