I want to make a quick post on a new article by Nezih Seven. You can find it here.
A Genetic Analysis of Historical Population Movements Around The Zagros Mountains
He analyzed the NW Iranian samples from the 'Southern Arc' paper and reached some interesting conclusions.
 |
| Model for Hasanlu_LBA_A by Nezih |
The p-value of the above model is ~0.20 (therefore passing as p>0.05). His qpAdm output file can be seen here.
Armed with this output, he concludes that the Hasanlu LBA and IA people were Mitanni.
This is most likely an Indo-Aryan (Mitanni) migrant from Central Asia. Although the authors of the paper argue against an Indo-Iranic presence in Hasanlu, the lack of Y-DNA R1a, which is their main argument against the Indo-Iranic presence, does not prove anything especially considering the quite low R1a presence among the Iron Age Swat Valley samples. Therefore;1. There is no sufficient evidence to think that heavily BMAC-derived Indo-Aryans were all paternally steppe descendent and2. The absence of the abovementioned evidence (R1a) is not the evidence of the absence of Indo-Iranics. Also we have additional reasons to think otherwise: well-known names of the field like Parpola, Dumézil, Kuzmina, C. Young, Francfort and Kurochkin all have interpreted some of the foundings from the site as Aryan-related (either Indo-Aryan or Iranic).
He further models DinkhaTepe_B and Hasanlu_IA samples with Hasanlu_LBA_A as a source. I agree with this and in my Southern Arc post, I had come to a similar conclusion.
Thus we may conclude that there were at least three different genetic clusters in the Early Iron Age northern Zagros: DinkhaTepe_BIA_A-like (Hurrians?), Hasanlu_LBA_B-like (probably speakers of something related to the unknown language of their mostly local predecessors) and DinkhaTepe_BIA_B-like (Matienis?). Since it is possible for the earliest Iranic migrants to the region to be genetically similar to Hasanlu_LBA_A, it might be impossible to detect them solely based on autosomal DNA, but it might well be the case that they had already began penetrating increasingly; especially considering that the generally accepted entrance date for them is somewhere between 1200 and 900 BCE.
I have no issue with the claim that these people were the Mitanni or Indo-Iranian in general. In fact, I have made such a claim myself here, quite emphatically at that. I also agree with him that R1a or not cannot decide whether a population spoke Indo-Iranian, only material culture can. This R1a glorification needs to stop, it reeks of Eurocentricity. That R1a and R1b males explosively spread steppe autosomal ancestry in Europe does not mean that the same is applicable to Asia.
The issue with Nezih's claim is that the qpadm model is faulty, and I will prove this later. I have known Nezih to be a strong believer in the Sintashta Indo-Iranian origin theory and I suspect that has led him to be partial to one particular model.
There are various reasons (apart from good qpAdm models which reject Sintashta as a source) why Sintashta ancestry cannot be a source for the Hasanlu samples.
1. The LBA_A sample itself has Y hg J1a2a1b1a~(Oxus/BMAC related) and mtDNA R2 which is generally related to Iran Neolithic ancestry. Neither of these is Sintashta related.2. There is no Sintashta-related R1a (frequency ~100% at Sintashta) at Hasanlu LBA/IA, or at Dinkha Tepe. Out of the 16 samples dated to LBA or IA at these two locations, there are 0 Sintashta-related R1a samples but 10 Yamnaya-related R1b samples. It is clear that the steppe ancestry in these samples is from its immediate north (yamnaya-related) across the Caucasus mountains rather than from Sintashta.3. There is no Sintashta-specific mtDNA at either of the two locations (spreadsheet here). Of all the samples at Sintashta, the only mtDNA that is common with NW Iran is from a Sintashta_outlier sample (I0983) with mtDNA J1b1a3. However, J1b1a3 is first found at the BMAC site of Gonur (I6118). The oldest J1b1a1 is from Geoksyur Eneolithic, and is also very frequent at Geoksyur, suggesting that the common maternal ancestor J1b1a was from the SC Asian region.4. Nezih tries to hide this by giving the example of Swat valley samples, but the fact is that 2 R1a samples were found from Swat_IA, along with multiple samples with Sintashta-related mtDNA markers (Narasimhan et al, 2019 supplement pg 318-321). So these cases are not equivalent. Sintashta related contact with Swat is without question but the same cannot be said for Hasanlu or Dinkha Tepe because not a single such connection is found there.
qpAdm models for Hasanlu_LBA_A
The sample id and label list that I used are available here.
