The recent avalanche of ancient DNA data from across Eastern Europe, including modern-day Bulgaria, Estonia, Latvia, Romania, Ukraine and western Russia, has revealed prehistoric hunter-gatherer populations indigenous to the region harboring a remarkable diversity in Y-chromosome lineages belonging to haplogroups R1, R1a and R1b.
Neolithic transition in the Baltic
Baltic Corded Ware: rich in R1a-Z645
The genetic history of Northern Europe
The genomic history of Southeastern Europe
A few more ancient genomes from the Balkans and Iberia
So the once popular idea that these Y-haplogroups were instead native to Central Asia, the Near East and/or South Asia now looks very wrong.
R1a probably first arrived in South Asia during the Bronze Age with highly mobile Yamnaya-related pastoralists. These people were expanding in almost all directions from the Pontic-Caspian steppe at the time, and it's difficult to imagine that they weren't the ones who first spread Indo-European languages to peninsular Europe and the Indian subcontinent.
It's likely that almost all interested parties will soon agree that this was indeed the case. So the focus in the debate on the expansion of the Indo-Europeans, including Indo-Aryans, into South Asia will soon have to shift from whether it actually happened to how it happened. For instance, was it simply a migration or potentially violent invasion?
I already strongly believe that it was an invasion, or rather a series of invasions. I'll change my mind if, at the end of the day, the evidence says otherwise. But if you favor a migration scenario, then consider these points:
- the population in the northern part of the Indian subcontinent during the Bronze Age, even after the collapse of the Indus Civilization, was likely to have been very large for its time, and yet there was a massive pulse of admixture across South Asia from the steppe and a turnover in Y-chromosomes, especially amongst the ruling classes, suggesting that something very dramatic took place that had a major impact on the social and political fabric of the region
- early Indo-Europeans in the Near East, from the Hittites to the Scythians, are often recorded as warlike and expansionist, with a habit of invading and subjugating other peoples, like the Hattians, Hurrians and Mitanni (who apparently ended up with an Aryan elite)
- if early Indo-Europeans outside of South Asia had a penchant for invasions, then there's no reason to believe that the M.O. of the early Indo-Europeans in South Asia would have been any different, unless some sort of direct empirical evidence says so, but what kind of direct empirical evidence?
Please note, I agree that the suggestion of a potentially violent invasion of South Asia by Indo-Europeans, and, indeed, Aryans, sounds provocative, and will always be politically controversial no matter how much evidence is gathered in its favor. But what if it really happened?
Ancient herders from the Pontic-Caspian steppe crashed into India: no ifs or buts
Indian confirmation bias
"the once popular idea that these Y-haplogroups were instead native to Central Asia, the Near East and/or South Asia now looks surprisingly out of touch with reality."ReplyDelete
Only for you, no geneticist would agree with what you said. We have no Paleolithic and Mesolithic aDNA from Central Asia or South Asia. There is no genetic study out there that claims R1a is indigenous to E.Euro. NONE.
"yet there was a massive pulse of admixture across South Asia from the steppe"
Lazaridis modeling Indo-Aryans 30%-50% Yamnaya admixed is unrealistic and is not based on ancient DNA from South Asia. And, you classifying such admixture to "Indo-Europeans" is even more unrealistic since we have no idea what language these Yamnaya people spoke.
"early Indo-Europeans ... Hittites"
When we will have aDNA from Anatolia and Caucasus cultures, things will be more clear for PIE. Since steppe is no longer PIE, they most likely went through language shift themselves.
Yamnaya looks LPIE.ReplyDelete
Something like Khvalynsk and/or Sredny Stog was probably PIE. There's nothing suggesting otherwise, and it fits with the Y-DNA data.
You should probably keep in mind that there's very little clear archaeological or literary evidence of any migration leave alone an invasion. Granted that there was violence involved between Indo-aryans and Non-Indo-Aryans and between various Indo-Aryan speakers themselves, but to use the term invasion seems to suggest a pre-planned conquest besides an extremely large amount of violence. There'simply no evidence to support this. There is no evidence that the cities of the IVC fell to sword and fire. Rather it seems that increasing aridity around the beginning of the 3rd millenium seems to ve responsible for its collapse. Its incumbent upon those making claims of an invasion to provide evidence in its supportReplyDelete
But what makes you think that the massive genetic turnover in South Asia wasn't accompanied by the same sort of things that went on in the Near East at the hands of the Indo-Europeans, including Aryans?ReplyDelete
Why is South Asia special in this regard? Spicy food and warmer climate mellowed them out, or what?
R1a probably first arrived in South Asia during the Bronze Age with highly mobile Yamnaya-related pastoralists. These people were expanding in almost all directions from the Pontic-Caspian steppe at the time, and it's difficult to imagine that they weren't the ones who first spread Indo-European languages to peninsular Europe and the Indian subcontinent.ReplyDelete
If R1a spread Indo-European languages to peninsular Europe then why is R1a so rare in Western Europe?
I'm just trying to find out how and when Proto-Indo-Europeans or R1a, according to the "Out Of India" fans, migrated ?
Even in an Out of India IE migration scenario, the breaking up of PIE unity and subsequent migrations out of the homeland, does not need to go further back than 4000 BC.
According to the standard linguistic theory, Anatolian was the 1st to break off but the 2nd one in line was Tocharian and it was found in Central Asia, in a region that was historically, very much influenced by India and was very much an outpost of Indian civilization.
More to the point, Anatolian, Tocharian, Celtic, Italic, Germanic and Balto-Slavic are said to have moved out from the original homeland and come together in a secondary homeland, where they shared linguistic innovations to the exclusion of those who were still in the original homeland i.e. the Indo-Iranians, Greeks, Armenians, Albanians etc.
In an Out of India scenario, this secondary homeland is Central Asia, north of the Himalayas. Anatolians would be the 1st ones to move out of this secondary homeland towards the Near East while Tocharians move towards the East or Northeast eventually ending up in the Tarim. This would be around the time bracket of about 4000 BC - 3500 BC or thereabouts.
Central Asia in this period, does show signs of influence from South Asia but mostly, the archaeology of this period needs more excavation, examination and analysis. The research is not so extensive as it is for Europe or the Near East.
Incidentally, Mariya Ivanova, the archaeologist, argues for an influence from Central Asia into the formation of the Maykop. If true, this would give you the archaeological evidence for a movement of IE people from Central Asia into Europe.
It is also pertinent to note that Anthony and Ringe, in a paper published a couple of years ago, had argued for the wheeled vehicle vocabulary being present across all IE groups except the Anatolians. Yamnaya do show evidence of wheeled vehicles and so do the Maykop. However, as Ivanova notes, wheeled vehicles were also present in the Indus civilization from atleast 3500 BC if not earlier. This innovation certainly looks like to have spread from the Indus civilization into Central Asia and Eastern Iran. This is also accompanied by the spread of Indian Zebu cattle into those regions in that timeframe.
The Wheeled vehicles could therefore have spread from Central Asia into the Maykop and from there on the steppe among Yamnaya. It is also noteworthy that the most complete wheeled vehicle vocabulary is found among the Indo-Aryans. Such an excellent preservation of all the PIE terms for wheeled vehicles also argues against the Indo-Aryan migration into South Asia and instead argues in favour of long term presence in South Asia.
The R1b-Z2103 found among the Yamnaya is present in South Asia but in very minor proportion. However, it has a significantly higher percentage in Central Asia.
For this theory to hold true genetically, the Y-dna of the Tocharians and Anatolians needs to be R1b rather than R1a.
The last to leave the secondary homeland of Central Asia would be the Balto-Slavs who would also by then have been more in contact with those in the South such as the Indo-Iranians and the Greeks and would therefore have linguistic innovations shared with them to the exclusion of Anatolians, Tocharians, Italo-Celtic and the Germanic. These latter groups were perhaps spread by R1a carriers.
