Every published ancient human genome in the Allen Ancient DNA Resource, 16,496 people, placed where and when they lived. Press play and watch the first farmers spread out of Anatolia, people from the steppe sweep across Europe after 2500 BCE, and the archive's own blind spots appear.
Bones keep secrets for a very long time. In 2010 scientists published the first genome of an ancient human, read from a man who lived in Greenland around 1935 BCE. Since then labs have read the DNA of thousands more: hunters from the last Ice Age, the first farmers, Bronze Age herders, Roman townspeople, Viking raiders, medieval villagers. The Allen Ancient DNA Resource (AADR), curated by Swapan Mallick, David Reich and colleagues at Harvard, gathers every published genome in one free file. Release v66.p1 describes 17,629 ancient individuals, and 16,496 of them pass its quality checks.
The map below puts each of them where they were buried, at the time they lived. Press play and the archive comes alive: a scatter of Ice Age hunters, then a wave of farmers moving from Anatolia into Europe after 7000 BCE, then, around 2500 BCE, a sudden change in the male lines across western Europe. Colour the dots by the Y chromosome to see it. Along the way the map also shows something about the science itself: where the bones come from, and where they do not.
Each dot is one person, at the place the archive gives for their burial and the mean of their date. The time slider runs on a stretched scale, so that the last few thousand years, where most genomes come from, get more room; a window shows people within a few centuries of the date at the top (wider for older periods). Faint grey dots (Earlier people, in Map layers) lived before the window. Click a dot for the person. Colours: broad period from the date; culture label from words in the archive's group name; Y and mtDNA lines grouped into major branches (pale: no call, or a woman for the Y line). Map layers also shows the 8 Neanderthal and Denisovan genomes and the 1,133 people who did not pass the quality checks. Source: AADR v66.p1 (Mallick et al., doi:10.7910/DVN/FFIDCW), CC0. Outlines: Natural Earth. Labels: OpenFreeMap, © OpenMapTiles, © OpenStreetMap contributors. Relief and hillshade: Esri.
The AADR file has 23,089 rows. 3,970 of them are people alive today, used by geneticists for comparison. The other 19,118 describe 17,629 ancient individuals, many of them sequenced more than once. The archive grades each set of data, and I keep a person if at least one of theirs passes: 16,496 people (94%), from 3,664 sites in 127 present-day countries and territories, reported in 334 publications. The other 1,133 are flagged for contamination, low quality or other problems, and are left out of every count below.
Ancient genomes by the year they were first published
People passing the quality checks, by the year in the label of the paper that first reported them. 276 have no year in the label. Hover a bar.
The field is young and growing fast. 80% of these genomes were first published in 2020 or later, and 3,757 in 2025 alone. Single papers now report hundreds of people at a time: the largest, PattersonReichNature2021, accounts for 741.
The oldest modern humans in the archive lived around 43,925 BCE: a group from the cave of Ranis in Thuringia, Germany, and a woman from Zlatý kůň in Czechia, given the same broad date. Ust'-Ishim man, from Siberia, follows at about 42,416 BCE. But the archive is overwhelmingly recent. Half of everyone in it lived after 450 BCE; 80% lived after 3000 BCE; only 102 people lived before 10,000 BCE, at the end of the last Ice Age.
How old are they?
People in each 500 years, from 10,000 BCE to the present (102 older people are not shown). Hover a bar.
Of the 16,496, molecular sex (read from the ratio of X and Y chromosome DNA, not from the skeleton) is male for 8,601 (52%), female for 6,835 (41%) and unknown for 1,060. That is 1.26 men for every woman, a skew that reflects which burials archaeologists dig up and labs choose to sample, not the people of the past.
The archive also holds our archaic cousins. 8 archaic genomes pass the quality checks, from the Denisova and Chagyrskaya caves in the Altai mountains of Siberia, Mezmaiskaya in the Caucasus, Vindija in Croatia and Grotte Mandrin in France. The oldest, Denisova 25, is dated to about 183,050 BCE. They are on the map as large dark dots (Map layers) but left out of every count of modern humans here.
Ancient DNA does not come evenly from the past. Of the 16,400 modern humans on the map, 11,217 (68%) were buried in Europe, a continent with 9.2% of the world's people and 7.5% of its land. Africa, where our species began, has 1.9% of the genomes for 17% of the people; Asia 22% for 60%. Per head of today's population, Europe has about 66 times as many ancient genomes as Africa.
