An AI reads DNA patterns to flag who carries a higher inherited stroke risk

Most genetic risk scores read DNA one letter at a time. A University of Technology Sydney team took a different view: their AI model looks at where genetic changes sit and how they interact. Tested on the UK Biobank and the US All of Us datasets and published in Computer Methods and Programs in Biomedicine, the tool outperformed earlier genetic risk methods, identified 18 genetic variations associated with ischaemic stroke — the kind caused by a blocked artery — and found people in the highest inherited-risk group had a 22% greater chance of stroke than lower-risk groups, with one gene, EDEM2, standing out. The team is careful: a genetic score is not a diagnosis. Next, they want to combine DNA with medical history, scans and lifestyle — so the people at highest risk hear about prevention first.

SDG 3 · Good health and well-being — Mechanism: tech innovation