The timeframe for the migration and dispersal of these latter groups would co-incide with the phase when the Indus civilization was entering into its advanced phase in the 3rd millenium BC. Therefore we would expect the latter migrants to have a more refined and materially more advanced culture than the earlier migrants. Historically the Greeks, Iranians and the Indo-Aryans have been more sophisticated among the Indo-Europeans.
I hope this partly answers your questions.
An alternative hypothesis is that there was no sudden 'invasion'. Rather, steppe people were incorporated into 4th millenium interaction sphere including steppe, central Asia and south Asia. Their earliest presence would thus already be present in 2500 BC BMAC, and arrived in India toward the later phase of the Mature Harappan period as groups of people shited southeast.ReplyDelete
So they were just incorporated, and then systematically rose to the top of society at the opposite ends of the Indo-European range, from the British Isles to India, because why?
Go and see a good psychologist.
"So they were just incorporated, and then systematically rose to the top of society at the opposite ends of the Indo-European range, from the British Isles to India, because why?"
Noone saying that they were non-violent or that the establishment of their supremacy was entirely peaceful. Rather arguing about the magnitude of violence and pointing out that a lot of other factors were important in the establishment of their supremacy as well
The same thing happened in Western Europe, South Asia and the Near East, and we have records of it from the Near East.ReplyDelete
Well according the direct arxhaeological evidence, the elite in Swat valley were women.ReplyDelete
Modelling the entire IE as the arrival of BB in Britain would be a case of the tail wagging the dog.
Curious to know which group in particular you keep referring to in the near east
It is a shame that this topic is so sensitive, particularly for Indians. Hopefully the truth will out.ReplyDelete
Curious to know which group in particular you keep referring to in the near east.
Well, from my reading, the founding of the Hittite Empire on Hattic lands wasn't exactly a friendly acquisition sort of a deal.
They also sacked Babylon, from memory.
An important issue in India is genetic stratification by social status/caste. When we were analyzing Y chromosome data from India about 10 years ago, I inquired about why there were no Dalit samples in our study. The response was that the Dalit are "invisible" are not of interest. Since Dalit comprise probably at least 20% of the population of India, I'm wondering whether anyone more recently has published autosomal/R1a1-Y data on Dalits (non-ST/non-SC). Such data would be helpful in teasing apart the degree of possible movements from the steppes into South Asia.ReplyDelete
chamar, dharkar, mala, pallar, pulaya are all dalit castes. Dalits = SCDelete
In the true words of the late, great, Long John Baldry,ReplyDelete
'.....let the flame wars begin,'
'.....I can't help it, I can't win'
Dalit apparently are not SC nor ST. It was the Sengupta et al 2006 paper that we wrote that did not include any Dalit groups--we published both SC and ST results. The collaborators from India did not sample Dalit for the reasons mentioned above.
Not related to the argument at hand but just to clear things up for you, there are no non-SC Dalits, SC (Scheduled Caste) is just the official term for Dalit. Of course, there are a whole host of other 'subaltern' populations that are not necessarily Dalit, they are grouped under OBC (Other Backward Classes). ST (Scheduled Tribes) are tribal populations some of which have been 'Hinduised' only in the last 100 years, while some others haven't. A lot of STs are also Christian as the missionary presence is strong among them.
Thanks for the clarification! Are there studies--autosomal and/or Y chromosomal of OBC--Other Backward Classes/Subaltern populations?
I was always under the impression (along with my Dalit friends) that the Dalit were beyond the four classical varnas and were thus "untouchables". That was how my colleagues termed the social categories in the paper.ReplyDelete
If sengupta implied dalit =/ SC he's mistaken. Dalit essential refers to ethnicities classified formely as 'unclean' oweing to their association with tabooed profession such as undertaking, butchery, leather-working etc.(Though many were in reality landless agricultural labourers) SC is the legal term for these groups.
I'm guessing ur referring to the large grouping referred to as the OBCs. However in comparison to the later grouping the OBCs are heterogenous including amongst their ranks various upper caste landed castes, merchants, middle ranking peasants and artisans to even a few 'upper untouchables'
Yes, the Dalits, previously known as Untouchables, were beyond the four classical varnas. Gandhi had also coined the term 'Harijans' (children of god) for them. After independence, the government of India created a separate category for them, known as Scheduled Castes (SC), mostly for affirmative action purposes. I understand you not getting this right, but your Dalit friends should have known better.
A quick look suggests that OBCs mostly appear to be occupational castes and the list differs from state to state. I think some groups like Yadavs, Kurmis, Gujjars and Jats (both of whom aspire to OBC status for affirmative action benefits) have been covered in studies.ReplyDelete
So the R1a people found at Derievka 10 000 years ago had nothing to do with todays European population ? Not ancestral in any way, not even distantly related ?
Thanks again for the sociological clarification. The individual who considered herself Dalit was working on a PhD in religious studies. She was from Kerala, and from a long line of Christians. She must have been part of OBC, since she was neither SC nor ST, like many Christians in India. Her family was definitely "untouchable".
If she was from Kerala and from a long line of Christians, it's possible she may be a Syrian Christian. This is not the time and place to go into the details of Christianity in India, but Syrian Christians are a very old line dating back to St Thomas, or at least so they claim. A lot of people from Kerala, especially their educated classes, come from that lineage, and the most common surnames are Joseph, Mathew, Vincent and Thomas.
But we are digressing too much. Let us not take this any further here.ReplyDelete
She's probably right. Kerala is sociologically unique. Except for the Brahmin and Nair aristocrats most other major groups including the peasantry (ezhavas for eg. who have been included in some older studies) were considered 'untouchable' though they socio-economically they are closer to the middle castes of surrounding regions. They have now been included amongst the OBCs. Also xtians even those from dalit ethnic groups aren't legally SCs, they cease to hold this status on conversion.
"An alternative hypothesis is that there was no sudden 'invasion'. Rather, steppe people were incorporated into 4th millenium interaction sphere including steppe, central Asia and south Asia. Their earliest presence would thus already be present in 2500 BC BMAC, and arrived in India toward the later phase of the Mature Harappan period as groups of people shited southeast".
In my view, there were two or three waves of Indo-Aryan to South Asia, the lattest could be the one of Rigvedic people. The first in Harappan times (more or less as Rob comments). This continuous flow maybe contributed to massive y-DNA record.
I think we have to be clear about exactly what we mean when we say "invasion(s)" vs "migrations". Those are vague terms that could encompass a plethora of historical scenarios.ReplyDelete
Looking at documented historical cases, we can see a huge variety when it comes to population introgressions falling along what you could term a spectrum between peaceful assimilation and organized genocidal invasion.
Germanic peoples migrated into Roman territory and both assimilated and invaded to the extent that the late Roman Empire saw Roman armies with Germanic soldiers led by Germanic generals fighting other Germanic invaders.
Turkic Mamluks started as foreign slave soldiers in Muslim armies and eventually ended up as the ruling caste of much of the Muslim world. At the same time they were fighting other Turkic armies.
I guess what I'm saying is that we should expect complexity, including the possibility that Indo Aryans migrated both peacefully and violently and probably fought each other at least as much as they fought the native post-IVC population.
"You should probably keep in mind that there's very little clear archaeological or literary evidence of any migration leave alone an invasion. "
I've heard many people say this about different many subjects. Then later ancient DNA confirmed there really was a migration. Maybe archaeology isn't good at detecting migrations that occurred within a short period of time and even involved an invasion.
Not many expected Balkan WHGs to be rich in R1b1a. Considering R2 existed in Neolithic Iran, it isn't too crazy to think R1 did as well.
It's all the LNBA data like R1a M417* in Corded Ware which debunks any theory that R1a Z93 didn't come to South Asia from Eastern Europe.
