search this blog

Thursday, April 20, 2017

Swat Valley "early Indo-Aryans" at the lab

For a while now I've been hearing rumors that the Reich Lab was working on Late Bronze Age and Iron Age samples from Pakistan's Swat Valley for a new paper on the Indo-Europeanization of South Asia. This has now been confirmed officially in a newsletter released by Padova University. See here.

Anyone want to guess how they'll turn out? I'm betting they'll be modeled as well over 50% Steppe_EMBA or Yamnaya-related. In other words, similar to the Kalasha people of the Hindu Kush, but even more Yamnaya-like. Exciting times ahead.

The archaeological paper mentioned in the newsletter is available behind a paywall here. I skimmed through it and didn't really understand it. But the authors seem to agree with the general consensus that these samples represent some of the earliest Indo-Aryan speakers in South Asia; likely descendants of recent migrants from the Central Asian steppes.

Abstract: The protohistoric graveyards of north-western Pakistan were first excavated in the 1960s, but their chronology is still debated, along with their relationship to broader regional issues of ethnic and cultural change. Recent excavation of two graveyards in the Swat Valley has provided new dating evidence and a much better understanding both of grave structure and treatment of the dead. Secondary burial was documented at Udegram, along with the use of perishable containers and other objects as grave goods. The complexity of the funerary practices reveal the prolonged interaction between the living and the dead in protohistoric Swat.

Massimo Vidale and Roberto Micheli, Protohistoric graveyards of the Swat Valley, Pakistan: new light on funerary practices and absolute chronology, Antiquity, Published online: 04 April 2017, DOI:

See also...

The peopling of South Asia: an illustrated guide


Gioiello said...

Si tratta dello scavo di alcune necropoli megalitiche datate col metodo del radiocarbonio dal 1400 al 900 a.C., lo stesso periodo temporale in cui gli specialisti collocano la diffusione, dall'Asia centrale verso il Subcontinente IndoPakistano, delle lingue Indo-Arie, una branca della grande famiglia Indo-Europea.

Se per determinare il DNA dei resti ossei l’analisi è stata affidata a David Reich del Department of Genetics della Harvard Medical School, è solo grazie alla microstratigrafia, una "specialità" della Scuola archeologica dell'Università di Padova, che è stato possibile ricostruire nel minimo dettaglio i gesti rituali compiuti da queste antiche popolazioni dell'età del Bronzo durante i loro lunghi funerali.

MaxT said...

"I'm betting they'll be modeled as well over 50% Steppe_EMBA or Yamnaya-related. In other words, similar to the Kalasha people of the Hindu Kush, but even more Yamnaya-like."

I agree

any clues on when the study will be out?

rozenfag said...

Excellent! I hope they will also test older samples for comparison.

Davidski said...

any clues on when the study will be out?

Nick Patterson did a talk on South Asia a while ago at Max Planck, and I'm pretty sure that this was connected to the new paper with these Swat Valley samples. But it might be months before the paper is published.

Nirjhar007 said...

Haha . They found indications for a Matriarchial society there . Quite like the case of Kalash , who still have some matriarchal elements: freedom of women, and goddess cult:

So much for the Immortan Joes coming down from steppes , that only happens in movies , like of Dave and Anthony etc .

Genetically they should show similarity with EBA groups with minor ASI admixture , they will be not like Andronovo or Sintashta .

But we need older samples from India proper to close the case and bring peace.

Davidski said...

It won't be very difficult to prove that these samples have a lot of ancestry from EBA Eastern Europe, with or without older samples from India.

The annoying thing is that some people here will simply choose to ignore the results and claim that they show something totally opposite to what they really show.

Jijnasu said...

@nirjhar goddess worship doesn't necessarily imply matriarchy

Nirjhar007 said...

I am not implying anything , I am just saying whats gonna happen , keeping out any theory or hypothesis .

Nirjhar007 said...

Also for Ulug depe , although its not Nuclear aDNA , but authors found 100% West Eurasian Hg's, and of course Yamnaya type ancestry which again is similar that of Kalash and Ulug Depe also gives :

The archaeologists have discovered a set of geometric ornaments and stone figures with depictions of the goddess of fertility.

Again, all these hints are not in favor of a Patriarchal society . This is not implication but pointing the facts . Needs further digging of course I admit . But first we need to get the genetic data in hand, to establish a trustworthy pattern .

Nirjhar007 said...

It won't be very difficult to prove that these samples have a lot of ancestry from EBA Eastern Europe, with or without older samples from India.

Type yes, from where its another issue . Your only hope is Andronovo , but if it itself is not the ideal fit ,then you have nothing . Also to verify the time depth of such ancestry you need the older samples from India of course and you need Hg pattern time to time .

Davidski said...

Your only hope is Andronovo, but if it itself is not the ideal fit, then you have nothing.

I couldn't care less what type of Eastern European steppe population moved into South Asia.

It's either gonna be Steppe_EMBA or Steppe_MLBA, or a mix of the two. Whatever.

Salden said...

There isn't a single example of a matriarchal society. Having lines of descent traced from a mother isn't matriarchy. Women lack both the strength and psychology to rule.

Nirjhar007 said...

Or nothing , a pre-existing local ancestry, that is creating the pseudo-steppe effect .

Which is more likely , as large scale migrations are ruled out by everybody .

It will be a miracle! if such Steppe Invasion get proven , well I have no probs in accepting something miraculous , but for now I stick to the reality .

Davidski said...

Or nothing, a pre-existing local ancestry, that is creating the pseudo-steppe effect.

That's hilarious.

Nirjhar007 said...

Good for you mate .

Nirjhar007 said...

Anyway, you say they support Steppe Scenario , well they mention it, as its the most popular theory , that is not their fault , you have to mention that, but they like me don't take any shit , for example on a Shitty claim by Kuzmina they clearly point out :

''The interpretation by Kuz’mina (2007: 309) of double burials as
an “Indo-Aryan antecedent” of the sati custom of Hindu aristocratic tradition—the ritual
suicide of the widow after her husband’s death—has no supporting evidence, as it is not
backed by the required microstratigraphic records, nor by osteological studies.''

As every time someone closely observes the AIT situation , they find it not reliable ....

P Piranha said...

A small note on some very important aspects of fsts which anyone interpreting them must appreciate. Given a dataset separated into groups, fst is a calculation of the variation between groups over the variation within groups. Thus there is an extra knob to turn: if the group is very homogeneous internally, as is possible with high drift and endogamy, then its fst to all other populations is uniformly elevated; if the group is very diverse and and has a large long-term population size, then its fst to all other populations is uniformly depressed. If we simply compare fsts its possible to get weird results like Kalash or Karitiana being the most diverged Eurasian population from everyone else. To interpret fsts properly you have to look at relative fsts, i.e. are Zoroastrians relatively closer to Steppe and further away from Iran_N compared to Iranians? What is the ratio of distances to Steppe and Iran_N, etc. Or you have to pass the figures through something that will take care of that for you, e.g. PCoA.

P Piranha said...


