Punjab sits on the upper Indus, where the great ancestry streams of South Asia met. The Indus Civilisation's heartland lies across it, the Swat valley cemeteries that gave South Asia its only substantial Bronze and Iron Age genome transect are a few hundred kilometres north, and Rakhigarhi, the one directly sequenced Harappan individual, lies just beyond its southeastern edge.
Ancestry is not identity. A genome describes which ancient populations a person's ancestors resembled; it does not say what language they spoke, whether they were Sikh, Muslim, Hindu or Christian, or which biradari or caste they belonged to, and DNA cannot recover any of that. There is no "Punjabi gene", no "Aryan gene", no ranking of ancestries and no purity to measure.
The short answer: Punjabi genomes are, to first order, a mixture of two ancient streams: the Indus Periphery ancestry of the Harappan world, itself a blend of an Iranian-related lineage and the deep South Asian ancestry called AASI, and a Steppe Middle to Late Bronze Age ancestry that arrived from Central Asia after 2000 BC. Punjabis sit near the northwestern end of the subcontinent's ancestry cline, and the Swat valley Iron Age individuals are the closest ancient reference for the region.
The Indus Periphery and Rakhigarhi#
The foundation was laid by Narasimhan and colleagues (2019). Among their 523 ancient genomes were eleven individuals from Bronze Age Shahr-i-Sokhta in Iran and Gonur in Turkmenistan who were genetic outliers at those sites. The authors called them the Indus Periphery: people of Indus Civilisation ancestry buried in its trading partners' cities. Their ancestry was a mixture of an Iranian-related lineage and AASI, the Ancient Ancestral South Indian stream, in proportions that varied along a cline.
The same year, Shinde and colleagues published the one genome recovered from a Harappan cemetery itself, a woman from Rakhigarhi in Haryana, just southeast of Punjab. She matched the Indus Periphery profile. Both papers showed that the Iranian-related lineage in these people had split from the Iranian plateau's population before the plateau's farmers existed, so it did not arrive with a Neolithic migration from Iran: it is related to the Zagros farmers but not descended from them.
AASI: the stream without a sample#
AASI is the deep indigenous ancestry of South Asia, distantly related to the Andamanese. No unadmixed ancient AASI individual has been sequenced; it is inferred from its presence in Indus Periphery individuals and from its high share in southern and tribal populations today, and every model of it uses a proxy. See our post on qpAdm models for South Asian ancestry. For Punjab it means the AASI share, typically the smallest of the three streams in the northwest, is the least precisely estimated.
The Steppe MLBA layer and the ANI/ASI cline#
After 2000 BC a new ancestry appears in South Asia: Steppe Middle to Late Bronze Age ancestry, carried by populations of the Sintashta and Andronovo horizon of Central Asia, descended from the Yamnaya-era herders mixed with European farmers. Narasimhan and colleagues traced its arrival into the Swat valley, where it is present in individuals dated 1200 to 800 BC, with admixture dates placing the mixing between roughly 1900 and 1500 BC.
The result was the cline that Reich and colleagues (2009) first described as Ancestral North Indian and Ancestral South Indian. In the 2019 framework, ASI was Indus Periphery ancestry mixed with additional AASI; ANI was Indus Periphery ancestry mixed with Steppe MLBA, formed in the northwest. Punjabis, with other Pakistani and northwestern Indian groups, sit toward the ANI end, with steppe-related ancestry in the neighbourhood of a fifth to a third of the genome depending on community and model. Moorjani and colleagues (2013) dated the ANI-ASI mixing across India to between roughly 1,900 and 4,200 years ago.
A dated timeline for ancestry in Punjab#
- 7000 to 2600 BC: Mehrgarh and the pre-Harappan farming villages of the Indus basin.
- 2600 to 1900 BC: the mature Indus Civilisation: Iranian-related plus AASI, no steppe.
- 1900 to 1500 BC: Steppe MLBA ancestry arrives and mixes in the northwest.
- 1200 to 800 BC: the Swat valley Iron Age cemeteries at Loebanr, Katelai, Aligrama and Udegram, the first directly sampled population with all three streams.
- 500 BC to 500 AD: Achaemenid, Greek, Mauryan, Kushan and Gupta Punjab; the Gandharan and Saidu Sharif individuals of Swat extend the transect.
- First millennium AD onward: Hun, Turkic, Afghan and Mughal rule with little genome-wide trace.
The Swat valley: the best local ancient reference#
The Swat valley of northern Pakistan, upstream of Punjab, holds the protohistoric grave cemeteries that supplied most of the ancient South Asian samples. Individuals from Loebanr, Katelai, Aligrama, Udegram and Barikot dated between 1200 and 800 BC carry Indus Periphery ancestry, a further share of AASI beyond what the Indus Periphery individuals had, and Steppe MLBA ancestry at a level the authors put around a fifth on average. They are the first people in South Asia sampled with all three streams together.
Nothing has been sequenced from Punjab itself for this period, but the Swat population is the closest sampled ancestor of the northwestern cline, and modern Punjabi and other northwestern groups fall near it in most analyses.