Kalash have the highest steppe dna but have mainly local South Asian Y dna, a similar trend is seen with Central Asian Pamiri Tajiks, where L and J are also common among them. Groups like Jats which have also have the highest steppe ancestry among Northern Indians are not even in the Varna system. With Kalash women , U4 which is typically considered a Steppe Mtdna is the most common, which means to your horror , they are local South Asian men marrying Aryan women omg. So clearly it far more complicated than that juvenile brain of yours is painting it. Another thing to note is of the thousands of excavations sites in the Swat and in Northern Pakistan, they did not find 1 chariot, or any fragments . Which is strange considering that was the main advantage groups like Hittites and Mitanni , Kassites had over the urbanized city states they ruled. Which is another point to note , that the IVC city states were long abandoned a few hundred years earlier.ReplyDelete
"- the population in the northern part of the Indian subcontinent during the Bronze Age, even after the collapse of the Indus Civilization, was likely to have been very large for its time, and yet there was a massive pulse of admixture across South Asia from the steppe and a turnover in Y-chromosomes, especially amongst the ruling classes, suggesting that something very dramatic took place that had a major impact on the social and political fabric of the region"ReplyDelete
Isn't that a point in favour of migration? You'd need a large movement of people to have a large demographic effect.
They used various Dalit samples, it's not called as such.
Lazridies in his study modeled Mala (dalit) from South, as 18% steppe. Reich models ANI admixture in Mala 38.8% & another south dalit group Madiga 40.6%. Northern dalits like Chamars be more ANI shifted. They are not homogeneous group.
This modeling will most likely change, things will be more clear with ancient DNA from South Asia.
I think that posing the question as an either/or is not a very productive one. I'd favor framing it as "how did this demographic transition happen?"ReplyDelete
In the vein, probably the most notable point is the fairly strong evidence that Harappan society had collapsed or was in an advanced state of collapse when the Indo-Aryans arrived, probably due to climate reasons that among other things diverted one of their major river systems.
Indo-Aryans may have conquered, but they probably conquered a fragmented and demoralized set of ad hoc warlords rather than a unified Harappan empire. It was probably more like conquering Somolia than Egypt.
I wanted to get these on the last comment thread, so let me take them on right now.
'- the population in the northern part of the Indian subcontinent during the Bronze Age, even after the collapse of the Indus Civilization, was likely to have been very large for its time, and yet there was a massive pulse of admixture across South Asia from the steppe and a turnover in Y-chromosomes, especially amongst the ruling classes, suggesting that something very dramatic took place that had a major impact on the social and political fabric of the region'
Stilicho & Aetius. Collapsing civilisations show an increase in violence, but that does not always mean an invasion. One group of barbarians, like Stilicho, Aetius & Ricimer can come to power because the local men all die in war, and they are needed to fight other barbarians.
"- early Indo-Europeans in the Near East, from the Hittites to the Scythians, are often recorded as warlike and expansionist, with a habit of invading and subjugating other peoples, like the Hattians, Hurrians and Mitanni (who apparently ended up with an Aryan elite)"
I do not mean this to be personal, but this is some terrible historic analysis. Having studied History at a University that is excellent in these matters, and worked with those whose theses were published on Near Eastern History during this period, this doesn't really fit with the historical method as it applies to ancient societies.
1) The Indo-Europeans were no more violent or expansionist than any other society at this time. The difference is that they were remarkably successful in imposing their language and certain cultural attributes, though not always.
2) The early Hittites are first attested as vassals of the Assyrians. This is around 2000 BC, while their conquest period only peaks around 1300 BC. This isn't some Indo-European trait, since Near Eastern polities were constantly at one another's throats. We do not know how they came to power.
3) The theory of a violent takeover of the Hattians has been largely disproven. It was based on bad archaeology that postulated an apparent 'migration of peoples fleeing the Hittites.'
4) We do not know how the Hurrians wound up with a Mitanni Indo-Aryan elite. In fact, we only have a few words used by Kikkuli, references to gods in a treaty, and ruler's names. As a whole, their society was basically Hurrian, and outside the treaty of Qadesh, they seem to be worshippers of the local deities. Surprisingly, it seems that Hurrian society was largely the same, and that their Indo-Aryan language and culture was not very successfully imposed.
"if early Indo-Europeans outside of South Asia had a penchant for invasions, then there's no reason to believe that the M.O. of the early Indo-Europeans in South Asia would have been any different, unless some sort of direct empirical evidence says so, but what kind of direct empirical evidence?"
Good reason - we don't know how the Indo-Iranians arrived. They show up at the Medes, on the Iranian plateau, but we don't see a conquest of the Elamites. The Iranian arrivals are actually vassals of the Assyrians, which betrays the idea of a powerful conquering group.
Also, the steppe component in India does not peak with the warrior castes, but instead with the priests. Brahmins have more of the steppe component than Rajputs.
Further, the steppe component correlates with the Iranian Neolithic components, which indicates some continuity with the local ruling elites. If this was the invasion you propose, the numbers would line up differently.
I see in the Jones et al 2015, both Mala and Vishwabrahmin, who are SC/Dalit can be modeled as an admixture of Onge and Kotias (CHG) and not with Afanasievo_BA admixture, unlike some of the Gujarati populations. So perhaps the Mala and Vishwabrahmin were influenced broadly from CHG and ASI. That said, Mala are almost 50% R1a without steppe input?
Honestly, considering the fact that high-caste Northern Indians have more Steppe_EMBA than all Southern Europeans and some Western Europeans, and considering the fact that the Kalash have around the same amount as Lithuanians, I'm pretty sure that the Indo-Aryan expansion was fairly "explosive".ReplyDelete
Although, with regard to whether it was an "invasion", or just a massive folk "migration", that I don't think we can really know, and both "invasion" and "migration" are very loose/vague terms in this context.
That being said, considering that there is very little steppe-related mtDNA in India, I guess a primarily male expansion might make more sense.
Then again, the Kalash do actually display a fairly substantial amount of steppe-related mtDNA, as do many Pashtun tribes.
So, the lack of steppe-related mtDNA in South Asia proper could simply reflect "selection" (not an expert on this angle, but if I'm not mistaken, mtDNA may be subject to natural selective dynamics).
Or, the presence of decent-to-substantial steppe-related mtDNA in South Central Asia (Tajikistan, Afghanistan, western + northern Pakistan) could indicate that the Aryans arrived in this region via a huge folk migration, and then expanded into South Asia proper via a male-mediated movement that could be better described as an invasion.
I think that this scenario (a two-pronged movement; first a folk migration into the southern portion of Central Asia proper, and than a male-mediated expansion into northern India) is the most parsimonious.
Yep, I don't think PiE expansion was an ineffable, unstoppable phenomenon.
It was a series of events
Despite claims to the contrary; it's not even understood yet for Europe, let alone Asia
@Sein and everyone:ReplyDelete
In also believe their expansion into India was rather explosive and I arrived at this conclusion by looking at results from Lazaridis as well as some of Kurds's calculators.
No, I don't believe that most of the steppe DNA was brought to India by a mediating group(say, an even mix of Iran_Neo and Steppe from Bactria). If we take the steppe fraction from those runs and try to couple it with Iran_Neo, what ends up being left over is an ASI-heavy "pre-steppe" component. Smth like 70% ASI and above. That is a hard scenario to wrap my mind around. Judging by how much impact the farmers in Europe had, uniparental markers in NW India and archaeological evidence, Iran-related DNA was certainly substantial and widespread.
Besides, we have historical post-cedents of this sort of thing occuring----in the middle ages, to be exact. Why can't people deal with it? lol That's right, ~1100-1300 AD and into sophisticated urban landscape. I am referring to the migration of Seljuks and Mongol-triggered Oghuz migrations into Anatolia and Iran.
