What the research found

Polygenic scores—statistical tools that aggregate genetic variants to estimate disease risk—have historically performed worse in non-European ancestry groups because they were built mainly from European genetic data. Researchers tested seven different type 1 diabetes (T1D) polygenic scores across large biobanks representing diverse ancestries to see which ones bridged this gap.

Two scores stood out: T1D GRS2 and T1D MAPS. The newer T1D MAPS model showed a meaningful advantage: it maintained performance similar to T1D GRS2 in European ancestry populations but substantially improved its ability to identify T1D risk in non-European ancestry groups. This improvement was consistent across multiple independent biobanks and statistically robust.

Why it matters for you

If you're using genetic risk information to shape your health strategy—whether tracking metabolic markers, adjusting training intensity, or monitoring inflammatory biomarkers—the ancestry background of any polygenic score matters. A tool that performs poorly in your genetic ancestry group essentially gives you noise rather than signal.

This is especially relevant if you're non-European ancestry and considering genetic screening for T1D risk, or if you have family history and want to understand your actual risk profile. Better polygenic scores could support more targeted monitoring of fasting glucose, HbA1c, C-peptide, and autoimmune markers if your score suggests elevated risk. It also underscores why you shouldn't assume a genetic risk tool developed in one population will work equally well for you.

Caveats

  • This is observational biobank work; the scores predict who has T1D, not whether someone will develop it
  • Improvement in non-European groups was meaningful but moderate (0.049 AUC gain)
  • The analysis is retrospective; validation in prospective screening would strengthen the findings
  • Access to these newer scores in clinical or consumer settings remains unclear