Population Genetics Atlas

Method

Each page uses one population average from the G25 modern sheet (modern3.txt, the site's reference panel). Where a country has no single national average, the page uses a central regional average, one of the closest to the mean of all that country's regional averages (for example Italian uses Marche, Turkish uses Central Anatolia, Greek uses the Peloponnese). Closest modern populations: the 10 nearest averages by Euclidean distance over the 25 G25 dimensions, skipping outlier rows, admixed sub-groups, overseas diaspora and settler rows such as Acadians or Afrikaners (except on pages for the Americas) and rows that are the same samples under another name. At most two other regional samples of the population itself are listed (four when the extra ones have their own atlas page), and at most three samples of any other population without a page of its own, so the table also shows the neighbouring peoples. Closest ancient populations: the 10 nearest averages in g25_ancient_averages.tsv from regions that make sense for the population (only groups found in the Americas for Native and Latin American pages), skipping rows flagged as outliers, contaminated or low coverage. The period is read from the label (dates in years BP, or the era it names) and is approximate. Ancestral makeup: a non-negative least-squares fit that sums to 100 percent, against a fixed set of reference populations: the same deep ancestral sources as the paid Focus reports (ancient Western, Eastern and Caucasus hunter-gatherers, Anatolian, Iranian and Levantine Neolithic farmers, Natufians, Yamnaya steppe herders, Iberomaurusians) plus modern West African (Yoruba) and East Asian (Han) stand-ins. Populations that are mostly outside West Eurasia, or that this model fits clearly worse, use a global model that adds ancient Southern Chinese and Southeast Asian Neolithic, Northeast Asian, Ancient North Eurasian and Native American genomes, and modern stand-ins where no suitable ancient genome exists: Dinka (Nilotic), Mbuti and BiAka (rainforest foragers), Ju/'hoan (San), Nganasan (Arctic Siberian), Irula (South Asian hunter-gatherer related) and Papuans. A population that is itself one of these modern stand-ins comes out at 100 percent of its own component by construction, which says nothing about its history. Components under 1 percent are dropped. The fit distance is the distance between the population and its fitted model: under 0.03 is a close fit, 0.03 to 0.06 is usual for populations with their own later drift (Northwest Europeans, Arabians), above 0.06 means the sources only describe the population roughly (Balts, East Slavs, Volga and Siberian peoples). These are models of deep ancestry, not a list of the peoples someone descends from.

Data generated 2026-10-01.