Continuous ancestry profiles from genomes and genealogies

Predicted continuous ancestry profiles from genomes and deep genealogies place 78% of documented ancestry in the correct region, far exceeding categorical labels in a validation set of 28,332 individuals.

• Why it matters: Single regional labels oversimplify ancestry, which is multidimensional and often reflects contributions from multiple ancestral sources. The study tests whether continuous profiles better capture this complexity than categorical assignments.
• What they did: Relatedness among 28,332 genotyped individuals was computed via haplotype painting and ordination, then regressed on regional ancestry proportions from deep genealogies for one-third of the cohort. A validation set was used to compare predicted profiles with training-set averages and categorical labeling.
• The result: Predicted profiles correctly localized 78% of documented ancestry in the correct region (vs 48% for the training-set average), while single-label assignment achieved 32% and even perfect “largest-region” classification reached 38%. This supports representing ancestry as a continuum rather than a label.

AI summary of the abstract · may contain errors · not medical advice · read the paper.