Machine learning tool trained on functional assays improves missense variant prediction
Researchers at The Rockefeller University developed FuncVEP, a machine learning tool that predicts the impact of missense variants using functional assays instead of clinical or population data. The predictor achieved 84.6% accuracy on functional benchmarks and 92.4% on clinical benchmarks, outperforming 48 existing tools. The team, led by researchers including Kayaalp, Çil, and Casanova, published their findings on January 14, 2025. When applied to the UK Biobank and Mount Sinai Million Health Discoveries Program cohorts, FuncVEP identified 210 new gene-phenotype associations involving 494 genes linked to inborn errors of immunity. The tool improved the discovery rate compared to current state-of-the-art predictors.
FuncVEP may reduce diagnostic uncertainty for patients with rare genetic variants by relying on experimental evidence instead of indirect clinical patterns, addressing a critical bottleneck in research and diagnostic genetics.
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].