The first order of business is to validate Nezih's qpAdm model. He runs the model with Sintashta and Dzharkutan_BA (BMAC bronze age) as sources and with 25 populations on the right (as references). First, the program qpfstats() from Admixtools 1 is run on a Linux/MAC system with a list of all these labels. The f4 stats output is then passed as input to the qpadm() program with the parameters 'allsnps: YES' and 'inbreed: NO'.
The left populations (sources) are the same in our runs - Dzharkutan_BA and Sintashta.
Of the 25 right/reference populations, I exclude Tarim_EMBA as the latest Harvard v54.1 does not include them so I can't use them. Instead of Cameroon_SMA.DG (African outgroup), I use Mbuti.DG (African pygmy). The other difference is that I club Hasanlu LBA_B and Hasanlu_MBA under a single label Hasanlu_MBA since both these labels have the same ancestry (This is also confirmed by Nezih in his post and also in my Souther Arc post. I did this to reduce qpfstats runtime). So, instead of 25 reference populations used by Nezih, I use 23. All the other labels are the same as those used by Nezih. Within a given label, there might be minor variations depending on the number of samples used in each label, and I have not checked for those.
None of these minor changes should affect the qpadm() output, and good models are resilient to such minor variations. They certainly should not make good models fail. Good models should be robust to different sets of reference populations.
With the above setting, the qpAdm() output of Nezih's model gives me the same coefficients as his model.
Dzharkutan 85%
Sintashta 15%
However, the p-value is 0.0128, and therefore Nezih's model fails. The output file is here.
I ran another model but replaced Sintashta as the source with Armenia_MBA and added Sintashta as the 24th reference population.
This model passes with a p-value of 0.30. Armenia_MBA 32% + Dzharkutan_BA 68%. The output file is here.
Lazaridis et al 2022 show that Armenia in the MBA saw an increase in Yamnaya-related ancestry. There are multiple R1b found from Armenia_LBA sites as well. Therefore, the steppe and R1b in Hasanlu_IA can be best explained via Yamnaya-->Armenia-->NW Iran rather than via Sintashta-->BMAC--->NW Iran. This is relevant because Yamnaya/Armenia cannot be the source of Indo-Iranian languages and no mainstream theory proposes such a model. The mainstream theory pushed by Harvard considers Sintashta culture to be the origin and carrier of Indo-Iranian languages, and Harvard also seems to possess some special technology which allowed them to verify that all ancient R1a-Z93 descended men spoke Indo-Iranian because they sometimes use Indo-Iranian and 'R1a descendants' interchangeably.
Simpler static qpAdm() models
The Sintashta + BMAC model fails even with just 10 right populations with a p-value of 0.015. Output here.
However, the Armenia_MBA + BMAC model passes with the same 10 references. P-value is 0.37. Output here.
So regardless of the number of references, the Armenia_MBA+ BMAC model is always more robust than the Sintashta model.
Rotating qpadm() models
The major issue with Nezih's analysis is that he does not seek to compare various models for Hasanlu_LBA_A, rather he seeks to push a particular model which he favours. The unbiased way to do this is to run various rotating models as proposed by Harney et al, 2021. In this setup, various populations which could be likely sources are chosen, and when some of those are not in the source list of the qpadm model, they are put in the reference population. So, if we have a target and a list of 25 source populations and we want to test all 2 source models, we will have a combination of 300 2-source models (25C2). For each of the 300 qpadm runs, the other 23 populations are put in the reference list.
The idea is that if any of the populations in the reference list is a true source of the target, then this affinity should make the model fail (via statistically significant generated Dstats) unless that true source is placed in the source list. This allows for head-to-head competition between possible sources, ceteris paribus.
A slight improvement to this model is the concept of a fixed right list. The populations in this list always sit in the reference list and are not tested as a source (because they are old neolithic populations, or African outgroups, and need not be tested as a potential true source anyway). This allows us to reduce the number of combinations to be tested and save time while allowing us to use the discriminatory power (in choosing the correct source) provided by the populations in the fixed right list. This concept is also independently implemented in the qpadm_rotation() function from the Admixtools2 R version, although I wish I could take credit for using it as part of my analysis before Admixtools2.