Historical records indicate a forked, simultaneous scenario. Moreover, Dienekes demonstrated with exhaustive analysis(involving Roloff, etc.) that this was a largely undiluted 50-50 ENA-WestEurasian wave. In other words, more like a pulse of modern Kazakhs than like modern Turkmenistanis(who represent a mixture of Turkified, assimilated locals and "original" Turkics).
This total central asian admixture was estimated at ~20%. This is consistent with a higher Yamna-related ancestry in Turks compared to Armenians, who would be interchangeable otherwise. Not insignificant considering how densely populated and urbanized Anatolia was!
yeah but the Turks came in tens of thousands and were galvanised military unitsReplyDelete
Anthro SurveyNo, I don't believe that most of the steppe DNA was brought to India by a mediating group(say, an even mix of Iran_Neo and Steppe from Bactria). If we take the steppe fraction from those runs and try to couple it with Iran_Neo, what ends up being left over is an ASI-heavy "pre-steppe" component. Smth like 70% ASI and above.ReplyDelete
Problem with that though, if you had an independent entry of Steppe and Iran_Neo into ASI India, you'd expect to see some independence of the components actually existing today.
The way it is in Europe, where WHG and Barcin and Steppe_EMBA have some independence, and you have Bulgarians and Romanians having similar Steppe_EMBA to France (or high caste North Indian Brahmins), but less WHG than Spanish and the triangular shape of the PCA fits this independent relationship.
In India you just have a single cline, with the only exceptions really being Baloch, Makrani, etc, who live outside India anyway. You don't seem to have Dravidians who have a high ratio of Iran_N:Steppe relative to their ASI or anything like that. South Asian populations - Kalash, Indo-Aryan, Dravidian, tribal - basically model with the same ratio of Iran_N:Steppe. (Differing ratios of Iran_N:Steppe among Indian populations are a scenario I've tried to explore as it seemed to "make sense" but there doesn't seem to be any populations that actually *work* or require it).
It's hard to see scenarios where a population with a variable IranN:ASI ratio all happened to take on exactly the right amount of Steppe to have the same Steppe:Iran_N ratio.
For what it's worth, the Indo-Aryans didn't encounter a thriving civilization; IVC had already collapsed pretty hard, prior to the IA influx.
So, even if it was an invasion, Indra doesn't stand accused.
At the end of the day, I guess there was a vacuum to be filled.
I wonder if some kind of X-autosome comparison likeReplyDelete
Mbuti Test Dai/Onge Yamnaya/EHG
for Indian populations would reveal whether it was a male-dominated invasion when the result is compared to that of some population with ASI but low steppe ancestry. These comparisons work for Caribbean admixed populations (autosome more European, X more Amerindian) and also revealed that 1000genomes Europeans don't have sex bias admixture relative to each other but Sardinians do have it (likely male-mediated steppe) relative to them. Just need a proper reference.
I have no problems with whether it was an "invasion" or "migration" (i'm no 'apologist'), as the terms imply a false dichotomy anyhow.
A post-1700 BC arrival is easy to envisage given the "pull" of empty land, previously cleared, thus good for pasturizing.
But I am merely allowing for other scenarios - such as my pre-admixed BMAC one. The reason for this are twofold,
1) First of all genetics - Matt's & others comments, and my own skylarking suggests that the steppe in "south Asia' precedes Kura-Arax-type admixture, thus might be from before 2500 BC. Soft & intuitive, sure, but not unreasonable.
2) The obvious lack of genetic trail which demands explanation; and the explanations offered by some commentators to get around this issue don't suffice because they're either factually inaccurate to begin with, or are false analogies.
(E.g. the apparent lack of steppe trail into BB territory sometimes conjured up is simply a false statement if one merely eyeballs the archaeological material)
I do agree that choosing between "invasion" and "migration" is somewhat forced; neither are mutually exclusive.
Although, my personal bet is that BMAC was probably equivalent to being a mix of Iran_Chal and Iran_N, while IVC was probably equivalent to being a mix of Iran_N and ASI.
I think qpGraph is very powerful, and by far the best method for analyzing these questions (much better than TreeMix, qpAdm, and ADMIXTURE).
That being said, the output is still a highly stylized abstraction of (a possible) reality, and that too based on current aDNA.
With actual aDNA from IVC, I think we may possibly find that the ratio of Iran_N to Steppe_EMBA does vary in South Asia.
Anyway, totally off-topic, but I was really interested in hearing your thoughts on Razib's tweets concerning Tianyuan:
"models: admixture btwn proto-euro and proto-asian early on. or ancient structure before separation into eurasian west and east"
I guess ancient West Eurasian-ENA gene flow is now something that has to be recognized in any modelling of these relationships. Or, at the very least, some sort of complex structure.
In conceptual terms though, don't these deep linkages between ENA and ancient European/Siberian hunter-gatherers make the very distinction between ancient West Eurasia and ENA rather problematic?
I mean, we already knew about BEu in modern West Eurasia, so this isn't a shocking thought.
If we want to be analytically on the mark, I think distinguishing between ENA and the non-BEu portion of West Eurasian ancestry is no longer a very effective line of thinking. At the end of the day, ANE, WHG, and the huge "ENA" group (East Asians, Siberians, Native Americans, ASI, Onge, and Oceanians) don't seem to be separate clades. They're mixtures of each other; a massively complex rhizome involving lots of intersecting ancestry and rich structure (ANE admixture in WHG, UP European-related admixture in MA1, ENA admixture in ANE, ENA admixture in WHG, ANE admixture in East Asians/Siberians/Native Americans, ASI admixture in Southeast Asia, Oceanian-related admixture in Southeast Asia, East Asian admixture in Melanesia, etc).
In the end, the fact that they look like clades could just turn out to be a function of sampling (we have very little in the way of Upper Paleolithic/Mesolithic aDNA across Eurasia), and with more extensive samples we may have to revise our terminological apparatus.
"35,000 year old goyet from belgium a lot like tianyuan compared to other ancient europeans."
I forget; was there any hint of this in Qu et al.?
Yeah Goyet 116-1 , looked 'distant' in the MDS.
How close is Goyet to Tianyuan? Apparently as close as East Eurasians are and closer than Mesolithic Europeans:ReplyDelete
MYang: Tianyuan - 40ky individual more closely related to modern E Asians than to all modern & ancient Europeans except Goyet #smbe17
It's also closer to modern East Asians than Kostenki14, but not closer to East Asians than Loschbour and company. This adds some complexity.
So Goyet has some 'retained' East Asian affinity by way of affinity to Tianyuan, which is partly ancestral to modern East Asians.ReplyDelete
Yet Loschbour has closer East Asian affinities probably due to arrival of new 'eastern' admixture in the post-glacial stage, perhaps via an intermediary which admixed into Mesolithic Europeans and pre-modern East Asians. This could be Siberian/ ANE admixture, right ?
That's interesting. I think I should re-read that paper.
It get's even more cool:
"Shared ancestry btw Amazonians, Papuans, and Tianyuan - persistent ancestral population structure"
Mala are mostly a mixture of ASI and Iran_Neolithic, but they do have around 9% of Steppe_EMBA, of which around 50% is CHG. So that makes less than 5% of CHG.
The problem with the stats in the Jones paper is the lack of Iran_Neolithic. This results in an inflated signal of CHG in the Mala. And it's impossible to model details such as minor steppe ancestry with such stats, because they only allow for two potential sources of mixture.
Guys, I'm working on the Goyet angle, and I think I'm getting close. I will share what I can, when I can.ReplyDelete
This is what I said: So the once popular idea that these Y-haplogroups were instead native to Central Asia, the Near East and/or South Asia now looks surprisingly out of touch with reality.
So I'm not saying that R1a and R1b can't also be native to Central Asia, the Near East and/or South Asia, but that it was assumed until recently that they were native to these regions instead of Eastern Europe where they were seen as intrusive.
But admittedly, the chances now of finding R1a in foragers from anywhere south of Eastern Europe and Siberia are very, very slim. South Asia looks about as likely as the moon in this context.