Razib treated you very roughly because even if old R1a clades are found in India, the vast majority of Indian R1a clades coalesce with Central Asian and European R1a clades within the metal Ages. In other words a single man gave rise to these millions of Y chromosomes after 6kya, including the vast majority of Indian R1a, and the old R1 or R2 in India is of little consequence for recent population movements that coincide with the IE expansions. A sample with a Y chromosomal haplogroup belonging to a clade just upstream of Indian R1a has already been found in Europe just a few generations before the calculated origin time of recent Indian and European R1a, and Z93 outright has also been found later in the same region, with genetic continuity throughout the period and no trace of South Asian ancestry. This indicates that Z93, and therefore the R1a found in the vast majority of Indians, originated in an individual of Steppe MLBA background, no matter the origin of R or R1 as a whole. Contra this one can postulate that ASI was not even present in Balochistan as late as the metal ages, making the Indus a viable--if odd--location for the origin of the Southern ancestry in Steppe MLBA, with a massive northward movement of ASI-carrying populations in the historical period, such that we reach the 15% ASI we find in the Indus Valley today. Possibly with the importation of millions of Andamanese or Tamil slaves in historical India, so that 15% of the recent ancestors of present day Indus peoples are actually pure Andamanese Tribals or 30% of them are pure Tamils or 70% of them are Gujaratis--but no one has ever heard of such a thing.

P Piranha said...

Going back to the topic of fsts, very possible that the high diversity of Indian populations will cause them to have low fst to ancient genomes that does not reflect real connection.

Rami said...

The earliest IA grave culture sites are from around 2100-1900 BC, this corresponds with the granary grave burials. So getting dna from those would provide an exact answer. By Iron Age, Indo Aryan migrations had ended, so they will probably resemble Kho or Kalash populations, I would think.

EastPole said...

Nirjhar007, listen to this:

Very archaic religion. Pagan Slavs and early Indo-Aryans also believed in One God only and worshiped his manifestations.

Kalash goddess Dizane may be related to Slavic Dziewanna (Dievana) and Latin Diana.

Rob said...

Interesting article, which doesn't overreach in conclusions. Critiques traditional appraisals of 'IA archaeology", but highlights this was a period of significant cultural flux.
The dating offered is now lower to 1500 BC, upper limit. The primary interred was a female, followed shortly by younger males, showing evidence of de-fleshing or disarticulation. ? slaves, ? boy-toys :)
Whatever the case, there is no clear analogies to the steppe ? perhaps steppe admixture had arrived earlier, or was incorporated into a wholly new societal/ cultural outloook

Rob said...

Although that side crouched inhumation in Fig7 looks similar to that seen in steppe cultures

Gioiello said...

@ EastPole
"Kalash goddess Dizane may be related to Slavic Dziewanna (Dievana) and Latin Diana"

η Άρτεμις εν Ταύροις της Σκυθίας τιμωμένη; η από μέρους, των ποιμνίων επστάσις. η ότι η αυτη τη σελήνη εστι καί εποχειται ταύροις
Difficult to say how from * diwiana Kalash Dizane came.
"Diana (pronounced with long 'ī' and 'ā') is an adjectival form developed from an ancient *divios, corresponding to later 'divus', 'dius', as in Dius Fidius, Dea Dia and in the neuter form dium meaning the sky. It is rooted in Indoeuropean *d(e)y(e)w, meaning bright sky or daylight, from which also derived the name of Vedic god Dyaus and the Latin deus, (god), dies, (day, daylight), and " diurnal", (daytime).
On the Tablets of Pylos a theonym διϝια (diwia) is supposed as referring to a deity precursor of Artemis. Modern scholars mostly accept the identification.
The ancient Latin writers Varro and Cicero considered the etymology of Dīāna as allied to that of dies and connected to the shine of the Moon" (Wikipedia).

Anyway it seems that Latin Diana is linked to pre-Greek διϝια (diwia), i.e. to the Balkans and western Europe rather than to the Steppes.

P Piranha said...

David, there's a chance you might be disappointed, not because the migration didn't happen, but because these burials look almost nothing like steppic burials... what with the defleshing and reuse of the same tomb, wooden-above ground structure showing that people must have returned repeatedly to or had continuous cultic presence at the same site, i.e. a decidedly non-pastoral pattern, and the female-centric burial form. Either there was extreme transformation of steppic burial practices, or we are looking at the wrong set of tombs if we're searching for the migration from the steppe.

Nirjhar007 said...


Thanks for the video , I will give it a look .Yes Vedics also suggested everything comes from one , the gods are the different manifestations of the one . You may know this famous line of RV. 1.164.46 :

एकं सद विप्रा बहुधा वदन्त्यग्निं यमं मातरिश्वानमाहुः ||
कर्ष्णं नियानं हरयः सुपर्णा अपो वसाना दिवमुत पतन्ति |

46 They call him Indra, Mitra, Varuṇa, Agni, and he is heavenly nobly-winged Garutmān.
To what is One, sages give many a title they call it Agni, Yama, Mātariśvan.

Alberto said...

So finally it's coming! Great news. I think these samples will give us some very good clues, even if they are a bit late and will always leave some doubts.

I agree that if we have to guess to whom should they resemble, the Kalash seems like a good guess. But I also agree that there's little evidence of any steppe connection here. Another bit from this article:

"What we can understand from the graves is that they were a very powerful civilisation. They were socially well organised and apparently very peaceful because no weapons were found from the site, unlike most civilisations.

Apparently no war chariots, horses, weapons, kurgans... Maybe these are the rumored samples who were L-M20? Let's wait and see how they are autosomally and their Y-DNA.

Nirjhar007 said...

Although that side crouched inhumation in Fig7 looks similar to that seen in steppe cultures

This is not Yamnaya style, IINW it is like Iranian Neolithic and Mehrgarh. See for example here :
see also this:
Here are Yamnaya burials:

Nirjhar007 said...

The link is fine . Hit the 'yog' button in the end to get the research .

Nirjhar007 said...

Thank you EastPole . That was a nice documentary ! :) .

Davidski said...

I gotta run. Going out for a steak and craft beers. So please behave in the comments while I'm gone.

In fact, let me just add that there's nothing to get emotional about here, because it's actually all over.

The Aryan invasion did happen, and it was big. Just sit back, relax, and wait for the relevant papers and data to come out later this year. There's nothing you can do to change the outcome.

Olympus Mons said...

@ Davidski.
Yes it did. To the east.
And to the west...kinda. To poland. But poland is still east, right?
From that point westward...all they did is provide females to Bell beaker males.
The rest is confabulation...which is a mental disorder....kinda.

postneo said...

"A sample with a Y chromosomal haplogroup belonging to a clade just upstream of Indian R1a has already been found in Europe just a few generations before the calculated origin time of recent Indian and European R1a, and Z93 outright has also been found later in the same region, with genetic continuity throughout the period and no trace of South Asian ancestry."

1) I think this "few generations" is being taken too literally.

2) no trace of south asian ancestry in a z93 is meaningless. Yamnaya ancestry supposedly shows up in CW with no trace of uniparentals like r1a. So why suddenly a double standard of ascribing r1a as a pan continental deep ancestry marker marker when its not even satisfied locally.

Taymas said...

"Genetically they should show similarity with EBA groups with minor ASI admixture , they will be not like Andronovo or Sintashta"

Nirjhar, do you mean Steppe EBA groups? If so, your prediction can't be differentiated from the steppe hypothesis. If you meant some other group, like Harappa EBA, I suggest you spell out your prediction just like Davidski has.

Davidski said...


Isn't there plenty of both R1a and R1b on the Eneolithic/Bronze Age Steppe in maternally and autosomally closely related populations?

Yes or no?