The results of the rotating runs along with the source and reference list are provided here. A snapshot of the results is given below.
 |
| Hasanlu_LBA_A rotating models. Models with p>0.05 (green) pass |
 |
| Rotating models without Sarazm as a source. Models with p>0.05 (green) pass |
The rotation includes Sarazm Eneolithic as a potential source. P-values>0.05 are green, and below 0.05 are rejected models in red. We can see that either 3600BCE Sarazm, 2500BE Shahr-Sokhta-BA1 from East Iran, or 2000BCE Bustan_BA (BMAC) are selected as sources. The second source is local - Hasanlu_LBA_B, MasisBlur Neolithic, DinkhaTepe_IA_1, Hajji_Firuz_N or a steppe-rich source like Armenia_MBA. However, Sintashta IS rejected as a source.
In the second run, I removed Sarazm Eneolithic from the list of possible sources because it is 2000 years older than the Hasanlu samples. Even in this case, Sintashta is rejected as a possible source.
The working models that remain are Armenia_MBA + Shahr-Sokhta / Bustan_BA.
Again, the R1b in Hasanlu_IA samples favours Yamnaya/Armenia as a source rather than Sintashta culture and this is validated by qpAdm as well.
CONCLUSION
While I enthusiastically accept Nezih's proposition that the Hasanlu samples signify the presence of Mitanni/Indo-Iranians in the region, the claim of Sintashta ancestry in the samples fails to hold up under closer scrutiny. My conclusion that Mitanni / Indo-Iranians are unrelated to the Sintashta culture has not been forced to change.
Related
REFERENCES
Seven, Nezih. “A Genetic Analysis of Historical Population Movements Around the Zagros Mountains.” 26 Nov. 2001, nezihseven.substack.com/p/a-genetic-analysis-of-historical.
Lazaridis, Iosif, et al. “The Genetic History of the Southern Arc: A Bridge between West Asia and Europe.” Science, vol. 377, no. 6609, 2022, https://doi.org/10.1126/science.abm4247.
Éadaoin Harney, Nick Patterson, David Reich, John Wakeley, Assessing the performance of qpAdm: a statistical tool for studying population admixture, Genetics, Volume 217, Issue 4, April 2021, iyaa045, https://doi.org/10.1093/genetics/iyaa045
Narasimhan VM, Patterson N, Moorjani P, et al. The formation of human populations in South and Central Asia. Science. 2019;365(6457):eaat7487. doi:10.1126/science.aat7487
Personally, I'm not Indo-Aryan nor am I interested in South Asian culture beyond the Vedic period.
ReplyDeleteI'd have to say Nezih isn't that wrong, he just has a selective approach to the data. Let's assume his model is correct. Why is that 15% Sintashta and not the 85% BMAC ancestry supposed to prove that Sintashta was Proto-Indo-Iranian? His claim :"Hasanlu_LBA_A is most likely an Indo-Aryan migrant from Central Asia" is 100% percent wrong however, as the current consensus is that the Mitanni arrived in the18th BCE at the latest. At best he is an Early West Iranian.
I find that the blogosphere is only following really outdated works of Kuzmina, Mallory, Parpola and Sarianidi. I don't know how you can argue with current aDNA and archaeological works of Vidale, Bonora..etc that Indo-Iranian comes from Sintashta.
As someone who previously accepted the "Yamnaya Horizon" and Maykop models that no longer work I always believed that Indo-Iranian couldn't be from Sintashta and so took a west to east route. The reality is that there is nothing in Sintashta that is related to Proto-Indo-Iranians that isn't found elsewhere. On the other hand, Anahita and Mitra-Varuna with their thousand pillared palaces cannot be derived from Sintashta. The fire altars, the Soma, the Garuda...etc are all 3rd millennium South Central Asian innovations from previous traditions. Also, the Sintashta "chariots" do not resemble the Mitanni ones as per Anthony as they were used in close quarters with javelins (mind you there are claims they were only for ceremonial and hunting use).
The Western Scythians, who are used in various circular explanations, are the earliest attested Indo-European group in the good ole steppe. Yeah, an 8th century bce group that mostly worshipped female deities and had enerees priests that culturally barely resemble any other known Indo-European group is used as solid proof of Sintashta Aryans as if they don't have a good chunk of BMAC related ancestry.