And I also doubt that we'll see any R1b in foragers outside of Europe and Siberia, but R1b is an old haplogroup, so some basal, albeit irrelevant, lineages might turn up in surprising places with enough sampling.
In the Ganges part of India past Punjab, IE's seemed to have admixed with a population that was 50% ASI and 50% Iranian Farmer. Davidski can you make a tree using the Chamar as a target to confirm this?ReplyDelete
The couple of Chamar samples that I happen to have in this dataset are showing a ratio of ~70/30 ASI/Iran_N before steppe admixture of ~20%.ReplyDelete
Other Chamar individuals might show something different, although I'm not sure if 50/50 is realistic.
That's very interesting.
I guess this shows that modelling different South Asian populations together (in the same topology) is somewhat problematic, as it forces all of them to have ancestry from a hypothetical population that was 50/50 Steppe_EMBA/Iran_N.
But 20% Steppe_EMBA for the Chamar makes good sense.
As an experimental exercise, can you substitute the Kalash for Chamar (so, the same exact topology)?
I would really appreciate this. As always, only once you find the time, and only once you have the inclination.
Thanks in advance.
The R1a-Z93(Z94+) in the Poltavka outlier is really a smoking-gun. He lived around 2925-2536 calBCE, and this haplogroup is hardly any older. Yet, what made him autosomally an outlier to other Poltavka samples isn't South Asian admixture - he didn't have any - but EEF admixture from Europe. All the Indian opponents to the AIT can do with this important piece of evidence is to ignore it.ReplyDelete
And even if just the Late PIE was from the steppe and the original PIE homeland was in West Asia, then Indo-Iranian would still be derived from the steppe, because it's derived from Late PIE, it's not as distinct as Anatolian or Tocharian. So the possibility that the ultimate PIE homeland was in West Asia cannot really spare you an AIT scenario.
Granted, Yamnaya and related EMBA Steppe cultures were nearly half CHG, and you might speculate they got their IE language from the CHG part of their ancestry. But then I wonder: Where is all the (yDNA) J1, J2 and G2a in their elite burials?? Ancient DNA has shown that West Asia, including Iran, was full of these. Yet there was just one single J2 in an EHG forager in remote northeastern Europe that had very little to do with Yamnaya. The rest is R1b-M269 and (in the Corded Ware and later cultures) R1a-M417.
@ Sein and Davidski: I guess this shows that modelling different South Asian populations together (in the same topology) is somewhat problematic, as it forces all of them to have ancestry from a hypothetical population that was 50/50 Steppe_EMBA/Iran_N.ReplyDelete
The problem OTOH is that it seems like if you're modelling them separately, then the drift edges between the Iran_N like, Steppe_EMBA like and Onge like ancestor can slide around.
Let's try another co-fit model: https://pastebin.com/TEeFWgfU
This has each population with a unique individual ANI population to which ASI is added. The problem may be that modeling with a single ASI and a single ANI compromises all the Iran_N to be in the ANI population, somehow.
@ Sein, I think Tianyuan actually particularly makes it pretty tricky to define Basal Eurasian as a separate thing also. We can't use East Asian groups to capture that, can't use Tianyuan (although ongoing ancestry from Goyet is very slight in West Eurasia today), and I think I have some doubts about using Ust Ishim as well, since it's harder to be confident there was no geneflow between Ust Ishim like populations and others.ReplyDelete
Which LGM Refugia are currently known ? Balkans, Crimea, Southern Urals, Southern Siberia/Dzungaria ?ReplyDelete
Btw, what's this about Basal Eurasian? No longer a separate thing? So what is it now?
Hey, I have an update on your mtDNA. Since you emailed your mtDNA last year my European database has grown to 20,000.
Your haplogroup; K2b1a1, only pops up in Spain, France, Italy, Denmark, and Poland. Though a mere 1.2% of my 3,000 Spanish samples belong to it, it peaks in Spain by a long shot.
Very nice; seems quite sensible.
Basically, the non-Steppe_EMBA ancestry of the Kalash should be much more Iran_N-related than the non-Steppe_EMBA ancestry of the Chamar, and your topology shows exactly that.
Going purely by geography, one would expect the non-Steppe_EMBA related side of the Kalash to be something like 80%-75% Iran_N + 20%-25% ASI, and with the Kalash themselves at around 10% ASI, compared to the non-Steppe_EMBA side of the Chamar being 70% ASI + 30% Iran_N. The difference in Steppe_EMBA ancestry is also rather sensible (45% vs 21%).
For what it's worth, with something simple like nMonte (in conjunction with PCA data), higher ASI tends to correlate rather strongly with a higher Iran_N-to-Steppe_EMBA ratio (in India).
Perhaps, this could be due to Neolithic Indians being akin to a mix of Iran_N and ASI?
So, I think that your topology probably reflects the actual patterns of gene-flow and ancestral structure.
I think that your topology is of great interest, but it still leads to ANI being around 50/50 Steppe_EMBA/Iran_N for everyone, which doesn't strike me as plausible.
I think the issue might be that "ANI" wasn't a real population; Iran_N + ASI preceded Steppe_EMBA admixture in this part of the world.
"I've heard many people say this about different many subjects. Then later ancient DNA confirmed there really was a migration"
this is a very vague statement. Please provide an example.
Let me give you some...The invasive nature of CW in Europe has been written about by archeologists for decades for example. DNA seemed to agree with what archeologists were already saying, so in this case dna did not lead but lagged.
Akkadians replaced Sumerians. archeological evidence is clear. Here DNA evidence is not even needed, for all we know the two populations may have been indistinguishable genetically.
Similar proof is needed to prove massive invasion in South Asia. What we have is a relatively arid, deurbanized IVC phase. So its a transition at least. Archeology suggests a population shift from arid western regions eastwards. Still tell tale chariot and horse remains remain non existent. and CW or Yamnaya trail to IVC trall is missing. Andronovo borders BMAC but no andronovo traits are found beyond that.
The only way that there was no massive migration to South Asia during the Bronze Age from the steppe is if a population like Yamnaya was native to some part of South Asia.ReplyDelete
And anyone who argues this obviously can't be taken seriously.
"@The R1a-Z93(Z94+) in the Poltavka outlier is really a smoking-gun. He lived around 2925-2536 calBCE, and this haplogroup is hardly any older. Yet, what made him autosomally an outlier to other Poltavka samples isn't South Asian admixture - he didn't have any - but EEF admixture from Europe."
by your logic this is a smoking gun that the original R1a were EEFs. Since obviously uniparentals are the same as the autosome and the correlation is infallible. Especially in this case its so close to the birth of the lineage so it must be so.
Which South Urals refugia was there ?
"The only way that there was no massive migration to South Asia during the Bronze Age from the steppe is if a population like Yamnaya was native to some part of South Asia"ReplyDelete
why so, whats the proof that the only possibility is that ancient South Asia has to yamnaya like?
Clearly yamnaya uniparentals don't match with modern south asians. Nor any kurgan trail.
So you cannot use that.
Do you have statistical proof and arrow of time arguments for all recombination events since 5k BP in all concerned and intervening populations.
Clearly yamnaya uniparentals don't match with modern south asians.ReplyDelete
Some early Corded Ware samples are identical to Yamnaya and some belong to R1a-Z645, which is directly and recently ancestral to "South Asian" R1a-Z93.
Nice try. Haha.
I may have misunderstood what you're saying, but this model isn't locked into showing the West Eurasian (ANI) portion of ancestry in South Asians as 50/50 Iran_N/Yamnaya. It can vary this ratio in a significant way if necessary, like for the Paniya.
Generally speaking, that ratio is close to 50/50 for the more ANI South Asian groups, and closer to 60/40 or even 70/30 for the more ASI groups.