Regional and community variation within Punjab#
Punjab is not one population. South Asian genome-wide structure follows endogamous communities more than geography, with strong founder effects in many groups (Nakatsuka et al. 2017). Within Punjab, groups such as Jats, Arains, Rajputs, Khatris, Gujjars, Brahmins and Scheduled Caste communities differ modestly in their positions on the cline: Jats and some other landholding groups tend toward the highest steppe-related shares, while other communities carry more Indus Periphery-related and AASI ancestry. Punjabi Muslims, Sikhs and Hindus of the same community are genetically indistinguishable; the differences run along community lines, not religious ones.
What the Y-DNA and mitochondrial DNA add#
Punjabi paternal lineages reflect all three streams. Haplogroup R1a, almost entirely its Z93 branch, is the most frequent lineage in many Punjabi communities, on the order of a third to a half of men in the groups with the highest steppe-related ancestry. Haplogroups L-M20, J2, R2 and H represent the Indus-era and older South Asian lineages, and G and Q occur at a few percent (Sengupta et al. 2006 and later surveys). Maternal lineages are predominantly South Asian branches of haplogroup M, with West Eurasian lineages such as U2, U7, W, HV and R2 a substantial minority, as is typical of the northwest.
R1a is far more common on the Y chromosome than the steppe-related share of the autosomes would predict, while maternal lineages remain mostly South Asian, a sex-biased pattern. No haplogroup is a "Punjabi gene" or an "Aryan gene"; R1a-Z93 is one paternal line among many and carries no information about identity.
Limitations#
The Punjab plain itself has almost no ancient genomes: Rakhigarhi is a single individual just outside it, and the Swat transect is in the mountains to the north. AASI has no ancient sample and every estimate of it is proxy-dependent. The Steppe MLBA proportion in modern groups varies by model and by which steppe reference is used. And no result here bears on identity, caste status or descent from any named group: ancestry and identity are different facts.
Frequently asked questions about Punjabi DNA#
Do Punjabis have steppe ancestry?#
Yes. Steppe Middle to Late Bronze Age ancestry from Central Asia arrived in the northwest of South Asia after 2000 BC and is present in essentially all Punjabi communities. Published models place it in the neighbourhood of a fifth to a third of the genome, depending on the community and the reference populations used.
Are Punjabis descended from the Indus Valley Civilisation?#
In large part. The Indus Periphery ancestry that characterised the Harappan world, a mixture of an old Iranian-related lineage and the deep South Asian AASI stream, forms the largest component of Punjabi genomes. Steppe ancestry was added on top of it after 2000 BC, forming the northwestern end of the South Asian cline.
Are Pakistani and Indian Punjabis genetically different?#
Not by nationality or religion. Punjabi communities on both sides of the border share the same ancestry streams, and Muslim, Sikh and Hindu members of the same community are genetically indistinguishable. The differences that exist run along endogamous community lines, such as Jat, Arain, Rajput or Khatri, and along a west to east gradient.
What is the closest ancient population to Punjabis?#
The Iron Age individuals of the Swat valley in northern Pakistan, dated 1200 to 800 BC from sites such as Loebanr and Katelai. They are the earliest sampled people with all three streams, Indus Periphery, AASI and Steppe MLBA, and they sit near the northwestern end of the modern cline.
How to model Punjabi ancestry with qpAdm#
A qpAdm model tests whether a genome can be written as a mixture of chosen ancient sources and reports a p-value and standard errors. For a Punjabi genome the Ancestrify catalog is strongest at the proximal level, because it holds the Swat transect directly: Loebanr Swat Iron Age (1200 - 800 BC) is the Iron Age reference described above, and Gandharan Swat (BC 200 - 300 AD) and Saidu Sharif (25 - 500 AD) extend it into the historical period. Peninsular South Indian (BC 600 - 600 AD) supplies the AASI-rich southern end of the cline, and Parthian Iran (BC 250 - 200 AD) the Iranian-plateau direction for western Punjabi genomes.
At the distal level the catalog represents the streams more broadly. Ancestral South Indian (10000 - 2000 BC) is a proxy for AASI. Iranian Neolithic Farmer (8000 - 5000 BC) stands in for the Indus Periphery's older Iranian-related lineage. And Western Steppe Herder (5000 - 2800 BC) is a Yamnaya-era source, whereas the ancestry that reached South Asia was Steppe MLBA, already mixed with European farmer ancestry, so the steppe weight in a distal model is an approximation. Read the model record for which sources the data require. A Global25 analysis is a good first pass, placing the genome on the cline before any formal test.
Sources and further reading#
- Narasimhan, V. M. et al. (2019). The formation of human populations in South and Central Asia. Science 365, eaat7487.
- Shinde, V. et al. (2019). An ancient Harappan genome lacks ancestry from steppe pastoralists or Iranian farmers. Cell 179, 729-735.
- Reich, D. et al. (2009). Reconstructing Indian population history. Nature 461, 489-494.
- Moorjani, P. et al. (2013). Genetic evidence for recent population mixture in India. American Journal of Human Genetics 93, 422-438.
- Nakatsuka, N. et al. (2017). The promise of discovering population-specific disease-associated genes in South Asia. Nature Genetics 49, 1403-1407.
- Sengupta, S. et al. (2006). Polarity and temporality of high-resolution Y-chromosome distributions in India identify both indigenous and exogenous expansions and reveal minor genetic influence of Central Asian pastoralists. American Journal of Human Genetics 78, 202-221.
Editorial note: the hero artwork in this article was generated with AI as conceptual illustration. It does not reproduce a scientific figure, an ancient individual or a measured migration route.
