AI model boosts accuracy in ranking harmful genetic variants
The model identified novel disease genes, offering clinicians clearer guidance in complex unsolved cases.
Researchers have unveiled a new AI model that ranks genetic variants based on their severity. The approach combines deep evolutionary signals with population data to highlight clinically relevant mutations.
The popEVE system integrates protein-scale models with constraints drawn from major genomic databases. Its combined scoring separates harmful missense variants more accurately than leading diagnostic tools.
Clinical tests showed strong performance in developmental disorder cohorts, where damaging mutations clustered clearly. The model also pinpointed likely causal variants in unsolved cases without parental genomes.
Researchers identified hundreds of credible candidate genes with structural and functional support. Findings suggest that AI could accelerate rare disease diagnoses and inform precision counselling worldwide.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