Share of ancient genomes, people and land, by continent
Genomes: modern humans passing the quality checks, by where they were buried. People: present-day population (Natural Earth estimates, 2019). Land: area without Antarctica. Russia west of 60°E (the Urals) counts as Europe. Hover a bar.
The reasons are partly natural and partly human. DNA survives best in cold, dry ground, and badly in the tropics. But money, laboratories and long traditions of excavation matter as much: the countries with the most people in the archive are Russia (1,394), Hungary (1,317) and the United Kingdom (1,151). Europe's lead holds in almost every period: it has 82% of the people from 6000 to 4000 BCE and 79% from 1 to 1000 CE. A map of ancient DNA is, first of all, a map of where geneticists have looked.
The countries with the most ancient genomes
People passing the quality checks, by the present-day country or territory the archive gives. 35 have at least a hundred. Click a row to see it on the map.
Farming began in the Near East, and ancient DNA studies have shown that it reached Europe mostly with people who moved, not just as an idea that spread. The archive's dates trace that movement. The earliest people it labels Neolithic come from Turkey, from 8750 BCE. Then, country by country, the first Neolithic burials get later the further they lie from Anatolia: Greece 6366 BCE, Hungary 5990 BCE, France 5413 BCE, Portugal 5352 BCE, Britain 3981 BCE, Ireland 3796 BCE.
The earliest Neolithic burial in each country, against its distance from central Anatolia
For each country in Europe and Anatolia with at least three people labelled Neolithic: the earliest of them (by mean date) and its distance from Çatalhöyük. Line: least squares fit, Turkey left out. Hover a dot.
Across 24 countries the earliest dates fall later by about 0.88 years for every kilometre from Çatalhöyük (correlation 0.75). Put the other way round, the farming frontier, as the sampled burials record it, moved at about 1.1 km a year, roughly a lifetime to cross a county. The first Neolithic burial in Ireland comes 4,954 years after the first in Turkey. By latitude the story is the same: the first Neolithic burial north of 45°N is from 6035 BCE (Serbia), north of 50°N from 5250 BCE (Czechia), and north of 55°N from 4110 BCE (Denmark).
Two cautions. The earliest sampled burial is not the moment farming arrived: it depends on which sites happen to have been dated and sequenced, so the true front was earlier, by an unknown amount. And "Neolithic" here is the archive's label. In the Baltic, Russia and Ukraine it often marks pottery-making hunter-gatherers, so I leave those countries out, along with groups whose labels also name hunter-gatherers and dates that are very uncertain or inconsistent (1,597 people remain).
The Y chromosome passes from father to son almost unchanged, so its branches, or haplogroups, trace male lines. In the archive, men of the Yamnaya culture of the Pontic and Caspian steppe, most of them buried between 3052 BCE and 2672 BCE, mostly carry R1b (79% of 163). Men of the Corded Ware culture, which followed in northern and central Europe, carry R1a (50%) and R1b (41%). Men buried with Bell Beaker pots carry R1b (95% of 114). The earlier farmers of Europe mostly did not: among Neolithic men in Europe, 8.7% carry R1b or R1a.
Y lines over time, by region
Share of men in each 500 years whose Y call reaches a major branch, by branch; the number above each bar is how many men. Calls too broad to place (such as R or CT) are left out. Pale bars: fewer than five men. Choose a region.
Britain shows the change most starkly. Of the 76 men in Britain before 2500 BCE whose Y line the archive resolves, none carried R1b; 71 (93%) carried I2. Of the 56 from 2500 to 1500 BCE, 50 (89%) carried R1b. The earliest R1b man in Britain in the archive, I3071, was buried with Beaker pottery at Canada Farm in Dorset, 2468-2294 calBCE (3900±30 BP, SUERC-32210). R1b still made up 91% of British men in the Iron Age, before the Romans came.
In Britain, none of 76 men before 2500 BCE carried R1b. In the thousand years after, 89% did.
The same turn happens across western and central Europe within a few centuries. In every one of the four regions below, R1b and R1a together go from a small minority before 3000 BCE to at least half of all men from 2500 BCE: in Iberia from 2% to 80% (2500 to 1500 BCE), in central Europe from 5% to 82%, and in Scandinavia from 3% to 51%, where I1 also rises.