Also, the Sintashta model doesn't work based on simple time and space constraints Van Koppen, Von Dassow, and Jesper Eidem all put the appearance of the Mitanni in the Zagros at the early 18th century BCE. So Proto-Indo-Iranian splits 2000 BCE and "Proto"-Indo-Aryans (who somehow have the sapta > satta change like their Sindoi cousins) magically teleport to Urmia bypassing and leaving no trace in Hissar, BMAC, Helmand, Jiroft, Elam...etc. For example, even Wouter Henkelman who wishes to Elamize everything including Cyrus can't deny the early arrival of the Persians: "Given the above framework, one cannot but conclude that the (Indo-)Iranians, who are assumed to have emerged on the plateau sometime between 1,500 and 1,000 BC, must have been exposed to Elamite culture for some 500 to 1,000 years prior to the rise of the Achaemenid Empire". And Gathic is much older than even the written Old Persian of Behistun which is by the way much older than the Old Persian spoken in the Elamite and Achaemenid periods.
It fails on every aspect, onomastics, archaeology, DNA, and the laws of time and space . It's a bogus theory that came up with Soviet archaeologists who did not want to follow carbon dates because it ruins their chronology. And now it is heavily defended by gatekeepers in and outside academia even though there's a current consensus that Kuzminas favorite Namazga VI phase predates Andronovo. Honestly, no difference between it and the OIT theory both require special pleading and are based on preconceived notions of magical people imposing magical languages. Yes, that R1b sacrificed little boy 100% proves Proto-Anatolian came from the steppes lol.
@Vara
ReplyDelete"I'd have to say Nezih isn't that wrong, he just has a selective approach to the data. Let's assume his model is correct. Why is that 15% Sintashta and not the 85% BMAC ancestry supposed to prove that Sintashta was Proto-Indo-Iranian?"
That's a hypothetical question because there is no Sintashta ancestry at Hasanlu. But if it were actually present, I too would ask why the 85% BMAC is discounted as the source of Indo-Iranians in the region.
The problem that I face is that most of the Kurganists online are zombie-like, they keep coming even after the limbs are chopped off. I have to ensure that the head is chopped clean, which is what I want to do by proving a complete absence of Sintashta ancestry in this case.
"His claim :"Hasanlu_LBA_A is most likely an Indo-Aryan migrant from Central Asia" is 100% percent wrong however, as the current consensus is that the Mitanni arrived in the 18th BCE at the latest. At best he is an Early West Iranian."
That's possible. We do have a sample (I3912) from Dinkha Tepe dated 1881-1693 cal BCE who has a similar migrant profile with loads of BMAC ancestry. Maybe that sample better fits Mitanni.
Target: IRN_DinkhaTepe_BIA_B:I3912
Distance: 2.1148% / 0.02114841 | R3P
52.6 IRN_DinkhaTepe_BIA_A
40.4 TKM_Gonur1_BA
7.0 ARM_Tavshut_Trialeti_MBA
"It fails on every aspect, onomastics, archaeology, DNA, and the laws of time and space . It's a bogus theory that came up with Soviet archaeologists who did not want to follow carbon dates because it ruins their chronology. And now it is heavily defended by gatekeepers in and outside academia even though there's a current consensus that Kuzminas favorite Namazga VI phase predates Andronovo. Honestly, no difference between it and the OIT theory both require special pleading and are based on preconceived notions of magical people imposing magical languages. Yes, that R1b sacrificed little boy 100% proves Proto-Anatolian came from the steppes lol."
Yeah, the gatekeeping is at an insane level. As I said - zombie-like. No amount of evidence matters to them. Just saw a tweet with the IE language tree which asserted that Anatolian came from Cernavoda in the Balkans.
It would be nice if you could drop links to the relevant works of some of the scholars you mentioned in your post.
ReplyDelete"I have to ensure that the head is chopped clean, which is what I want to do by proving a complete absence of Sintashta ancestry in this case."
The problem is that this will never work because "dilution", language can be picked up randomly and other mental gymnastics. The blogosphere is filled with lame passive aggressive amateurs so I don't bother. Makes you really appreciate the straightforwardness of our boy David. TBH, It's the people of academia that baffle me with their outlandish models.
On one side you have R1 = IE even in a clear non-IE context on the other you have people out there believing that language is picked up magically without any cultural or genetic impact because it's from the steppe. Yeah, let me change my culture and start worshipping new deities with new rituals for fun or because they traded horses with me even though I belong to the higher culture and they most likely use my language as lingua franca. Also, how come there are iron age Iranians with Hurro-Urartian, Elamite, and Kassite names but not a single one that can be reliably related to a BMAC non-IE language? Also, what super power did the Aryans of South Central Asia possess to entirely erase the original non-IE toponymy of one of the most powerful cultures of its day but could not do that in the little leftover Kassite kingdoms of Western Iran or deeper in India?
"Dinkha Tepe dated 1881-1693 cal BCE"
Yes, this is more likely since Dinkha Tepe was an outpost for SCA tin.
