Genetic Evidence — Y-DNA Q1b (L275) and the Steppe Connection

Modern population genetics testing of the Varaha lineage

Modern genetics offers a line of evidence independent of the textual record. The book reports the author’s own Y-chromosome haplogroup — Q1b (Q-L275) — and analyses it against published reference panels using the qpAdm framework. What the results establish is a lineage with deep roots on the Iranian plateau and the Bronze Age steppe, overlain by South Asian ancestry. What they do not establish is the date or the agent of its arrival in India, and the distinction is the whole point of this page.

Why genetic evidence matters here — and where it stops

Y-chromosome lineages descend strictly through the male line and do not recombine, so a haplogroup signature carries unbroken from father to son for hundreds of generations. That is the strength of the method, and it is real. The limits are equally real, and three of them bear directly on this case.

First, a haplogroup is not a people. Q-L275 is carried today by Iranians, Afghans, Pashtuns, Punjabis, Jews of several communities, and Arab lineages of the Gulf. No one of these owns it, and sharing it establishes common male-line descent at a depth of several thousand years, not membership of a common historical group.

Second, Q-L275 in north-western South Asia is older than the Huns. The Iron Age cemeteries of the Swat valley, published as part of the Reich lab’s South and Central Asia dataset, contain Q lineages — including Q-Y1150, within L275 — at around 900 BCE. Whatever brought this haplogroup to the Indus borderlands had done so before the Achaemenids, let alone the Alkhans. A clan that carries it today may descend from Huns, but the haplogroup itself is not the evidence of that.

Third, there is no Hephthalite reference panel. Ancient DNA from the Xiongnu is now abundant, and the published Xiongnu Y-chromosomes are dominated by other lineages altogether — R1a, N, C2 and branches of Q other than L275. For the Hephthalites proper there is effectively nothing: no published, securely attributed Hephthalite genome series exists at the time of writing. An earlier version of this page implied otherwise. Until such a series is published, “Hephthalite ancestry” is not a quantity any test on this page can measure.

A note on nomenclature

The same branch is written Q1b-L275 in the ISOGG-derived trees used by most commercial laboratories and Q3-L275 in Balanovsky et al. (2017). They are the same clade; the numbering schemes differ. Readers should also note that L275 is not downstream of L54, the marker that defines the large American and north-Eurasian Q branch; the two are separate major lines within Q-M242, and any chart on this site or elsewhere that nests one inside the other is in error.

The qpAdm framework

qpAdm — published by the David Reich lab and refined by collaborators including Choongwon Jeong — is a tool for testing whether a target population can be modelled as a mixture of “source” populations against a panel of “reference” outgroups. Where simpler tools simply report admixture proportions, qpAdm explicitly tests model fit (a p-value above the conventional 0.05 threshold means the model is not rejected) and rules out plausible alternative ancestries. A successful qpAdm model — especially one that survives rotation against multiple reference panels — is a serious test of a migration hypothesis. It is also only as good as the sources available to put into it, which is where the missing Hephthalite panel bites.

Key components

The qpAdm models reported in the book identify five recurring ancestral components in the Varaha sample:

• Eastern Hunter-Gatherer (EHG) — the Mesolithic Siberian taiga component carried into the steppe.
• West Eurasian Steppe — Yamnaya / Andronovo / Sintashta-derived ancestry.
• East Asian / Siberian (Mongolia North Neolithic) — the component that distinguishes steppe populations such as the Xiongnu from southern Iranian ones.
• Iranian Neolithic (Iran_GanjDareh_N) — the substrate of the Iranian plateau.
• South Asian (AASI / Indus Periphery) — the substrate the lineage encountered in the subcontinent.

Q-L275: what it anchors, and what it does not

Y-haplogroup Q1b (Q-L275) is the marker that anchors the male line. Its phylogeography is published in Balanovsky et al. (2017), BMC Evolutionary Biology, as Q3-L275: the modern frequency centre lies in north-eastern Iran, Afghanistan and adjacent Central Asia, with secondary clusters among Pashtun and Punjabi populations and a well-known presence in several Near Eastern lineages. That distribution is West and South Asian. It is not a steppe distribution, and the clade’s coalescence is reckoned in millennia, not centuries.

Mascarenhas et al. (2015), in BioMed Research International, proposed an “ancient lineage” of Q1b entering north-west India from inner Asia in the late first millennium CE, and this page formerly treated that as corroborating the Hephthalite narrative. It should not have. The Iron Age Swat samples place L275 in the same region some fifteen hundred years earlier, which means the 2015 date cannot stand as an arrival date. The honest statement is narrower and still worth making: the author’s male line belongs to a clade whose homeland is the eastern Iranian and Afghan highlands, and which was present on the Indus frontier long before the medieval period. That is compatible with a frontier-dynasty origin for the clan. It is not proof of one, and it would be equally compatible with several other histories.

What would actually test the thesis

A haplogroup label cannot settle this. Two things could move it. The first is a Big Y-700 or equivalent full-sequence test on the author’s sample, which would place the line on a dated terminal branch and give a time to the most recent common ancestor with the Afghan and Khorasanian clusters — a TMRCA in the right centuries would be genuinely probative; one of three thousand years would close the question the other way. The second is sampling across the streams: if the Puadh Varyāh, the Jalandhar Warya, the Cholistan Varyáh and the Ambota line share a recent terminal branch, the clan is one lineage and the traditions are describing a real dispersal; if they do not, they are separate groups that share a name. Neither test has yet been done. Both are more informative than any model on this page.

