Researchers use ChatGPT-style technology to predict stroke risk from a person’s DNA

Researchers have developed an AI model that analyses a person’s DNA to catch genetic risk factors for stroke — the second leading cause of death worldwide. ADVERTISEMENT ADVERTISEMENT A team at the University of Technology Sydney (UTS) has identified specific genes that show a higher predisposition to ischaemic stroke — the most common type of stroke, in which a blood clot blocks blood flow to the brain. “When someone has a stroke, particularly at a young age, they often want to know whether other family members are also at risk,” said Kairui Guo, the study’s corresponding author. “These new AI tools will provide the ability to better help families understand that risk.” Genetics account for roughly 38% of an individual's stroke risk, according to previous research. “Traditionally, in thinking about stroke risk, genetics hasn’t been a main consideration; however, this study highlights the significant impact of genetic factors,” said Guo, adding that processing that much data at an individual level to provide a personalised assessment would be impossible without AI. To improve stroke risk prediction, the team created a new AI model based on transformers — the same technology that underpins large language models such as ChatGPT — and trained it on genetic variations to identify links to ischaemic stroke. The model identified 18 genetic variants associated with ischaemic stroke. It pointed to key genes, such as EDEM2 (active in the brain cortex) and KANK2 (in whole blood), as major contributors to stroke predisposition. The team tested the model on two major medical datasets, which together included more than 40,000 people. It successfully categorised individuals into low, medium and high genetic risk groups. Over time, individuals placed in the high-risk group were 21% more likely to suffer a first-time ischaemic stroke compared with those in the low-risk group. Until now, scientists have estimated genetic risk using polygenic risk scores (PRS) — a statistical tool that estimates a person’s genetic risk of developing a disease, based on the combined effects of many genetic variations. However, this method overlooks how different genetic variations interact with one another or how their specific positions along the DNA sequence affect overall risk. Towards more personalised medicine The new model could pave the way for future personalised and targeted treatments. “AI can be used to find associations between human genes and disease, and build personalised risk assessments, to help prevent a potentially life-changing or fatal event,” said Jie Lu, director of the Australian Artificial Intelligence Institute. Lu said the next step is a multimodal AI tool that can combine a genetic risk assessment with other types of data, including health and lifestyle surveys, scans and imaging, to better assess stroke risk. “Genetics are only one part of stroke risk. The ability to combine a genetic score with clinical and lifestyle information, including established risk factors such as blood pressure, age, atrial fibrillation, type 2 diabetes and cholesterol, could enhance monitoring and prevention,” said Lu.