A new study reports a multi-ancestry polygenic risk score for late-onset Alzheimer's disease that outperforms earlier versions, with the biggest gains among people whose genetic ancestry is not European. A polygenic risk score condenses the effects of many genetic variants into a single number estimating a person's inherited susceptibility to a disease. Existing Alzheimer's scores have worked unevenly across populations, and the researchers behind this work set out to close that gap.
The reason for the inconsistency is straightforward: the genome-wide association studies that supply the raw material for these scores have overwhelmingly enrolled white participants of European ancestry. Variants that flag elevated risk in one population do not always carry the same weight in another, so a score calibrated on European data can lose accuracy when applied to someone of African American, Hispanic or East Asian background. By incorporating GWAS data from those groups, the new score is designed to travel better across ancestries.
That matters for anyone thinking about how genetic risk information will be used in practice. Late-onset Alzheimer's affects an estimated 6.9 million people in the United States, and its course typically unfolds over years, beginning with subtle cognitive changes long before dementia is apparent. Risk scores that work only for some patients would channel the benefits of earlier identification and research participation toward already well-served groups.
APOE e4 remains the single strongest known genetic risk factor for the disease, and polygenic scores are meant to layer additional common and rare variants on top of that signal rather than replace it. The researchers frame broader population representation as essential to building models that are robust and applicable across patient groups, which points to continued expansion of diverse genetic datasets as the next step.