Medieval-era admixture

The genetic picture is not static. Three later layers add complexity, none of them specific to this clan:

• Turkic migrations (6th–11th c. CE) — overlapping the Turk Shahi period and adding a small east-Eurasian component across the north-west generally.
• Mongol Empire (13th–14th c. CE) — a further east-Asian signal, particularly along the Silk Road corridor.
• Silk Road and medieval Iran — continuous gene flow with the Iranian plateau through the whole period, which is the simplest account of the Iran_GanjDareh_N component.

What the data does not say

The qpAdm models reported in the book do not include a direct Y-DNA “source” — qpAdm is an autosomal admixture tool, not a Y-haplogroup classifier. The Q-L275 finding is reported separately, on the basis of commercial Y-chromosome SNP testing and the Balanovsky et al. published phylogeography. The two lines of evidence are consistent with each other and with a north-western origin for the male line. Neither of them, singly or together, dates the clan’s entry into India or identifies the people who brought it. Anyone citing this page should cite it for that much and no more.

For the textual record on the same migration see Origins — The Altai Steppe and Migration West. For the dated documents see The Documentary Record. For the published references see the Sources page. For the full documentary-and-genetic synthesis see The Bárah–Waryah (Varāha) Rajputs.

Autosomal admixture — G25 model fit

A four-source admixture model fitted to the Global 25 PCA coordinates of a Lower Himachal Rajput sample

In addition to the qpAdm framework discussed above, the same lineage has been tested independently against the Global 25 (G25) coordinate system — a 25-dimensional principal-component space built from a few thousand ancient and modern reference genomes. The G25 vector for the Lower Himachal Rajput sample was modelled as a four-source mixture; the best fit returned the four ancient populations below, with a fit distance of 1.86% (under 2% is considered a solid fit, under 1% is excellent).

What each source represents

• IRN_Shahr_I_Sokhta_BA3 · 62.6% — one of the standard Iran-Bronze-Age proxies; late-3rd-millennium-BCE samples from the Helmand-basin urban site of Shahr-i-Sokhta in south-eastern Iran. Captures the broad “Iranian-plateau farmer + early eastern-Eurasian admixture” component that dominates South Asian autosomal ancestry.
• RUS_Srubnaya_Alakul_MLBA · 18.8% — Middle-to-Late-Bronze-Age steppe pastoralists of the Andronovo / Sintashta cultural horizon, c. 1900–1500 BCE, on the Volga-Ural steppe. The canonical Indo-Iranian / steppe signal that arrives in South Asia in the Late Bronze and Iron Age. 18.8% is in the typical band for North-Indian Brahmin/Rajput populations and slightly above the all-India average.
• UZB_Dzharkutan1_BA · 14.0% — BMAC (Bactria-Margiana Archaeological Complex) at the Dzharkutan type-site in southern Uzbekistan; the urban Oxus civilization, c. 2200–1500 BCE.
• NPL_Chokhopani_2700BP · 4.6% — an Iron-Age Tibet-related sample from Mustang, Nepal. The small East-Asian-via-Himalayas signal that turns up in Himachal and Punjab samples generally.

How this lines up with the Varaha narrative — and how far

The textual story sketched on this site runs Altai origins → Hephthalite Sogdiana / Bactria → Hindu Kush → Punjab → Shivalik. In autosomal terms the G25 fit gives roughly 77% combined Iran-BA and BMAC ancestry, 19% steppe Indo-Iranian, and a 4.6% Tibet-related trace. It is fair to say this is compatible with that story. It is not fair to say it confirms it, and an earlier version of this page did say so.

The reason is that these proportions are unremarkable for the region. A steppe component near 19% and an Iran-BA component near 60% is the ordinary profile of an upper-caste north-Indian sample from Himachal or Punjab, whatever its clan history. The sources are all Bronze Age — none is post-Hunnic, because G25 holds no Hephthalite panel — so the model simply has no way to register a fifth- or sixth-century CE arrival, and it would return much the same numbers for a lineage that had never left the plains. A result consistent with a hypothesis is not evidence for it unless it would have been inconsistent with the alternatives, and this one would not.

Caveats

The 1.86% fit distance is good but not perfect. Under 1% would be excellent; above 3% is suspect. The model is also under-determined: with only four sources one can get reasonable fits even when the true mixture is more complex, and a four-source fit on a single individual carries no confidence interval worth quoting. Substituting an Asian Hunnic or Xiongnu proxy for the Srubnaya-Alakul slot and comparing fits would at least ask whether the steppe signal is specifically Hunnic or generic Bronze Age Andronovo; that test has not been run, and should be before any weight is put on the steppe figure. One sample is also one sample: the clan-level question needs the multi-stream sampling set out above. The G25 numbers are offered here as context, not as corroboration.

In this era

• Y-haplogroup affiliation: Q1b (Q-L275), also written Q3-L275; not downstream of L54
• Attested in the region by the Iron Age: Q lineages within L275 in Swat, c. 900 BCE
• qpAdm autosomal admixture model — Mongolia North Neolithic + EHG + Iran_GanjDareh_N + AASI + West Eurasian Steppe
• Outstanding: Big Y-700 on the author’s sample; terminal-branch comparison across the Puadh, Jalandhar, Cholistan and Ambota lines
• Primary references — Balanovsky et al. (2017), Mascarenhas et al. (2015), Narasimhan et al. (2019, Swat Iron Age), Schmidt & Seguchi (2016), Habu et al. (2018)