Researchers at the Icahn School of Medicine at Mount Sinai have developed a new artificial intelligence system known as V2P (Variant to Phenotype), which aims to enhance genetic testing by not only identifying harmful genetic mutations but also forecasting the diseases those mutations may cause. The findings are published in the December 15 online issue of Nature Communications.
Traditional genetic analysis tools typically indicate whether a mutation is potentially harmful but do not provide insights into the associated diseases. V2P addresses this gap by utilizing advanced machine learning techniques to link genetic variations to their phenotypic outcomes—essentially predicting how a person's genetic makeup could influence their health. According to David Stein, PhD, the first author, V2P allows for a more focused approach to understanding genetic changes pertinent to individual patients, thereby improving the speed and accuracy of genetic interpretations.
The AI model was trained on a comprehensive dataset that included both harmful and benign genetic variants, alongside detailed disease information. In tests using de-identified patient data, V2P successfully identified true disease-causing mutations within the top ten candidates, suggesting its potential to streamline genetic diagnoses.
Beyond diagnostics, V2P has implications for researchers and drug developers, as it may help identify specific genes and pathways associated with particular diseases. The tool currently categorizes mutations into broad disease groups but could be refined to provide more detailed predictions and integrate additional data for enhanced drug discovery.
The research represents a step forward in precision medicine, where treatments can be tailored based on individual genetic profiles. The study was supported by various grants from the National Institutes of Health and other funding organizations, and the authors include notable researchers in the field of genetics and AI.