"The couple of Chamar samples that I happen to have in this dataset are showing a ratio of ~70/30 ASI/Iran_N before steppe admixture of ~20%.ReplyDelete
Other Chamar individuals might show something different, although I'm not sure if 50/50 is realistic."
Thank you for the feedback. There is a consistency of 70/30 ASI/Iran_N (pre-steppe) in both the Mala and Chamar.
On the other hand, Gujarat_D and others seems to be 50/50 before steppe.
Considering eastern and southern migrations after IVC collapse, there may have been absorption of additional ASI from the rest of India.
Examples in Europe would be additional WHG in Iberian farmers compared to the ones in Anatolia and South East Europe.
I would estimate the Harappans (at least in the Punjab) might be 40/60 ASI/Iran_N or 50/50 ASI/Iran_N.
Thanks for the Mala results! Can you do the same analysis for the CEPH Burusho? My guess is that they will have a large Iran_Neo proportion.
Because this, I'm Kazib Khan's fan:ReplyDelete
Clear, logical and precise.
Exactly; it does vary the Steppe_EMBA-to-Iran_N proportion, in your topology.
I was referring to Matt's co-fit topology.
In the co-fit topology, all the populations model as having West Eurasian admixture from a hypothetical group which was around 50/50 Iran_N/Steppe_EMBA, despite each one of them having separate ANI nodes.
Basically, it seems that co-fitting these populations always leads to the 50/50 ratio (it's happened before), so I guess it's probably best to run each modern Central/South Asian population individually.
Speaking of which, can you possibly swap Tajik_Shugnan for Paniya? So the same exact topology, just a different test population.
I'm very interested in seeing how they turn out.
As always, only once you have the time, and only once you have the inclination.
I'd really appreciate it. Thanks in advance.
Hardpans were 85% ASI and 15% Iran_NReplyDelete
I mean *harrapans*ReplyDelete
The invasive nature of CW in Europe has been written about by archeologists for decades for example. DNA seemed to agree with what archeologists were already saying, so in this case dna did not lead but lagged.
Akkadians replaced Sumerians. archeological evidence is clear. Here DNA evidence is not even needed, for all we know the two populations may have been indistinguishable genetically.
Similar proof is needed to prove massive invasion in South Asia”.
Tony Joseph answers Michel Danino:
“Archaeology has been indeterminate on the Bronze Age migration question, but there is no archaeological discovery that will militate against genetics joining linguistics in favouring migration as the right explanation.”
I'll just leave this here. Very interesting with Basal into West Eurasians after GoyetQ116-1. No Z>2. Adding Natufians to this tree brought up Basal by a couple percent. I am still working on this tree here. I'm not saying it is absolutely solid, but it works.
I am did not see the constant ratio across populations. Perhaps I misunderstood you. All three can vary quite a bit relative to each other----even within the confines of the IndoGangetic plain.
You have folks like Bengalis who will have extra steppe left over if you tether all IranNeo to steppe. And you also have folks like Sindhis who will have extra IranNeo left over.
In south Indian groups, you essentially have Iran+ASI+next to no steppe.
Even if there was/is a strict ratio, it does not exclude a long-range migration involving mainly steppic folks not substabtially admixed with sedentary IranNeo DNA in Bactria. It could simply mean than after their areival, there was a phase of homogenization w/local Iran-heavy north Indians and subsequent re-expansion across India. Basically, you'd have a mediating population. Sort of like with Beakers in Europe.
"I think distinguishing between ENA and the non-BEu portion of West Eurasian ancestry is no longer a very effective line of thinking. At the end of the day, ANE, WHG, and the huge "ENA" group (East Asians, Siberians, Native Americans, ASI, Onge, and Oceanians) don't seem to be separate clades. They're mixtures of each other; a massively complex rhizome involving lots of intersecting ancestry and rich structure (ANE admixture in WHG, UP European-related admixture in MA1, ENA admixture in ANE, ENA admixture in WHG, ANE admixture in East Asians/Siberians/Native Americans, ASI admixture in Southeast Asia, Oceanian-related admixture in Southeast Asia, East Asian admixture in Melanesia, etc)."
Agree, there is deep admixtures between Crown Eurasian ENA/ANE/WHG that we don't understand yet. Tinyanian might help uncover some of this for now but we have known for while that Crown Eurasians ENA/ANE/WHG are more related to each other than to Basal Eurasian/BEu.
@Sein: I think that your topology is of great interest, but it still leads to ANI being around 50/50 Steppe_EMBA/Iran_N for everyone, which doesn't strike me as plausible.ReplyDelete
I think the issue might be that "ANI" wasn't a real population; Iran_N + ASI preceded Steppe_EMBA admixture in this part of the world.
Still don't agree; PCA, ADMIXTURE show single clines and two component solutions in India - just doesn't seem compatible with those lines of evidence.
If anything to me, allowing all those degrees of freedom in the co-fit model and *still* coming up with more or less a single ANI strengthens the case for me that this is not just an abstraction.
If it best fit in the model to model Malayan's relationship to Iran_N, Yamnaya and Onge and all the other South Asian populations in the graph, by making the D13 node, say 70% Iran_N-like and 30% Yamnaya-like, it has freedom to do that. So that it doesn't means something...
qpGraph is a complex tool though, so I'll take your point seriously though, and modify the graph by adding in the Balochi - if the model then fits that as having the same ratio of Steppe:Iran_N there might be more cause for concern than I think - https://pastebin.com/gDKB98H6.
One point I do buy into is @Anthro Survey's that some period of homogenisation could've happened early on between populations in India who then re-expanded and mixed with almost purely ASI groups...
But at the same time, I'm not sure about assuming that the base population wasn't mostly ASI anyway. We can make analogies to Europe, but Europe today does not have populations of agriculturalists who are mostly WHG (contra the South Asian example), so how much does the analogy hold...?
Time will tell though of course.
Did you examine the link above yet?
@ Davidski - "Btw, what's this about Basal Eurasian? No longer a separate thing? So what is it now?"ReplyDelete
Err... nothing definite from me actually. Sorry to give the wrong impression! I'm just wondering aloud about how Basal Eurasian will be defined in future and if it will still work as an idea.
At first it was defined by differences in relatedness to East Asians by ancient West Eurasian populations, then by Ust Ishim as it was found that early Europeans had similar patterns of relatedness to East Asians as Anatolians but similar patterns of relatedness to Ust Ishim as MA-1 and WHG.
Now we have the complexity of relatedness to Tianyuan to add into the models, where the sample has more relatedness to the Goyet / Aurignacian group than present day East Asians do and less relatedness to WHG.
So will a Basal population still make more sense if we move to a model with lots of complicated substructure and admixture between early East and West Eurasian populations? I really don't know so just speculating aloud...
@ Chad: Just had a look now - so the model runs that Natufian and K-14 have mixture from the same basal clade (substantial and slight, respectively) and North from a slightly different basal clade (plus East Eurasian related), while Goyet and Ust Ishim are unadmixed members of the West clade?
Thanks for working all that out and testing it and I think I can see why all those fits happen. The fit looks very tight. Be cool to see what Tianyuan does to it.
We can make analogies to Europe, but Europe today does not have populations of agriculturalists who are mostly WHG"
Fair enough, of course, but key word: TODAY. Who knows what it was like in AD 500 in the Baltic forests?
It's not implausible to imagine scattered, few-in-number HG-dominant survivors hunting and foraging there. In the years following Charlemagne's death, Rome's legacy spread to that corner of Europe and would have certainly caused their extinction and/or assimilation.
In North India, you don't really have ASI dominant groups today, though, aside from Kharias but I consider them a Central Indian group anyway.
In the south, ASI-dominant groups like Paniyars do not make make up a substantial proportion of the overall population.
Nevertheless, eastern India(particularly central and south-east) is roughly analogous to modern European Baltic populations as a whole.
