Researchers from the Icahn School of Medicine at Mount Sinai have developed an innovative artificial intelligence system capable not only of detecting harmful genetic mutations but also of predicting which specific diseases these mutations may cause.

A New Approach to Genetic Data Analysis

The method, named V2P (Variant to Phenotype), was created to accelerate genetic testing and support the development of new treatments for rare and complex diseases. The V2P system uses advanced machine learning algorithms to establish connections between genetic variants and their possible phenotypic manifestations—that is, the diseases or symptoms that a mutation may trigger. This enables predictions about how changes in DNA can impact human health.

Advantages of V2P for Diagnosis and Therapy

As David Stein, who recently completed his training in the labs of Yuval Itan and Avner Schlessinger, notes, this new approach allows for the identification of the genetic changes most relevant to a patient's condition, eliminating the need to sift through thousands of possible variants. By determining not only the pathogenicity of a variant but also the type of disease it may cause, the system increases the speed and accuracy of genetic interpretation and diagnosis.

Training and Testing the System

To build the model, researchers trained V2P on a large dataset that included both harmful and benign genetic variants, as well as detailed information about diseases. This allowed the system to identify patterns linking specific variants to health outcomes. When tested on real, anonymized patient data, V2P frequently identified the true disease-causing mutation among the ten most likely candidates, demonstrating the system’s potential to simplify and speed up genetic diagnostics.

Implications for Science and Medicine

Professor Avner Schlessinger emphasizes that beyond diagnostics, V2P can help researchers and drug developers identify genes and pathways most closely associated with particular diseases. This opens up opportunities for creating genetically tailored therapies, especially for rare and complex conditions.

Currently, V2P sorts mutations into broad disease categories, such as neurological disorders or cancers. The team plans to further refine the system so it can make more detailed predictions and integrate its results with additional data sources to better support drug development.

A Step Toward Personalized Medicine

Researchers believe that this progress brings medicine closer to a personalized approach, where treatment is selected based on an individual’s unique genetic profile. By linking genetic variants to their likely health consequences, V2P can help doctors make faster diagnoses and enable scientists to discover new therapeutic targets.

Yuval Itan notes that V2P provides a clearer understanding of how genetic changes lead to diseases, which is crucial for both research and patient care. By connecting specific variants to the types of diseases they may cause, the system helps identify which genes and pathways require deeper study, facilitating the transition from understanding biology to finding potential therapeutic strategies and, ultimately, to individualized treatment.