Researchers at the Icahn School of Medicine at Mount Sinai have developed a method that combines artificial intelligence (AI) with routine laboratory tests to evaluate disease risk in individuals with rare genetic variants. Traditional genetic studies often provide only binary outcomes, indicating whether a variant is harmful or not. This approach integrates AI models with electronic health records to produce a continuous risk score, offering a more nuanced view of genetic risk.
The method relies on commonly collected lab data, including cholesterol levels, blood counts, and other routine measures, to estimate the likelihood of disease development. Researchers analysed more than one million electronic health records to train AI models for 10 common diseases. These models were then applied to individuals carrying rare genetic variants. Each individual receives a score ranging from 0 to 1, representing the probability of developing the disease. By combining AI and lab tests, the method allows researchers and clinicians to evaluate the penetrance of rare variants in a way that was previously difficult. This is particularly relevant for conditions that do not present clear clinical signs, or where rare variants’ effects are not well understood. The approach could inform clinical decision-making, helping to identify patients who may benefit from closer monitoring or preventative measures.
The study, led by researchers at Mount Sinai, was published in the journal Science under the title “Machine learning-based penetrance of genetic variants.” It demonstrates how AI and existing medical data can be leveraged to better understand individual disease risk and refine genetic insights for clinical use.
The article can be accesed on: MedicalXpress