One reason for this could be a more permissible environment: dense jungles and swamps, high humidity, hills(in the Deccan), etc. On the other hand, the environment of the Indus valley would have offered Iran-related farmers a better niche to thrive.
Steppe people could have heavily mixed with Iranian neolithic farmers in bactria/central asia before moving down to India. India was probably an ASI population before their arrival. That may explain why pca/admixture shows a single cline with two component solutions.ReplyDelete
I think it's possible that foragers survived unadmixed in the east baltic as late as 2000 Bc, but certainly not 500 AD.ReplyDelete
The thing is, those PCAs generate a "Han" vertex. With ASI, Han, and West Eurasian verteces, of course it will look relatively linear because nobody has substantial Han-like ancestry in India aside from some Austro-Asiatic speaking Adivasis like Kharia. As for "Mainstream" Indians, only Bengalis have any worth mentioning, clocking in at ~15%.
What in pre-Frankish times would have precluded the existence of small groups harboring direct (non-Yamna, non-EEF) HG ancestry in the 40-60% range in that corner of Europe?
It just seems extremely unlikely, as far as East Baltic is concerned. Maybe northern Scandinavia.
Looking at the archaeological correlates, the last islands of forager culture end c. 2000 BC, which is when the various pre-2000 BC elements (late TRB; CWC; Combed Ware; Narva & other 'sub-Neolithic' groups) finally fused to formed the Baltic Bronze Age. Recall that we have late B.A. samples from the Baltic which confirm this picture.
On the other hand, by 500 AD, actually 0 AD, the East Baltic was very much part of the wider European barbaricum: forts, swords, horse gear, spoils from campaigns with Attila, etc. I just can't think where unadmixed foragers might have hid.
"35,000 year old goyet from belgium a lot like Tianyuan compared to other ancient europeans."
DNA from AMH turned up in a Neanderthal fossil in Germany dated 124.000 years BP - clearly debunking the old guess ("estimate/opinion/hypo") that AMH didn't reach Eurasia before 70.000 BP...
"In the end, the fact that they look like clades could just turn out to be a function of sampling (we have very little in the way of Upper Paleolithic/Mesolithic aDNA across Eurasia), and with more extensive samples we may have to revise our terminological apparatus."
At times the most obvious points are the most important ones.
to the victor go the spoils and history is written by the victors. Science is always science. No interpretation.ReplyDelete
The youngest haplogroup spread across continents and controlling more societies. So they are writing the history. Dienekes proposed J2 people who started Indo Aryan culture and fire worship and than R1a merged into it.. You are proposing R1a . Both are with racist overtones.
We have seen this in linguistics. History says Israilites are slaves with not much education and Phonecians introduced script. Now Aramac is root of all Indo European scripts. May be after another 100 years, English will be root.
There will always be wars between tribes for resources. As long as there is upstream haplogroups like P in south and central Asia it is hard to get convinced with your theories. Europe is remote branch of human expansion. Just natural back migration and expansion of human society. They are not Terminators of I in Europe and H in South Asia and J in Middle East.
AMH 219.000 years ago??ReplyDelete
I'm a litytle concerned that the Neolithic_Iran component to India may contain a pseudo-Yamnaya influence. The Chalcolithic samples from Armenia circa 4000 BCE are all L1a-M27 which is frequent in Dravidian India and Southern Pakistan. These Armenia Chalcolithic date to the Late Ubaid period and have "Steppe-like" admixture but clearly predate Yamnaya.
From YFull and other studies L1a-M27 looks like it originates in the west-Northern Syria, etc...
The "Yamnaya-like component" may have moved into Southern Pakistan well before the I-A migration, probably bringing the linguistic association between Elamite and Dravidian. I do believe that Yamnaya-like folks moved into the the Indus region, but the quantitative effect could be occluded by an earlier Mesopotamian movement into IVC (or creating IVC) and later the L1a's moved south into India itself.
Speaking of Y haplogorups (someone also mentioned H before), Indian Y Hgs all look Western. Y Hg H is found throughout South Asia in high frequencies (north and South), followed by R1a, then J, and L (latter two more patchy).ReplyDelete
Nothing would qualify as ASI or Onge-like, apart from haplogroup O, which is restricted to obviously admixed groups in the east of India.
(I'm sorry for taking so long to respond; 4th of July stuff was/is happening, and after this I have to get back to it. We'll talk more tomorrow)
"Still don't agree; PCA, ADMIXTURE show single clines and two component solutions in India - just doesn't seem compatible with those lines of evidence."
I think therein lies the problem; in my experience, PCA data doesn't really show anything like that.
On the contrary, when modelling these populations using PCA data, one sees considerable independence between the Iran_N-related and Steppe_EMBA-related fractions, but a much more robust correlation between Iran_N and ASI.
Ditto with ADMIXTURE, which shows 0% Baltic-modal components in some Indians (which I assume tracks Steppe_EMBA affinity), but a Baloch-modal component in all of them (which I assume tracks Iran_N affinity).
Taking the PCA and ADMIXTURE evidence in consideration, I think David's topologies with only a single Central/South Asian population tend to make much more sense, in terms of how the gene-flow probably occurred in historical reality.
For example, David's most recent topology has the Kalash as being a 50/50 mix of a population that was 80% Iran_N + 20% ASI and another population that was a twin of Yamnaya, with the Kalash being around 10% ASI.
On the other hand, the Chamar end up being a 80/20 mix of a population that was 70% ASI + 30% Iran_N and another population that was a twin of Yamnaya.
Considering that the Kalash are an Indo-Aryan isolate who live in the southern edges of Central Asia, while the Chamar are a Dalit population from deep inside South Asia proper, I think the pre-Steppe_EMBA South Asian ghost population's Iran_N-to-ASI ratio for both of them make good sense (80% Iran_N/20% ASI for the Kalasha non-Steppe_EMBA ancestry, 70% ASI/30% Iran_N for the Chamar non-Steppe_EMBA ancestry).
At the end of the day, it makes more geographical sense, fits the archaeological record better, and is more in line with modelling based on PCA data (for the latter, I can post some examples tomorrow).
I think you've made a very important observation.
South Asia's Y-DNA heritage is almost completely West Eurasian.
For mtDNA, I think people of caste in South India have around 10%-20% West Eurasian mtDNA, while on the opposite end in South Central Asia different populations tend to be between 65% and 95% West Eurasian in terms of mtDNA lineages (although the Kalash have 100% West Eurasian mtDNA).
Burusho are similar to the Kalash in this regard, but with a lot more East Asian ancestry.
Written below has been my take on y-dna. I omitted the subclades for the sake of simplicity so assume the most relevant ones.ReplyDelete
R1a---mainly from steppe-DNA carriers
L---mainly from Iranian Neolithic farmers, some potentially from steppe carriers admixed with BMAC or similar populations
H---mainly from Iranian farmers
R2---mainly from Iranian farmers
F(xGHIJK)---definitely a strong ASI candidate, peaks on Paniya, present in most S.Indian groups. Virtually absent west of the Indus.
C-m130----another ASI candidate.
Yeah, North and Central Scandinavia are great candidates, too. Somehow overlooked them.
What you say is true, but Northeast Europe was also where deforestation lagged behind the rest. The vegetation could have provided cover for the last remaining forager groups on the continent. Also, it's fair to say it lagged behind the rest in complexity of metallurgy and societal organization in the millenia BC. There is a pattern there. So while you may well be right about the absence of HG-heavy Paniya analogs there by 500 AD, we can be sure that such analogs probably survived there the longest(as well as N. Scandinavia).
"Speaking of Y haplogorups (someone also mentioned H before), Indian Y Hgs all look Western. Y Hg H is found throughout South Asia in high frequencies (north and South), followed by R1a, then J, and L (latter two more patchy).ReplyDelete
Nothing would qualify as ASI or Onge-like, apart from haplogroup O, which is restricted to obviously admixed groups in the east of India."
