Indus Periphery map
Indus Periphery (Shahr-i-Sokhta) · Sistan, Bronze Age, about 3200 to 2100 BC
How to read these maps: pale yellow means little or none of the selected ancestry, orange a medium share and deep red the highest values. Each dot is the average of a modern population: hover or tap it to see its numbers. Areas with no samples nearby stay blank.
Indus Periphery (Shahr-i-Sokhta)
Sistan, Bronze Age, about 3200 to 2100 BC
Faded dots are populations the model fits less well: read their values with care. Click the map to zoom with your mouse wheel.
Who were the Indus Periphery people?
Shahr-i-Sokhta, the Burnt City, is a large Bronze Age town in Sistan, in southeastern Iran near the border with Afghanistan, occupied from about 3200 to 2100 BC. It was a busy crossroads of trade between the Iranian plateau, Central Asia and the Indus Valley.
Among the people buried there, geneticists found a few individuals whose ancestry looked very different from that of their neighbours and much closer to South Asia: a mix of Iranian farmer-related ancestry and South Asian hunter-gatherer ancestry. Researchers call this the Indus Periphery group and consider these individuals likely migrants from the Indus Valley Civilisation. The individual used here, I8728, belongs to that group, and a genome from Rakhigarhi, a major Harappan city in India, later showed the same kind of ancestry.
This Indus-related mix is the main foundation of most South Asians today, later joined by steppe ancestry, mostly in the north.
In our model, Indus Periphery ancestry peaks today in Telugu Vysya Andhra Pradesh (India), at 92.4%, followed by Balija (India) (90.3%) and Telugu Velama (India) (89.8%). All ten highest values come from India. It reaches at least 5% in 474 of the 1,998 populations on the map, and at least 20% in 418 of them. On the world scale the colours run from 0 to 85.1%, the level reached by the top 2% of populations, so the deepest red marks the real hot spots.
Where Indus Periphery ancestry peaks today
| # | Population | Share |
|---|---|---|
| 1 | Telugu Vysya Andhra Pradesh India | 92.4% |
| 2 | Balija India | 90.3% |
| 3 | Telugu Velama India | 89.8% |
| 4 | Telugu Balija India | 89.5% |
| 5 | Tamil Gounder India | 89.4% |
| 6 | Telugu India | 89.3% |
| 7 | Malayali Ezhava India | 88.3% |
| 8 | Tamil Vellalar India | 88.3% |
| 9 | Telugu Kamma India | 88.2% |
| 10 | Konkani Brahmin Gaud Saraswat Telangana India | 88.1% |
Averages of modern populations, modelled with 17 reference sources. Tap a name to find it on the map.
The 17 ancestral sources
About these numbers
These maps show model estimates for population averages, not ethnic labels and not the result of any one person. Each source stands in for a wider ancestral population, and ancestries that have no source of their own in this set, such as Oceanian or Khoe-San ancestry, are absorbed by their closest relatives among the 17. Populations the model fits less well are shown with faded dots and count less in the colour surface.
Method: Each population average (Global25 coordinates) is modelled as a mix of 17 ancient and modern reference groups with non-negative least squares, shares adding up to 100%. Fit is the distance between the population and its best mix: the lower the better, above 0.06 the mix explains the population poorly. Top lists leave out the few populations these 17 groups cannot model at all (fit above 0.15). Data generated 1 October 2026, 1,998 populations on the map.
Keep exploring
For the full picture of one population, the Population Genetics Atlas has a page for each of hundreds of modern populations, with its ancient make-up and its closest modern and ancient neighbours.
Curious about your own DNA? The free Report Finder tests which of our reports fits your G25 coordinates, before you order anything.
Frequently asked questions
Why is an Iranian site used for Indus Valley ancestry?
Very few genomes from the Indus Valley Civilisation itself are available. A few individuals buried at Shahr-i-Sokhta in eastern Iran had ancestry typical of South Asia and are thought to be migrants from the Indus world. One of them, I8728, is used here as a stand-in for that Indus-related population.
Where is Indus Periphery ancestry highest today?
In our model it is highest in Telugu Vysya Andhra Pradesh (India) (92.4%), Balija (India) (90.3%) and Telugu Velama (India) (89.8%). These are averages of modern populations modelled with 17 reference sources, so individuals within each population can differ.
How are these ancient ancestry maps made?
Each modern population average, taken from published G25 coordinates, is modelled as a mix of the same 17 ancient and modern reference sources with a non-negative least-squares fit, the same family of method our reports use. The map then interpolates the share of the selected source between the sampled places and leaves areas with no nearby samples blank.
Does a high value mean direct descent from these people?
No. Each source stands in for a broader ancestral population, and the 17 sources are fitted together, so closely related sources can trade shares between them. Read a value as how much of this kind of ancestry the model needs to describe a population average, not as a family tree.
Why are some areas of the map blank?
The colour surface is only drawn within a few hundred kilometres of real samples. Oceans and regions without data stay empty instead of showing invented values. Each dot on the map is a real population average.