BTW, 18th century Mitanni Aryans have been the consensus in the Near Eastern studies since Wilhelm 1989 but the Indo-Europeanist frauds always ignore evidence that upends their little cult. As far as I know only Parpola mentions the 1760 BCE date while everyone else pretends it doesn't exist.
For the latest stuff on 18th century BCE Mitanni:
Jesper Eidem - The Kingdom of Samsi-Adad and its Legacies < this one should be read with his Shemshara Archives, which is pretty heavy. It's pretty much about how the Mitanni migrated with the eastern Hurrians of Urmia.
Eva Von Dassow - Mitanni and Its Empire < all of Von Dassow's work is about how Hurrian the Mitanni are but agrees that the Aryan elements predates the formation of the empire. Also, always ridicules Anthony so it's a plus.
Frans van Koppen - The Early Kassite Period < Claims Aryans came with the Kassite migration 1750 BCE absolutely disagree with this premise but that's another story.
Vidale has been pretty adamant of Indo-Iranian religion in BMAC
https://www.academia.edu/39589650/Protohistory_of_the_vara_Exploring_the_Proto_Indo_Iranian_Background_of_an_Early_Mytheme_of_the_Iranian_Plateau
Thanks for the links.
ReplyDelete"Frans van Koppen - The Early Kassite Period < Claims Aryans came with the Kassite migration 1750 BCE absolutely disagree with this premise but that's another story."
What about the Vedic relation to Kassite Gods Suriash, Inda-Bugash, Maruttash? Sounds like they worshipped Surya, Indra-Bhaga, Maruts. No?
I don't disagree with the Kassite being influenced by Indo-Aryans rather the claim that their migrations brought Aryans with them. The Aryan influence is 100% related to the Greater Khorasan road. The Mitanni who come with the Eastern Hurrians who are closest to the Khorasan Road have a strong Indo-Aryan influence and can be seen with their throne names. The further away you go the less you see the influence as is the case of Zabshali/Shimashki. The Kassites are in between.
ReplyDelete"Also, what super power did the Aryans of South Central Asia possess to entirely erase the original non-IE toponymy of one of the most powerful cultures of its day but could not do that in the little leftover Kassite kingdoms of Western Iran or deeper in India?"
ReplyDeleteSteppe folks supposedly wiped out languages from BMAC region and the Indian subcontinent without a trace left of the older languages (including 0 non IE toponyms in RigVeda which has only 4% non IE vocabulary) but left the Tocharians alone to survive in a smaller region which had a much stronger archaeological and genetic impact from Andronovo.
"The Aryan influence is 100% related to the Greater Khorasan road. " Sounds about right. Do you think 3000bce shahr sokhta was Indo iranian?
ReplyDeleteIt's hard to say. Whatever, the main language of SiS was it's very likely extinct. SiS is the only confirmed proper SCA city as the others are considered forts for khans and arguably the most important trading center as Baluchi, Indus, Turanian and Trans-Elamite networks have reached the site.
ReplyDeleteI follow the same timeline as the Hamp model which I arrived to through a different method. If we follow the Eric Hamp model which is finally mainstream then yes I-I presence is very likely. It was built by NMG/Geoksyur colonists however, it has been burned and destroyed many times so who knows who ended up on top later. It's abandonment correlates with the beginnings of BMAC that later expanded towards that area (violently most likely).
Vara, are you sure the Soviets proposed the Steppe model? As far as I know, at least from the linguist point of view, the defended model is the one adopted on this blog, the Armenian one, which is proposed by the Glottalic theory. That is, that the PIE urhaimat was around the South of the Caspian sea. This is the place where you find the actual the areal zone that is rich in "glotallic" sounds and poor in vowels.
ReplyDeleteViktor Sarianidi, the archeologist puts proto Iranian in BMAC, in the 3rd millenium, that is 1500 years earlier than the usual account of 1000BC. That is not very different from locating proto Indic in India around the same time.
The Sintashta route was proposed by Diakonnof and Kuzmina as every other model at the time proposed a Caucasus route where the oldest I-I words are found in the Udi language. It was heavily critiqued (see Dani's book) until it was accepted in the west and somehow became the consensus. Even though many archaeologists independent of the Indo-European debate, like Lamberg-Karlovsky, have challenged the model.