If yes, where does that leave your "no R1a in Yamnaya so far" argument?

Nirjhar007 said...

Taymas ,

Having a steppe like ancestry and direct steppe ancestry are two different things .

Modern S Asian groups are similar of EBA Yamnaya type groups . But its not possible for an early migration from Yamnaya , since there is nothing for such scenario . The Best chance is Andronovo , but it itself not an Ideal fit for Modern -I-As ...

About R1a issue , yes neither Yamnaya or Afanasevo are candidates . They don't show any R1a-M417+ . Extinct clades of R1a1 were there all over Eurasia . They are not so important , the case of R1a-M417 will be I think solved with Indian data .

Matt said...

@ P Piranha, using relative Fsts as a control can be useful, but there are some strange questions around for me about adjusting by Fst distance to an African outgroup, as discussed with Sein and Alberto on another thread.

More detail: See - (the blog controls here don't like how many links I've put in the post, so I've put it on paste bin and linked).

(As an post script to this post: same kind of absolute and relative Fst for Europeans - and same kind of questions to a reduced degree; more Basal Eurasian populations look less related to others in the measure relative to Mbuti than the direct method, even net of any strong recent drift).

Taymas said...


I understand those are two different things, I wasn't trying to straw-man you. That's why I phrased it very precisely: your prediction (for LBA/IA Swat) can't be distinguished from the steppe hypothesis. That means your framework cannot gain on the steppe hypothesis regardless of what the Swat results return.

Nirjhar007 said...

Those burials have no relation to any hypothetical Aryan Migration . They will be similar to modern Kalash as said earlier . Those people were not ancestors of Indo-Aryans BTW , they were most likely Proto-Kafirs.

Taymas said...

Could you provide a hypothetical test for "Those burials have no relation to any hypothetical Aryan Migration" and "Those people were not ancestors of Indo-Aryans BTW , they were most likely Proto-Kafirs"?

Sorry, I know I'm not a prolific commenter here but if you go back you'll find I've been pretty consistent on this issue. We can argue til the end of time and get nowhere unless the advocates properly test their theories. I understand it's probably annoying that I, not providing my own tests, is asking for this, but I'm also not pushing any theories particularly hard (I don't consider myself sufficiently informed).

Al Bundy said...

The biggest fly in the steppe theory ointment for me is the early deep split of Anatolian.Everything else seems to be falling into place for a steppe homeland and by homeland I mean that all not just some IE languages comes from there.

Nirjhar007 said...

Taymas ,

Of course there are archaeological , linguistic data behind my suggestions and an article on Swat-Proto-Kafirs and their likely origins is under development . I will link you if you remain in discussion . But creating such articles takes patience and careful progress . Also we need the aDNA version, for an overall pattern . So you will have to wait , but you will have it :) .

Nirjhar007 said...

Everything else seems to be falling into place for a steppe homeland and by homeland I mean that all not just some IE languages comes from there.

You are under illusion . See here :

Al Bundy said...

I'll take a look at it thanks.

Atriðr said...


एकं सद विप्रा बहुधा वदन्त्यग्निं यमं मातरिश्वानमाहुः ||
कर्ष्णं नियानं हरयः सुपर्णा अपो वसाना दिवमुत पतन्ति |

46 They call him Indra, Mitra, Varuṇa, Agni, and he is heavenly nobly-winged Garutmān.
To what is One, sages give many a title they call it Agni, Yama, Mātariśvan.

This is an erroneous translation. Your first Skt. verse corresponds correctly to your second English verse. But your second Skt. verse is mistranslated. Your first English verse in Skt. should be: इन्द्रं मित्रं वरुणमग्निमाहुरथो दिव्यः स सुपर्णो गरुत्मान

Rami said...

Dr.Muhammad Zahir is an expert on this , from my conversation with him he said there was no drastic shift in grave styles, some pottery styles were similar to those found on the fringes of Iran. No Steppe pottery was found.
But he did say skull types in some graves were quite different from those found in others even though they are buried in the same. So I would assume those different skull types are Aryans. He said it did not seem like an invasion more of a demic diffusion.He said more archaic Indo Aryan sites will be found in Mohmand Agency , as whatever earlier digs did take place there , pointed to strong metallurgy , which is typical for Indo Iranians.

Nirjhar007 said...

EastPole, Atrior,

My deepest Apologies , here is the correct Sanskrit Version :
इन्द्रं मित्रं वरुणमग्निमाहुरथो दिव्यः स सुपर्णो गरुत्मान |
एकं सद विप्रा बहुधा वदन्त्यग्निं यमं मातरिश्वानमाहुः ||

In hurry I copied wrong :) .

Alberto said...


I think it's based on the archaeology of the site. After looking at the paper, and as Rob already pointed above:

"Thanks to the anthropological analysis by Maria Letizia Pulcini, we know that at Udegram, in the majority of cases, the primary burial was a mature adult female, resting directly on the floor of the grave, followed by a secondary burial of a male, of the same age or slightly younger, but not young enough to be considered her son. [...]

All of this suggests a female-centred pattern: women were buried in megalithic graves, thereby alluding to their leading role in the household, while the remains of a male relative (?) played an important but secondary role as an ‘accessory’

This sounds quite definitive. So if they're going by these samples to prove a steppe migration I'd guess it's because the samples don't have any Yamnaya-like admixture, and therefor it proves that Yamnaya-like admixture in modern populations arrived later from the steppe (after 1000 BC).

Or else, I don't know. If these samples do have a lot of Yamnaya-like admixture, it's going to be hard to argue that such ancestry came from the steppe. Unless one argues that South Asian women importing young males from the steppe as "stallions" to produce offspring. Which would be another weird theory to add to the collection.

Let's see what the DNA shows, and then we can debate on better grounds.

Nirjhar007 said...


Zahirs dates are too old .

Rami said...


No they are not

P Piranha said...

With haplotypes of a certain length, it is statistically impossible for it to be anything other than shared ancestry, such that the vast majority of segments in a haplotype analysis with some threshold are real chunks of recent shared ancestry whose age is some function of the length of the segment, so Truthprevails doesn't know what he's talking about. In fact, haplotypes are used in the field as the most direct and straightforward proof of genealogical links, if the genomes are of sufficiently high quality. Only the oldest and the most outdated research supports long-term genetic isolation in India, no one believes this, not even Indian geneticists today, e.g. Palanichamy or Basu.

That said, not sure that the genomes will turn out like Kalash. Perhaps a small subset of them will. The IE introgression into India is very different sociologically from the one in Europe, in India the remnants of the Indus Valley are very complex social formations on which Steppic practices were adduced--it was not a cultural replacement--and there is much more penetrance of the previous cultural groups, in religion and epic and possibly also anthropologically. We can already see this in cemetery H, where there is diversity in burial practices, pointing to subgroups or subcultures with fundamentally different self-conceptions and normative-religious worldviews. Its possible that there was continuous migration over a relatively long period and the proportion increased over time, or that there was stratification and differential fertility with a small number of initial migrants, which is definitely not the case in Neolithic Europe where the migration was much more punctuated and previous cultures were virtually leveled to relatively simple societies after the Chalcolithic, which is visible for example in the largest settlement sizes at each period.

Davidski said...


Do not post bullshit here. I will delete it as soon as I see it.

Al Bundy said...

