AI Tool Simultaneously Identifies Genetic Mutations and Disease Type

While existing tools can indicate whether a genetic mutation is harmful, they usually cannot determine what type of disease that mutation may cause. This forces clinicians to sift through thousands of variants, slowing diagnosis and delaying treatment decisions. Researchers have now developed an artificial intelligence (AI)–based method that not only flags disease-causing mutations but also predicts the category of disease they are likely to trigger, significantly improving the speed and clarity of genetic interpretation.

The AI framework known as Variant to Phenotype (V2P) has been designed by researchers at the Icahn School of Medicine at Mount Sinai (New York, NY, USA) to link genetic variants directly to disease outcomes. The system uses advanced machine learning models trained on large datasets of pathogenic and benign genetic variants, combined with disease annotations. By integrating genetic and phenotypic information, the tool predicts not just whether a mutation is harmful, but also the type of disease it is most likely to cause, such as cancer or nervous system disorders.

V2P was evaluated using large-scale genomic datasets and tested on real, de-identified patient data. In these evaluations, the tool consistently ranked the true disease-causing mutation among the top 10 candidate variants, dramatically narrowing the search space for clinicians. The findings, published in Nature Communications, demonstrate that incorporating phenotype prediction substantially improves the accuracy and efficiency of genetic diagnostics compared to traditional variant-ranking methods.

By linking mutations to disease categories, V2P has the potential to accelerate diagnosis for patients with rare and complex conditions and reduce uncertainty in clinical genetics. The approach may also help researchers identify disease-relevant genes and biological pathways, supporting more targeted drug discovery. The researchers plan to refine the system to predict more specific disease outcomes and integrate additional biological data sources, expanding its usefulness for precision medicine and genetically informed therapy development.

"Our approach allows us to pinpoint the genetic changes that are most relevant to a patient’s condition, rather than sifting through thousands of possible variants," said first author David Stein, PhD. "By determining not only whether a variant is pathogenic but also the type of disease it is likely to cause, we can improve both the speed and accuracy of genetic interpretation and diagnostics."

Related Links:
Icahn School of Medicine at Mount Sinai