All of Us diversity and scale yield context-dependent improvements in polygenic prediction.

Multiancestry polygenic risk scores (PRSs) trained on diverse datasets improve prediction accuracy for under-represented groups, especially with increased diversity in the All of Us cohort.

• Why it matters: Current large genetic datasets lack broad human diversity, limiting equitable prediction across populations. Improving PRS performance in diverse groups is essential for health equity.
• What they did: Using 245,388 whole-genome sequences from the All of Us program combined with UK Biobank data, researchers developed and evaluated PRSs for 32 traits and diseases across diverse ancestries, analyzing effects of ancestry, methodology, and genetic architecture.
• The result: Increased diversity in the All of Us cohort enhanced PRS accuracy for several traits in under-represented populations, and multiancestry training data reduced the decline in accuracy associated with ancestry divergence, highlighting the importance of representative biobanks.