@Nirjhar I read your post about Mallory and,as I said, the early split of Anatolian can't be explained yet because that's before Yamnaya and some of the other steppe groups.Anthony thinks Vucedol might have have something to do with it.All the recent papers have been about the spread of IE but for the origin we need to go further back to be sure about things.

Davidski said...


No more bullshit. I will delete it as soon as I see it.

bmdriver said...

Why is it bullshit David, why?

Let other people make there own mind up, instead of censorship.

.......but explain why it's bulkshit. Are you denying white Christians invaded, enslaved then converted those people and places to be subservient to white Christian rule?.....

Use your blue eyed blonde haired Aryan genes to give me an answer.

Indian migrants went out started satellite outposts in west Asia, Middle East and steppes, and some of those came back. And that authorship was changed by the cult of Abraham, to substantiate the bibical belief that civilisation started in west Asia, Middle East, and then spread by force to enslave the pagans and sons of Canaan. Aka Aryan theory.

Pagan history verses abrhamic Aryan theory.

Davidski said...


Let other people make there own mind up, instead of censorship.

I've tolerated enough nonsense from you and others of your ilk in these comments.

I need to clamp down on it now, because I don't want the comments section cluttered with bullshit when the new papers come out and it gets very busy here.

Think very carefully about what you post. It has to be logical and coherent and free from any insanity, otherwise I'll get rid of it as soon as I see it.

Your post above doesn't qualify, but I'll leave it. After this post, no more.

Chad Rohlfsen said...

No data will ever quiet some people. They will always say, "look here or there", or "it can't be so". Always with nothing to back their assertions. When the data comes out, get the programs and test it yourself. Your opinions mean nothing in the face of sound statistics. If you think it's wrong, prove it or put a sock in it.

EastPole said...

“You may know this famous line of RV. 1.164.46

46 They call him Indra, Mitra, Varuṇa, Agni, and he is heavenly nobly-winged Garutmān.
To what is One, sages give many a title they call it Agni, Yama, Mātariśvan.”

Yes, I know RV. What do you think about Greek mystery religions, Orphic Hymns like this one:

OF 543: “One Zeus, one Hades, one Helios, one Dionysus
one God in all them: why shall I tell you in two ways?”

Looks like there was a common source for Vedic, Hellenic and Slavic religion, poetry, philosophy etc.

Salden said...

bmdriver's babble just the old Critical Theory that replaces the Bourgeois and Proletariat with Whites/White Men/Straight White Men/Christian Whites and Whoever. Alongside a dash of Hindu Nationalism.

aniasi said...


Do we know the dating of these graves? If later, I suspect they will look closer to Andronovo, if earlier, then closer to the Poltavka outlier you mentioned a few months ago.

I don't think this is proof of an Indo-Aryan invasion, though it does prove a migration. We don't see mass signs of violence, and surprisingly, Meluhhas trading partners in Mesopotamia do not write of the invasion.

Davidski said...


Do we know the dating of these graves?

Like I said in the post, Late Bronze Age to Iron Age. The Padova Uni PDF I linked to says 1400 to 900 BC.

So some of the earliest graves from that site should contain skeletons with very high levels of steppe ancestry, while the later ones should be more similar to modern-day Kalash and/or northern Pakistanis.

They might create a nice cline on my PCA running from the Steppe_EBA cluster to modern-day Pakistanis.


If these samples do have a lot of Yamnaya-like admixture, it's going to be hard to argue that such ancestry came from the steppe. Unless one argues that South Asian women importing young males from the steppe as "stallions" to produce offspring. Which would be another weird theory to add to the collection.

We'll be dealing with ample genome-wide, Y-DNA and mtDNA data. It'll be very easy to show that these people had recent ancestors from the Bronze Age steppe if that's the case.

The idea that it's possible to somehow confuse native South Central Asian ancestry with admixture from the steppe to the extent that you're claiming is total bullshit and you know it.

Taymas said...

Nirjhar and Alberto,

Right, I understand that "Those burials have no relation to any hypothetical Aryan Migration" is a hypothesis coming from archaeological site analysis. I'm not even arguing it, I don't know enough to have an opinion on it. What I'm asking is: do you guys have a test in mind (I understand that you're not researchers and don't have the resources to implement such a test yourself, I'm speaking hypothetically, something for us to watch for) that would support or contradict that hypothesis?

"Those people were not ancestors of Indo-Aryans BTW , they were most likely Proto-Kafirs" seems a separate claim.

Rob said...

Well as long as they also have pre-2000BC samples from the ~ same region, and I suspect they would, then we can belt out the details..

MaxT said...

Looks like this 'swat valley' archaeological site is actually called 'Gandhara Grave culture'.

"It has been regarded as a token of the Indo-Aryan migrations, but has also been explained by local cultural continuity."

"According to Parpola, in the centuries preceding the Gandhara culture, during the Early Harappan period (roughly 3200–2600 BCE), similarities in pottery, seals, figurines, ornaments etc. document intensive caravan trade between the Indian Subcontinent and Central Asia and the Iranian plateau."

"According to Kennedy, who argues for a local cultural continuity, the Gandhara grave culture people shared biological affinities with the population of Neolithic Mehrgarh."

"This is contested by Elena E. Kuz'mina, who notes remains that are similar to some from Central Asian populations."

My guess is they will be turn out to be mix of steppe migrants + local population. Kalash, Nuristanis, Tajiks will most likely be closest population if these burials turn out to have high-steppe ancestry

Aram said...


Thanks for that citation. When Armenia was converting to Christianity the monks were saying that we are returning to our ancestral One God religion. That why the conversion was called 'Return'.

Jijnasu said...

Lol, neither the religion of the Indo-aryans nor the other IEs was monotheistic in any meaningful sense. (except to some extent that of the zoroastrian iranic people). Seems more like attempt by modern IEs to graft their religious views on their ancestors. Identifying a unity of the gods does not imply that they were monotheistic. Kathenotheism seems a better description. As an example an early post-vedic hymn to surya (the sun god) reads ' he is brahma, vishnu, shiva, skanda, prajapati, indra, kuvera (a god of wealth), time, yama ( the god of death), soma ( the moon), the lord of the water (varuna), pitrs (the forefathers), vasus, sadhyas (classes of minor gods), the ashvins (twin horse headed physician gods), maruts (storm gods), manu, vayu (the wind god), vahni (fire), and the lifeforce.' This doesn't mean that surya would be thought of as a water god, or wind god or that the ancestors would be worshipped with hymns addressed to surya or that it negates the tradition that identifies manu,yama and the ashvins as sons of surya.

Nirjhar007 said...

seems a separate claim.

And for a special claim you need equally special data . Not just from aDNA but surely others . You will find that in the article.

EastPole ,

Yes nice passage again. Apparently the IEs knew about an ultimate force behind all the divine aspects !.


I see .

Nirjhar007 said...

Chad ,

I don't know , the camp you guys are representing may get victory in future , but you are still an educated ass hole....

EastPole said...

“This doesn't mean that surya would be thought of as a water god”

But Agni who is identified with Savitar or Surya is also identified with water goddesses Ida and Sarasvati in RV.2.1.11:

“11. You, god Agni, are Aditi for the pious; you, as Hotrā Bhāratī, are
strengthened by song.
You are Iḍā, bestowing a hundred winters in return for skill; you, o lord
of goods, as obstacle-smasher, are Sarasvatī.”

