The population genetics explorer

An atlas of
human genetic diversity.

Explore genetic distances among 319 populations, using data from 10,665 individuals drawn from a collection of published datasets and the Allen Ancient DNA Resource (AADR). Use the dropdowns to inspect any pair, or browse the nearest-population table for the selected reference group.

319populations
10,665individuals
302,955shared SNPs
50,721population pairs
Download Matrix TSV Download Long TSVDownload CSVPopulation sizesPopulation namesDiploid sensitivity CSVMethods and provenance
Explore the world

A world of genetic connections

Choose populations above, or use the buttons below to pick them on the map. Every coloured dot shows Fst relative to your reference.

Click a dot to choose the comparison population.
Hudson Fst from reference
0 · closer0.15≥0.30 · more distant

Choose a selection mode, then click a dot. Scroll to zoom; drag to pan.
319 populations mapped. Solid markers use matched metadata coordinates; hollow markers use approximate country or region locations. Dots stay at their geographic anchors. Click overlapping dots to choose from the nearby populations, or zoom in. Locations are geographic guides, not population boundaries.

Select a marker to inspect its population.
Methods, estimation assumptions & sample filters

214 records have no heterozygous calls. The default counts these as one observed allele per site; this is an inferred ploidy assumption. Switch to the diploid analysis to inspect its effect. Only populations with at least 5 retained individuals are included. The threshold is applied after deduplication.

Buryat (41) and Mongol (9) are separate. Kalmyk (10) is also retained separately. Eight source-verified Tuscan individuals are labelled Tuscan. Duplicate and excluded records are omitted. Negative estimates are retained.

SNP coverage: Each SNP must have genotype calls in at least 80% of individuals and at least 5 individuals in each population being compared. Both estimation modes use the same coverage rule. Comparisons with fewer than 50,000 qualifying SNPs are unavailable on the map and in rankings; downloadable tables retain these estimates for audit.

Select two populations to view their Hudson's Fst.
-
Population 1 N-
Population 2 N-
Observed SNPs-
Jackknife SE-

Nearest Populations To Population 1

Lowest Hudson's Fst first
Population N Hudson's Fst SE Observed SNPs

Population names. Expanded 1000 Genomes codes follow IGSR descriptions. Recruitment locations describe the original cohorts; approved pooled groups may include samples from additional locations. Data and methods. A collection of published datasets combined with present-day samples from AADR v66, including Inner Eurasian Human Origins and West Javanese data. All pairs recomputed on the shared marker panel after individual deduplication. Pair-specific SNP counts reflect missing data. SE uses a block jackknife; it does not capture label or inferred-ploidy uncertainty. Downloads contain the default adjusted estimates unless labelled diploid sensitivity. Hudson estimator: Bhatia et al. (2013); AADR: Mallick et al. (2024).