ReplyDeleteSarianidi's interpretations of BMAC rituals are accepted now by Dubova and other specialists despite what Witzel and Anthony say. However, he still dated Gonur to 1500-1000 BCE to fit with the linguistic date of Zarathustra despite the actual carbon dates. His Proto-Zoroastrian ideas are refuted.
Was BMAC really Iranian or Zoroastrian?
Hint: Why is it when BMAC (2200-1700BCE) influence is seen we find nin-Indara, Indassu, Aguni, and maryannus?
Dani as in the archaeologist A.H Dani?
DeleteCan you tell the name of the book are you talking about?
Is it relevant to read it in this decade?
https://doi.org/10.1038/s10038-022-01099-w
ReplyDeletemtDna M maternal lineages from 2 Sri lankan mesolithic samples, older one is from 7500bce. Autosomal dna is in the works.
Potentially the first AASI samples.
@Vasist
ReplyDeleteAre both of them female samples?
Don't know. Could not find the gender after a cursory search.
ReplyDeleteThey haven't extracted nuclear dna yet so we don't have autosomal or X/Y chr data.
ReplyDeleteVara, I checked this paper, and it doesn't seem the question was settled in USSR. DOI: 10.2753/AAE1061-195923025
ReplyDeleteThe free version of the Sri Lankan paper should be available here
ReplyDeleteDaniel
ReplyDelete"As far as I know, at least from the linguist point of view, the defended model is the one adopted on this blog, the Armenian one, which is proposed by the Glottalic theory."
Before the anatolian dna (and lack of steppe signals) punctured the steppe homeland theory, Armenian (or east anatolian) homeland theory was in the minority and the kurgan theory was mainstream. Afaik the proponents were Gimbutas, Kuz'mina, Diakonoff, Anthony, Vinogradov, among others.
Vasistha, I am refering to USSR only... But maybe it was a minority? I don't know, but the Moscow school is formed by many people from the Caucasus and they have a sensitive perception to these sounds.
ReplyDeleteTin from Uluburun shipwreck shows small-scale commodity exchange fueled continental tin supply across Late Bronze Age Eurasia
ReplyDeleteBactrian Tin found in the Uluburun shipwreck (1320bce) off anatolian coast. Likely reached there via the Great Khorasan road.
Question: BMAC is source of Iranian??
ReplyDelete@vAsiSTha what is your opinion of Shrikanth Talageri’s work?
ReplyDelete"BMAC is source of Iranian" - for western iran, yes. BMAC region was likely indo-iranian even before BMAC culture started around 2200bce.
ReplyDelete"what is your opinion of Shrikanth Talageri’s work?" - he gives good food for thought. His work rebutting Witzel's claim that Mittanis used pre Vedic language is quite good. But I don't think his claim of Indians spreading IE languages as far as western Europe is accurate.
@vasistha Indeed, the names of the so called IE peoples that supposedly spread IE languages can, instead, be easily related to Iranian tribes
ReplyDeleteAmenhotep III's mother, Mutemwiya, was a Mitanni princess, the daughter of king Artatama I. Amenhotep III had mtDNA H2b, which is an eastern European, Sintashta lineage.
ReplyDelete"(mtDNA) H2b is a minor branch. It contains several ancient samples from Russia, all basal to the rest of the branch, including one individual from the Yamnaya culture and one from the Late Bronze Age Srubnaya culture, both from the Pontic-Caspian Steppe region, and five other Bronze Age samples from east of the Volga river: three from Sintashta and two from Krasnoyarsk. Also in a basal position, there are three modern Russian samples (two from the Altai region) and one Danish sequence. Interestingly, while the vast majority (70%) of H2 modern sequences in our dataset are of European origin, H2b displays a strong South Asian component, with seven samples from Pakistan, India and Sri Lanka. The newly published Sintashta and Middle Bronze Age Krasnoyarsk (Russian) sequences (Narasimhan et al.2018), together with the previously released Yamnaya and Srubnaya, span a period from 5 to 3.5 ka. These, plus the modern South Asian sequences, support our earlier suggestion that H2b was involved in movements east and southwards from the Pontic-Caspian region into South Asia, by documenting its progress eastwards across the Eurasian Steppe. The Sintashta Culture in the Ural Mountains, or a “Sintashta-derived” culture (such as the Andronovo), is thought to have expanded eastwards into Central Asia 3.8 ka, reaching South Asia"
(Silva et al. 2019, Untangling Neolithic and Bronze Age mitochondrial lineages in South Asia)
For Amenhotep's DNA see: Gad et al. 2020, 'Maternal and paternal lineages in King Tutankhamun’s family'