Not kathenotheism but monotheism, one God with many manifestations, many faces. Agni is One God but also is a “Hotar of many faces” (RV. 6.5.2.), is represented often as a policephalic deity , and is identified with many other god which are his manifestations.

But the same was observed among Slavs.
The 12th-century German missionary Helmold of Bosau recorded in Chronica Slavorum about Slavic religion:

“They also carve out many deities with two, three, or more heads. But they do not deny that there is among the multiform godheads to whom they attribute plains and woods, sorrows and joys, one god in the heavens ruling over the others. They hold that he, the all powerful one, looks only after heavenly matters; that the others, discharging the duties assigned to them in obedience to him, proceeded from his blood; and that one excels another in the measure that he is nearer to this god of gods.”

One god, not participating in the affairs of this world, which is ruled by his emanations or progeny, whose closeness to the father is the measure of their excellence.
The highest God is without evil, without duality, unmoved, not participating.

But the same we see in Greek mystery religions and later in the philosophy of Pythagoras, Plato, Aristotle or Plotinus.
Greek mystery religions came to them from the North, from Thracians and Hyperboreans.

It is not a PIE religion, as described by Dumezil, but later development originating north of the steppe and expanding from there. It explains all common things in Slavic, Greek and Indo-Iranian poetry, religion, philosophy etc. PIE theory doesn’t explain it.

Nirjhar007 said...

EastPole , this post of mine is rather a rough draft . But it speaks on Polycephalic aspects of divinities :

It will be great have your opinion there ! :) .

Alberto said...


Do you mean some kind of genetic model? It would have to be theoretical without ancient DNA, but let's see if this is more or less what you mean.

If you go to Lazaridis et al. 2016 and download the supplementary information here:

You can see that Iran_Neolithic is better modeled (using ADMIXTUREGRAPH) as 62% Basal Eurasian + 38% EHG (Figure S4.11, page 36).

And the theory is that as you get away (north or east) from the Arabian Peninsula the amount of Basal Eurasian should go down (and therefor EHG should go up). We have Iran_Hotu sample, from the South Caspian area (Mesolithic) to check this. On page 70, you can see:

"We can model Iran_HotuIIIb as 90.6±4.4% Iran_N and 9.4±4.4% EHG, with O9ALNW as outgroups (P-value for rank=1 is 0.45)."

And then all the modern populations from S-C Asia their West Eurasian part can be best modeled in the same way: Iran_Neolithic (or Late_Neolithic) + EHG. You can check this on Table S9.4 (page 127). For example, the Kalash are modeled as 46.8% Iran_LN + 31.7% EHG + 17.3% Onge + 4.2% Han. And this is true for South Indian Dravidians too. Mala (los caste Dravidian from Andhra Pradesh) are modeled as 28.1% Iran_LN + 11.9% EHG + 56,7% Onge + 3.3% Han.

And this is what I get using Global_10 PCA data (weighted values):

Iran_Neolithic:I1290 60.6 %
Karelia_HG:I0061 27.9 %
Agta 7.6 %
Yamnaya_Samara:I0231 3.9 %
Andronovo:RISE505 0 %
Han 0 %

Distance 0.005879

(There is no Onge, so I use Agta as the closest thing).

Iran_Neolithic:I1290 51 %
Agta 38.4 %
Karelia_HG:I0061 10.6 %
Yamnaya_Samara:I0231 0 %
Andronovo:RISE505 0 %
Han 0 %

Distance 0.019396

Iran_Neolithic:I1290 53.4 %
Agta 37.6 %
Karelia_HG:I0061 9 %
Yamnaya_Samara:I0231 0 %
Andronovo:RISE505 0 %
Han 0 %

Distance 0.021388

Agta 55.6 %
Iran_Neolithic:I1290 40 %
Karelia_HG:I0061 4.4 %
Yamnaya_Samara:I0231 0 %
Andronovo:RISE505 0 %
Han 0 %

Distance 0.023225

So you see there is a pattern there. The West Eurasian side of all South-Central Asians can be modeled as Iran_Neolithic + EHG. Which can create a pseudo-Yamnaya admixture (actually, in that same Table S9.4 you can see Mala also modeled as 23% Iran_N + 18.4% Steppe_EMBA + 55.2% Onge + 3.4% Han).

So yes, there are confounding factors there that without ancient DNA cannot be ascertained. I agree with Rob and Davidski above that with enough ancient DNA it will be clear and easy to see the difference, so I just hope that these samples (or better if they have more than just these ones) can clarify the situation completely, so we can finally know the truth and close this debate. Which is what we all want.

P Piranha said...

An issue you face when you use the purest and oldest unadmixed populations to model a modern population is overfitting. Because a population is always going to differ, by chance, from some combination of admixed populations, in other words no population is actually nothing else other than Yamnaya plus Agta plus Iran Neolithic but all have small fractions of admixture from elsewhere, or have slight deviations such as a few percents of CHG or old ANE such that they cannot be modeled exactly as a combination of those two populations plus the admixed Yamnaya, the best fit will always involve a combination of the purest unadmixed populations, the percentage of each fine-tuned to produce the exact combination that results in the best position in variation space.

If you attempt to model Central Asian and Siberian populations, for example, the best combinations are always Karelia Hunter Gatherer plus some Middle Eastern Ancient plus East Asian and ANE, as opposed to those proximate populations such as Scythians that we know for a fact actually contributed to their formation.

On the other hand, PCoA and other fst-based fittings are more sensitive to shared drift between recent admixed ancestors and their modern-day descendants, and don't make the same kinds of mistakes, and in those the pattern for EHG doesn't hold; South Asians quite clearly derive some EHG ancestry from Steppic populations.

Its still the case that only ancient DNA can confirm this, more so if samples before 1700BC or so to be there as well.

Rob said...

When adding less distantly ancient sources, I think Alberto's reasoning stands: as I pointed out several times, the best and most consistent fit across south Asians is Srubnaya Outlier.
And that's because it can be modelled as a very Eastern EHG + Iran Neolithic. Perhaps it would recall Chalcolithic central Asians, which
could be the actual source of admixture for south Asians.

Balaji said...

Exciting times indeed! The dating of these samples (1400 to 900 B.C) includes several hundred years after the supposed Aryan invasion began. If indeed there was such an invasion, we should expect the ASI to be fairly high in the earliest times and decreasing at later times with more invaders coming in diluting the ASI.

Moorjani et al. estimated a mixing date for the Pathans of 2100 years ago. See Table 1 of their paper.

If as I expect, the Reich lab find that the Swat samples have much less ASI than do Pathans and Kalash, this will not be good for the Aryan Invasion Theory.

Davidski said...


If as I expect, the Reich lab find that the Swat samples have much less ASI than do Pathans and Kalash, this will not be good for the Aryan Invasion Theory.

I see you're already in damage control.

Yep, the earliest Swat Valley samples will probably carry very little to no ASI. Some might even be identical to Bronze Age East Europeans/Central Asians.

Iran_N/ChL-related and ASI ancestry will probably be much higher in the later Swat Valley samples.

Why? Because it usually takes generations for migrant populations to blend with the locals, and the pre-Bronze Age locals in this case will turn out to be a mixture of Iran_N/ChL and ASI.

postneo said...