Men carrying R1b or R1a, four regions
Share of men with a resolved Y line in each 500 years who carry R1b or R1a. Only periods with at least five men are drawn. Hover a point.
Later movements leave marks too. In Britain from 400 to 1066 CE, 30% of men carry I1, a line that in the archive is commonest in Scandinavia (35% of men there from 1 to 1000 CE) and almost absent from earlier Britain. That fits the settlers from across the North Sea that historians and the genetic studies describe.
A Y haplogroup is one line of descent, father to father, out of the thousands of ancestors each of us has; an mtDNA haplogroup is one line, mother to mother. They are a proxy for ancestry, not ancestry itself. A change in Y lines can happen because newcomers' sons outnumbered others', without the whole population being replaced, and the published studies behind this archive measure whole-genome ancestry to show how much changed. R1b is also an old and large branch: a few earlier men carry older forms of it (10 in central Europe before 3000 BCE), so R1b alone does not mark steppe ancestry. I only count branches; I do not measure ancestry here.
Some of the archive's people are famous. Ötzi the Iceman, found in 1991 melting out of a glacier in the Alps, is here under the ID "Iceman", in the group Italy_TrentinoSouthTyrol_North_MN, dated 3371-3103 calBCE (4550±27 BP; ETH-8345.1-2). His Y line is G2a2a1a2a1a1, part of G2, the commonest Y branch among Neolithic men in Anatolia in the archive (25%). Not every famous name passes the checks. Kennewick Man, found by the Columbia River in Washington in 1996, and the Anzick child from Montana, the only known Clovis burial, are both in the archive but graded "Questionable", so they are not in the counts; turn on "People who did not pass the checks" in Map layers to see them.
Famous individuals in the archive
Searched for by ID. Grade: the archive's own quality assessment. Click a row to see the person on the map.
The busiest places are mostly cemeteries dug in large campaigns. The single most sampled site is Mödling-An der Goldenen Stiege (Lower Austria, Mödling District), with 447 people, followed by Rákóczifalva-Bagi-földek (Jász-Nagykun-Szolnok Coiunty, Szolnok) with 330; both are cemeteries of the Avar period, in Austria and Hungary, between 580 CE and 832 CE. Çatalhöyük, the great Neolithic town in central Anatolia, has 119. At the other end, 1,641 of the 3,630 sites (45%) have just one person.
The most sampled sites
Places with the most people passing the quality checks. A few "sites" are whole counties where the archive gives no closer place. Click a row to see it on the map.
Data. The AADR release v66.p1 public 1240K annotation file (23,089 rows, 49 columns). Ancient means a date mean above 0 years before present (1950). An individual is one Individual ID; the 19,118 ancient rows hold 1,489 extra rows for people sequenced more than once. A row passes if the archive's assessment begins "Pass", "PROVISIONAL_PASS" or "MERGE_PASS"; a person is kept if any of their rows passes, represented by the passing row with the most SNPs (if you take each person's richest row and then filter, 16,267 pass instead of 16,496). Dates are the archive's mean, from radiocarbon or from context; years are 1950 minus that mean. Places are the archive's coordinates; 88 people have none and are not on the map.
Labels. Sex is the archive's molecular sex. Y lines are the archive's automatic ISOGG call, or its hand call where one exists (224 people), grouped by the leading letters into major branches (R1b, R1a, I1, I2, G2, J1, J2, E1b and the letters N, Q, C, O, H, T, L and D); calls above that level are "too broad". mtDNA lines are grouped by their leading letter. Broad periods come from the date. Culture labels come from words in the archive's group names (for example Yamnaya, CordedWare, BellBeaker, N for Neolithic, BA for Bronze Age, IA for Iron Age, Viking), with named cultures taking priority over broad periods; the full rules are in the code. They are a convenience, not a classification.
Comparisons. Continents are Natural Earth's, with Russia split at 60°E; population is Natural Earth's 2019 estimate (Russia's split in proportion to its town populations on each side, 73% west), land is area without Antarctica. Present-day population is a yardstick for the sampling, not a claim about the past. Neolithic arrival uses the earliest dated individual per country, which depends on what has been sampled. Haplogroup shares count men, not ancestry.