A team of scientists in New York has developed a new model designed to help doctors interpret complex genetic test results and provide patients with clearer guidance on their health risks. The research, published in the journal Science, could improve early detection of serious conditions and reduce unnecessary medical treatments.
Genetic testing can identify changes, or variants, in a person’s DNA, but interpreting the results is often difficult. While some variants directly cause disease, many others have little or no effect, leaving doctors and patients uncertain about what the findings mean. The problem is compounded by the fact that most diseases result not from a single mutation, but from the combined influence of multiple genes and environmental factors.
To address this challenge, researchers at the Icahn School of Medicine at Mount Sinai built a model that draws on both genetic information and electronic health records (EHRs), which include lab results and a patient’s medical history. By combining these data sources, the model can calculate the likelihood that an individual with a specific variant will develop conditions such as breast cancer or polycystic kidney disease.
“Traditional genetic tests often leave patients in limbo, because the results don’t always provide a clear answer,” said Professor Ron Do, one of the study’s senior authors. “By using real-world medical data—like cholesterol levels and blood counts that are already part of routine care—we can make far more accurate predictions about disease risk.”
The researchers trained the model on more than one million health records and applied it to patients carrying rare genetic mutations. Each patient was assigned a risk score between zero and one, reflecting the probability of developing a particular condition. In total, the team calculated risk scores for more than 1,600 genetic variants.
In some cases, the tool clarified the significance of variants previously labelled as “uncertain.” For example, the model revealed strong links between specific mutations and known diseases, providing new insights for clinicians.
Dr. Iain Forrest, the study’s lead author, said the tool is intended to support, not replace, doctors. “This model could guide decisions on whether a patient needs further screening, preventive steps, or reassurance that their genetic result poses little risk,” he explained.
The team is now working to expand the model by including a wider range of genetic variants, more diseases, and a more diverse patient population to ensure broader accuracy.
“Ultimately, our work highlights a future where clinical data and genetic information can be combined to give patients more personalised and actionable answers,” Do said.
If widely adopted, the approach could change the way genetic testing is used in medicine—helping patients avoid unnecessary interventions while ensuring those at higher risk receive timely care.