Also Pagani's recent abstract talks about an ancient "steppe like" gradient from Iran westwards. Not clear what that means. If we imagine CHG in the caucasus and ANE stretching btw the Pamirs and Altai...

then you get similar CHG + ANE in the south east caspian and also north east caucasus black sea where CHG combines with a forest steppe component carrying ANE.
These populations could have some shared drift primarily due to CHG. Is this borne out by data?

higher ASI in modern Pathan will perhaps be explained thusly: AIT happened and then there was a resurgence of ASI .... of course it makes no sense but so what?

My opinion is that whatever ASI/Steppe/ANI contribution occurred would have had to have happened during the IVC.

Davidski said...

Pagani's claims are bogus.

And obviously there was no direct migration of CHG from the Caucasus to South Asia. South Asians have CHG-related ancestry from ancient Iran, and CHG ancestry from the Bronze Age steppe.

Of course, the Aryan migrants to South Asia and their immediate, recent descendants won't show any ASI. Why should they? Where would they pick up this ASI, in Eastern Europe or Central Asia? Why?

Only after the early Aryans mix with native South Asians will you see ASI in Aryan remains.

Postneo, you're not very good at this. Another, less intellectually demanding, hobby beckons.

Seinundzeit said...


Those are good points, and interesting models; perhaps, it might be the case that the best way to get around these issues (when we're using nMonte with PCA coordinates) might be the use of Iran_Hotu, as that sample seems to be around 85% Iran_Neolithic-related, but with an extra dose of approximately 15% native Central/South Asian ANE.

Also, I’ve heard the it’s mtDNA haplogroup can still be found in South Asia, while there is no mtDNA match between the Neolithic Iranian samples and contemporary South Asia.

With those two facts in mind, here is an exploration of South Asia and Central Asia, using Iran_Hotu and the Onge.

But before we begin, I do want to articulate a few thoughts (in the clumsy manner typical of me, lol).

For some time now, I've been using nMonte with Fst-based PCoA data (Matt really deserves our gratitude, for allowing us to add this sort of analysis to our general arsenal of methods), so I've become used to certain patterns which don't appear in the PCA data.

Also, I don't think that the Onge are useful proxies for ASI, especially after seeing the haplotype analysis in the Zoroastrian heritage paper. I'll add more on this, once I've posted all the results.

Basically, the following models are (in my estimation) not exactly on the mark. Still, I think they are of great interest, and once one is dealing with populations that are overwhelmingly West Eurasian in terms of genetic ancestry (upper caste northwestern Indians, many Pakistanis, Afghans, Tajikistanis, Iranians, etc), things become quite straightforward.

To be continued...

Seinundzeit said...

Continuing from where we left off...

Scaling was applied, and I also took advantage of some suggestions posted at Anthrogenica.

South India:


74.6% Onge
25.4% Iran_Hotu



64.2% Onge
35.8% Iran_Hotu



52% Onge
48% Iran_Hotu



51.5% Onge
45.2% Iran_Hotu
3.2% Srubnaya_outlier



50.6% Onge
49.4% Iran_Hotu


Piramalai Kallar

54.1% Iran_Hotu
45.9% Onge



55.35% Iran_Hotu
42.30% Onge
2.35% Srubnaya_outlier


Tamil Brahmin

54.4% Iran_Hotu
31.6% Onge
14.1% Srubnaya_outlier


Seinundzeit said...

North India:


49.2% Onge
47.5% Iran_Hotu
3.3% Srubnaya_outlier



48.7% Iran_Hotu
39.4% Onge
11.9% Srubnaya_outlier



50.6% Iran_Hotu
30.0% Onge
19.4% Srubnaya_outlier


UP Brahmin

49.6% Iran_Hotu
27.2% Onge
23.1% Srubnaya_outlier


Seinundzeit said...

South Central Asia:


38.25% Iran_Hotu + 11.90% Iran_Chalcolithic
24.90% Srubnaya_outlier
15.90% Onge
9.05% Mongola


Northeastern Sarbani Pashtun (KPK)

38.80% Iran_Hotu + 21.60% Iran_Chalcolithic
24.95% Srubnaya_outlier
14.10% Onge
0.55% Mongola


Northeastern Sarbani Pashtun (FATA)

35.3% Iran_Chalcolithic + 20.2% Iran_Hotu
30.8% Srubnaya_outlier
13.2% Onge
0.5% Mongola


Me (Pashtun; ancestry from Afghanistan, FATA, and KPK)

33.2% Iran_Hotu + 26.1% Iran_Chalcolithic
24.7% Srubnaya_outlier
12.8% Onge
3.2% Mongola


Ghilzai Pashtun (Afghanistan)

43.0% Iran_Chalcolithic + 13.6% Iran_Hotu
32.0% Srubnaya_outlier
11.5% Onge



47.80% Iran_Chalcolithic + 6.35% Iran_Hotu
32.90% Srubnaya_outlier
11.40% Onge
1.55% Mongola


Karlani Pashtun (Afghanistan)

50.8% Iran_Chalcolithic + 4.0% Iran_Hotu
34.5% Srubnaya_outlier
10.1% Onge
0.7% Mongola


Southwestern Sarbani Pashtun (Afghanistan; has 12.5% Hazara ancestry)

45.40% Iran_Chalcolithic + 6.35% Iran_Hotu
31.25% Srubnaya_outlier
9.80% Onge
7.20% Mongola


Southwestern Sarbani Pashtun (Afghanistan)

46.8% Iran_Chalcolithic + 16.2% Iran_Hotu
27.4% Srubnaya_outlier
4.9% Mongola
4.7% Onge



44.15% Iran_Chalcolithic
43.75% Srubnaya_outlier
10.45% Onge
1.65% Mongola



36.55% Srubnaya_outlier + 8.55% Sarmatian + 4.50% Poltavka_outlier + 2.05% Sintashta
40.75% Iran_Chalcolithic
5.95% Onge
1.65% Mongola



50.40% Iran_Chalcolithic
19.05% Sarmatian + 17.85% Srubnaya_outlier + 8.75% Poltavka_outlier + 0.40% Scythian_Samara + 0.05% Karasuk_outlier
3.50% Mongola


(With exception to Yaghnobi people, all South Central Asians tend to be between 5% and 15% Onge, with most Pashtuns at around 10% and most Pamiri folk at around 5%. This makes the Yaghnobi people quite unique; if we assume for a moment that ASI is Onge-related ENA, the Yaghnobi people are the only South Central Asians who lack this kind of ancestry. This is a consistent pattern; also appears when I use my ASI simulations).

Seinundzeit said...

General Observations:

Although I have reservations, I think the patterns here make sense.

Basically, even once we account for extra ANE (via ancient populations closely related to the Neolithic Iranian plateau samples), there is still a Steppe_EMBA-related impact in South Central Asia, best represented by the Srubnaya_outlier.

The Srubnaya_outlier is quite similar to Yamnaya, just much more shifted towards ANE on the WHG-to-ANE cline.

As a demonstration:


46.10% Karelia_HG
26.65% Kotias
26.50% MA1
0.75% Ulchi


Compared to Yamnaya and early Corded Ware (the Latvian sample)


44.9% Karelia_HG
40.2% Kotias
14.9% Loschbour



44.3% Kotias
34.9% Loschbour
20.9% Karelia_HG


Obviously, these are closely related populations, with Latvia_LN1 being the most genetically western, Srubnaya_outlier being the most genetically eastern, and Yamnaya being intermediate.

If we account for the fact that the Srubnaya_outlier was probably some special migrant from just east of the Urals (she was clearly a person of some importance, considering the circumstances of her burial), we are probably looking at a geographic cline among ancient IE populations from east to west.

As a sidenote, I included Iran_Neolithic and Iran_Chalc, but Srubnaya_outlier prefers CHG (just like Yamnaya), when one applies scaling to the data.

Anyway, I have a feeling that these Swat Valley Aryans will have loads of Steppe_EMBA ancestry, of the Srubnaya_outlier kind.

aniasi said...


To clarify, I meant more along the lines of what dating ranges do you think are most likely. As we know, samples have errors.

In terms of more newbie stuff, what do you mean by Kalash-like? Do you mean more Iran-Neolithic, or ASI?

Seinundzeit said...

Shifting gears; I have a few quick thoughts on ASI, after looking at some Fst distances, and after seeing the recent Zoroastrian paper.

Mainly, I’ve seen Fst distances for Ust-Ishim, and it turns out the closest living populations to him are North Indians, followed by South Indians. In addition, if my memory serves me right, Ust_Ishim apparently shared the most drift with Onge.

On top of that, in the Parsi paper, the South Indian cluster receives a hefty contribution from Ust-Ishim, even though the Onge (and East Asian ENA) are allowed as possible donors.

I think this is huge, because haplotype methods are supposed to be very sensitive, and they consistently fail to show an Onge or East Asian contribution to Indians, but rather show Ust-Ishim acting as “ASI” (the Iran_Neolithic paper showed the same result).

Also, as per Reich and Lipson, an adequate model of South Asians (in terms of deep ancestry) isn’t currently feasible, due to the complicated nature of Indian West Eurasian ancestry; so they’ve abandoned the cookie-cutter talk of simple ANI and ASI.

On top of that, we’ve only seen Y-DNA haplogroup K2e in South India.

In addition, the Paniya are dominated by Y-DNA F.

So, if we take into account the Fst closeness to South Asia above other contemporary populations, the haplotype-based results, and the Y-DNA data, I think a case can be made for ASI having some sort of relationship with Ust-Ishim-related populations.

And (as a reminder), as per the Reich and Lipson tree, Ust-Ishim was a very deeply diverged, and extremely basal, member of the broader West Eurasian/North Eurasian group.

Basically, I think there is some unique Eurasian ancestry in India (unique Eurasian ancestry which is perhaps closer to West Eurasia, rather than closer to ENA, in the deep phylogenetic sense).

All speculation though, until we see Upper Paleolithic/Mesolithic aDNA from inside India proper.

Nirjhar007 said...

Those samples are several centuries after the start of De-urbanization , so having ASI is always expected . The key is to find samples before 2000 BC .

Davidski said...


The date range for all of the burials is roughly 1400-900 BC. You need to read the paper if you're interested in the details.

Kalash-like = mostly Steppe_EMBA + significant Iran_N + minor ASI.

Davidski said...

The key is to find samples before 2000 BC.

SC Asian samples from before 2000 BC with be mostly Iran_N/ChL + ASI

Early Indo-Aryans will be mostly Steppe_EMBA + minor Iran_N/ChL and ASI.

Iron Age samples will look more or less modern-day SC Asian.

postneo said...

has anyone seen this preprint on Estonian CW I posted earlier?

Davidski said...

Of course.

capra internetensis said...


I think K2e ended up being the artifact of a bad primer. Do you mean pre-NO/'haplogroup X'?

Saqib said...


In Pakistan they have found skeletons from Mehrgarh which goes back to 7000BC and other IVC sites like Harappa or Mohejedaro. But I'm not sure if it's feasible to extract DNA from them? Maybe after Swat DNA.

"The postures of the skeletal remains in graves at Mehrgarh bear strong resemblance to those at Ali Kosh in the Zagros Mountains of southern Iran [19].[9] Clay figurines found in Mehrgarh resemble those discovered at Zaghe on the Qazvin plain south of the Elburz range in Iran (the 7th millennium BCE) and Jeitun in Turkmenistan (the 6th millennium BCE) [24"

Nirjhar007 said...

By De-urbanization I was talking about India BTW .

We will see all the aspects . We must remind ourselves again that during late phases on BMAC area Andronovo type settlements were observed . Which is natural because of the Geography.

Nirjhar007 said...

No Saqib . Pakistan is also well analysed , I heard a few days back .

Saqib said...


You mean no chance of DNA from Harappa, Mehergarh or Mohenjedaro?

Rob said...

@ Sein

"As a sidenote, I included Iran_Neolithic and Iran_Chalc, but Srubnaya_outlier prefers CHG (just like Yamnaya), when one applies scaling to the data"

Ah, okay, with weighted sets I get similar:

"Yamnaya_Samara:I0231" 64.9
"AfontovaGora3:I9050.damage" 28.9
"Itelmen" 3.5
"Iran_Neolithic:I1290" 1.4
"Paniya" 1.3
"Hungary_N:I1496" 0

Seems like a Steppe EBA mixed with 'local' central Asian groups.

Nirjhar007 said...

No! a very good chance :) , we are lucky ....

Saqib said...


wow, from which site? Harappa? :)

Nirjhar007 said...

I can't tell you that ! :) . But both India and Pakistan will give aDNA this year ...

Nirjhar007 said...

Now talking about pet theories , if BMAC is been analysed , I will expect R1a-Z2124 ad J2 from BMAC proper . I am certain also that with more sampling J2 will be found in Andronovo etc .

For Indo-Pak . The Civilization which had around 3-5 million people . I expect R1a, J2, L and R2a .

This year all will be tested ...

Seinundzeit said...


That's definitely a possibility.

Although, one big clue we have concerning the origins of the Srubnaya_outlier-related: the ancestry of Sarmatians and Western Scythians.

Basically, it seems that the primary differentiating factor (in terms of genetic ancestry) between Sarmatians/Western Scythians and the preceding Srubnaya Culture was an infusion of Srubnaya_outlier-related admixture.

Interestingly, the Eastern Scythians lack the Srubnaya_outlier affinity; the phenomenon is restricted to Sarmatians/Western Scythians.

I'd say that the process of admixture started before the Sarmatian/Scythian phenomenon; the Srubnaya_outlier was the biological mother of another Srubnaya sample.

Also, it seems that this individual held some importance in her society, which could provide some indications concerning why/how this gene-flow took place.

Since the Western Scythians/Sarmatians seem to have been consistently Srubnaya_outlier-admixed, while the Eastern Scythians weren't, I think parsimony would suggest that the Srubnaya_outlier was from much "closer to home" in relation to the Srubnaya people, rather than being all the way from Central Asia (which is why Steppe_EMBA people like her didn't leave much of an imprint on Eastern Scythians in north Central Asia).

So, at the moment, I don't think she was Central Asian.

Of course, I could be wrong. It all depends on where we eventually find samples like her.

capra internetensis,

I'm totally out of my element when it comes to uniparental data (you're the expert here), but if my impressions weren't incorrect, wasn't pre-NO/haplogroup X renamed K2e?

Slumbery said...


Speculative, but I'd say that the lack of Iran Chalcolithic ancestry in Indo-Aryan groups questions if an old favorite of some, BMAC, had any role in the formation of the Indo-Aryans whatsoever. It looks like the Aryans went trough that region quickly and pretty much bypassed them. Unlike the Iranian groups that mixed with Iran Chalcolithic people all over.

Seinundzeit said...


I completely agree; going by the current evidence, it really doesn't seem like BMAC had much of a role to play in the ethnogenesis of the Indo-Aryans, as the lack of Iran_Chl (or, at any rate, the much lower affinity towards Iran_Chl) is the most striking difference between upper caste North Indians/Indo-Aryan Pakistanis and Iranic South Central Asians.

Although, maybe interaction with native Central Asians did transform the Indo-Aryans in terms of culture/ritual/religion?

Perhaps, it could be that cultural influence from native Central Asians might even explain the unique burial practices ("unique" in comparison to steppe IE peoples) seen with these ancient Swat Valley people? Perhaps, we're looking at a case of cultural exchange without much gene-flow?

On the other hand, for reasons we don't currently understand, it could be that the Iranic-BMAC interaction involved much stronger entanglements, and thus significant gene-flow?

I mean, looking at the data which we currently have, these seem to be sensible ideas.

And who knows, differential levels of BMAC influence might even explain the ancient Iranic/Indic cultural rivalry that some people infer from Avestan and Vedic textual evidence?

Rob said...

@ Sein

RE: Srub Out; yes I agree actually. Her U5a suggests a Boreal northeast Euro/Russian Meso-Neolithic origin.

@ Sein & Slumberry

I have noticed the same. A distinct layering of non-steppe west Eurasian ancestry might be detectable, which helps differentiate them; an addition to the different levels of ASI ancestry.

So for Indian groups Iran Neolithic is clearly preferrable.
"Paniya" 43.25
"Iran_Neolithic:I1290" 26
"Srubnaya_outlier:I0354" 22.1
"Srubnaya:I0431" 8.65

For Kalash it's similar, but a Caucasus signal comes through (Kotias + Armenia EBA)

"Srubnaya_outlier:I0354" 30.4
"Iran_Neolithic:I1290" 27.5
"Paniya" 18.6
"Armenia_EBA:I1635" 9.75
"Kotias:KK1" 8.75

For Iranian Iranics, it is Iran Chalcolithic

"Iran_Chalcolithic:I1665" 62.25
"Srubnaya:I0431" 15.05
"Jordan_EBA:I1706" 9.05
"Srubnaya_outlier:I0354" 8.85
"Paniya" 3.1

"Iran_Chalcolithic:I1665" 57.55
"Srubnaya:I0431" 26.95
"Jordan_EBA:I1706" 7.85
"Iran_Neolithic:I1290" 4.6
"Paniya" 2.75

For other Iranics, it seems it is more complex; ? or layered, with variously different proportions of Arm EB, Iran Neol and Iran Chalc proportions. Other analyses might be needed to see if/ what it means.

"Srubnaya_outlier:I0354" 40.35
"Armenia_EBA:I1635" 19.05
"Iran_Chalcolithic:I1665" 16.85
"Paniya" 15.1
"Kotias:KK1" 6.6

"Srubnaya_outlier:I0354" 28
"Armenia_EBA:I1635" 24.25
"Paniya" 16.1
"Iran_Neolithic:I1290" 14.1
"Iran_Chalcolithic:I1665" 10.65
"Srubnaya:I0431" 5.2

Coldmountains said...

Very interesting article and discussion. Also Vedic Hinduism seems to be in many ways more archaic and steppe-like than early Zoraosterianism. The burial practises of Zoroastranism seem to be BMAC derived. Maybe this also correlates with less BMAC admixture among Early Indo-Aryans. The Vedic Dasas and archenemies of Vedic Aryans could be Bmacized Iranic tribes of Central Asia and not IVC people like many thought in the past.

Rob said...

What means is difficult to say. But it sounds a bit odd that Indo-Aryans just 'jumped" through BMAC without intermixture, whilst Iranics mixed. Whilst such a scenario is possible, I wonder if it means something different - such as that BMAC being more like Iranic Neolithic & 'steppe'; and Iran Chalc & K-A type ancestry being a later eastern drift.

Shaikorth said...


Ust-Ishim haplotype contribution appeared in Indians (alongside Iran_N so presumably as an ASI proxy) and Native Americans (mainly as a stand-in for ANE since the common fit was U-I + Han) in Broushaki 2016 too. I think the best way to reconcile this with other data is that it happens because Ust-Ishim is less temporally distant from the ASI/Han/Onge common ancestor than Han or Onge are (for this thought experiment let's say it has Z years of separation from the Crown Eurasian ancestor while ASI/Han/Onge node has 2Z and Onge has 8Z years from that, Z+2Z < 8Z) and haplotypes are sensitive to this. With Native Americans the alternative for U-I would have been Loschbour, resulting in a similar situation.

More support for this is that when ancient donors were allowed in the Broushaki paper, Onge was modeled as Ust-Ishim with pretty much nothing from modern India, Southeast Asia or Oceania. With only ancient donors so was Han as it wasn't allowed to be a donor for itself. I think Kostenki14 and MA-1 would also get very high Ust-Ishim with the limited set of high coverage ancients now available.

Seinundzeit said...


You make very important points, and I definitely agree that the possibilities you've outlined could easily explain things; although, it isn't just the haplotype evidence, but the Fst affinity (the closest living populations to Ust-Ishim, going by Fst, are all Indians. Interestingly, using Fst, East Asians/Australians/Onge are quite distant to Ust-Ishim. He's actually closer to BEA-rich Near Easterners, compared to ENA).

When subjected to the PCoA technique, having Ust-Ishim in the mix becomes essential for West Asians, South Central Asians, and Indians; Paniya turn out to be 25%, and it goes down to 5%-3% in South Central Asia, and is found at trace levels in West Asia/the Caucasus.

In addition, I always thought it was strange that people like Paniya and Pulliyar happen to be so rich in Y-DNA F; in terms of Y-DNA, South Indian tribal people are the most "Basal Eurasian" populations around.

Not to mention the relatively recent finding of K2e in South India, and the fact that C1b1a is found at low frequencies across West Asia, Central Asia, and South Asia, but is most common in South Asia.

Finally, I feel that it's quite telling that Reich and Lipson don't feel confident enough to model Indians in terms of very deep ancestry, and claim that the West Eurasian side of Indian ancestry is too complicated to be properly explicated at the moment (we don't have the adequate basal West Eurasian/North Eurasian references for Indians).

Adding all of this together, I think it is a plausible scenario (mainly: unique Eurasian ancestry in India; somewhat closer to ANE/WHG/Upper Palelithic Europeans than to ENA in terms of deep phylogenetic placement, and with some declining percentages into Central Asia and West Asia).

Of course, definitely a speculative proposal on my part; I'm certainly not married to the idea.

Honestly, when it comes unraveling ASI, I think only an Indian genome of Upper Paleolithic/Mesolithic provenance will give us clarity.

In this respect, the Paniya are a good population for these kinds of analyses. But, they have anywhere between 25% to 45% West Eurasian admixture (depending on the method, and depending on reference populations), so I still use the Onge, even though I think it's problematic.

With the Onge, ASI seems to be at 40%-50% in South India, and at a range of 5% to 15% in South Central Asia (Tajikistan, Afghanistan, and northern/western Pakistan), with around 0%-7% in Iranians (showing a peak in Bandari Iranians, with the other Iranians closer to 0%-2%).

Then again, who knows, maybe even those Rakhgarhi samples might help to clarify the nature of ASI? (That would be exceedingly awesome)