Although H isn't exclusively South Asian a lot of deep rooted subcaldes (previously classified as F) found in tribes suggest it being there since the upper palaeolithic and not the result of Iranian farmers.
Additionally haplogroup O likely doesn't qualify as ASI or Onge-like since the process of farming which brought L, J2 and R2 from the west also brought O from the east. There's no denying the east Asian affinity of Indian Austroasiatics which can be confirmed through autosomal tests.
"What you say is true, but Northeast Europe was also where deforestation lagged behind the rest. The vegetation could have provided cover for the last remaining forager groups on the continent. Also, it's fair to say it lagged behind the rest in complexity of metallurgy and societal organization in the millenia BC. There is a pattern there. So while you may well be right about the absence of HG-heavy Paniya analogs there by 500 AD, we can be sure that such analogs probably survived there the longest(as well as N. Scandinavia). "
No doubt, I garee with you relatively, just not the absolute time-frame. Agree that proper farming which made anthropogenic impact on the surrounds occurred later, as late as Iron Age, and maybe even with Slavs. Before that was basic 'slash & burn' agriculture and animal hubandry, which began with CWC days.
@ Davidski: https://drive.google.com/file/d/0B9o3EYTdM8lQbWRLZ2c1bmEtbGc/view?usp=sharingReplyDelete
Cheers. So in this topology the "ANI" for populations is (ordered by steppe):
Kalash - 55:45 Steppe:Iran_N
Brahmin_India - 55:45 Steppe:Iran_N
Gond - 49:51 Steppe:Iran_N
Balochi - 35:65 Steppe:Iran_N
Then you have (ordered by "ANI"):
Balochi - 84:16 ANI:ASI
Kalash - 81:19 ANI:ASI
Brahmin_India - 64:36 ANI:ASI
Gond - 22:78 ANI:ASI
It looks like the ratio of Steppe:Iran_N *can* vary appreciably in this model when it has to, as it does to accommodate the Balochi population, with quite a substantial difference in ancestry. The model just doesn't do that for the Gond. I would presume because that would give a worse model fit.
(Also would say that relatively even level in ASI between Kalash and Balochi is also more what I think we would expect to see as well, in contrast to the qpAdm model from Lazaridis et al 2016, where the Baloch had less than half the level of ASI as Kalash! https://1.bp.blogspot.com/-9k5UilpTSEA/V2TqxrBdxgI/AAAAAAAAFCE/awWnaVwRo6cPa2xjxEKuNKRLfmLIM5f3gCPcB/s1600/MODEL.png).
If I can get my latest general model for recent West Eurasia (from the comments to "Iron Age nomads vs Bronze Age herders") working, then I might be able to try and co-fit some Europeans between Yamnaya, Barcin_N and WHG, and then could be another check on whether co-fitting is a problem.
@ Sein: "On the contrary, when modelling these populations using PCA data, one sees considerable independence between the Iran_N-related and Steppe_EMBA-related fractions, but a much more robust correlation between Iran_N and ASI."
"Ditto with ADMIXTURE, which shows 0% Baltic-modal components in some Indians (which I assume tracks Steppe_EMBA affinity), but a Baloch-modal component in all of them (which I assume tracks Iran_N affinity)."
Well, on PCA there's always a single cline though? Main point I was trying to go for (by talking about PCA) was a single cline for South Asian populations. With the exception of the Makrani, Baloch, etc.
So really what we are disagreeing here on is the difference in how we model the end point of that cline. I guess you are saying that projected end points of that cline falls at points suggestive of Steppe_EMBA and a ENA related+Iran_N composite? ("considerable independence between the Iran_N-related and Steppe_EMBA-related fractions, but a much more robust correlation between Iran_N and ASI").
I think when I've tried doing cline projection on that single cline, as well, IRC end point was to a neutral ASI without an Iran_N dimension, and on the ANI end to a North Caucasus like population, rather than a steppe like one. That is, independence between ASI and either Steppe / Iran_N, but correlation between Steppe / Iran_N. If you had that single cline as a tradeoff between Steppe and an ASI+Iran_N composite, then cline projection would put the "north" point of the cline in the steppe.
I remember changes in the NE European related components, but those were pretty minor, IIRC, as a percentage of total ancestry.
Happy 4th of July also!
bit of both imoReplyDelete
i think herders could move into and through farmer territory on marginal land unsuitable for farmers and given the farmer's likely advantage in numbers it could have been symbiotic for a while and then at some point it turned violent - maybe after some kind of disaster weakened the farmers or after the herder's numbers got too big.
i think something similar may have happened with Hittites, Sumer/Akkad, LBK etc
While this thread has been informative and interesting, might we address the original point? I do believe it is key to understanding why a bronze age steppe migration is resisted so fiercely in South Asia.ReplyDelete
"Happy 4th of July also!"
Thanks! I really appreciate that (it was a very enjoyable day).
"Well, on PCA there's always a single cline though? Main point I was trying to go for (by talking about PCA) was a single cline for South Asian populations. With the exception of the Makrani, Baloch, etc."
Ah, so If I'm not mistaken, you have in mind the first 2 dimensions of a West Eurasian PCA?
In that case, the Kalash do cluster fairly close to Lezgins/Chechens, and other South Central Asians deviate from that point in the direction of the Iran_N samples.
But I would chalk that up to the broad similarity that exists between the deep genetic ancestry of South Central Asians and North Caucasians.
I mean, both the Kalash and the Lezgins have somewhat similar amounts of Steppe_EMBA (with the Kalash showing a little more), and both the Kalash and Lezgins have much of the rest of their ancestry being of the Caucasian/Caspian clade (CHG and Iran_N).
When using more than 2 dimensions, there is a clear independence of pattern for Iran_N and Steppe_EMBA, in relation to South Asia.
For example, non-Brahmin South Indians can be modeled as Iran_N + ASI, with virtually no Steppe_EMBA.
By contrast, North Indian Chamar show as much ASI as non-Brahmin South Indians, but have noticeable Steppe_EMBA.
So, with regard to Iran_N and Steppe_EMBA, one sees a clear geographic patterning that doesn't map unto a unitary ancient "ANI" population.
Below, I'll post some old nMonte fits, ones which are fairly typical, and pretty much demonstrate the heterogeneity.
These should also be of interest for the fact that David Reich and company seem to be moving away from formal methods.
Here are some examples:ReplyDelete
Southern Central Asia
50.55% Iran_Chal + 1.75% Iran_Meso/Neo
42.95% Steppe (an average of the Srubnaya_outlier, the Potapovka_outliers, and the Srubnaya people)
4.75% East_Asian_ENA (an average of Ulchi, Mongola, Yi, Han, and Buryat)
ASI=0% (note: ASI is an average of Andamanese_Onge, Andamanese_Jarawa, and Austroasiatic_Bonda)
40.75% Iran_Chal + 2.75% Iran_Meso/Neo
25.9% Iran_Chal + 14.2% Iran_Meso/Neo
26.60% Iran_Meso/Neo + 16.80% Iran_Chal
Sarbani Pashtun, South
36.8% Iran_Meso/Neo + 21.4% Iran_Chal
Karlani Pashtun, Central
27.6% Iran_Meso/Neo + 25.1% Iran_Chal
Batanri Pashtun, South
38.25% Iran_Meso/Neo + 15.00% Iran_Chal
51.2% Iran_Meso/Neo + 0.6% MA1
(a note: when using Yamnaya, the Kalash receive 45%, just like what we see with David's qpGraph topology. The lower steppe-related percentage is due to the much greater ANE shift seen with the Srubnaya_outlier)
As is quite evident, most of these populations are between 90%-85% West Eurasian (with both ASI and Siberian-related admixture), and the West Eurasian ancestry is a mix of Iran_N, Iran_Chal, and Steppe_EMBA, but the proportions don't suggest a unitary ANI base.
Also, the Kalash are outliers, as they have no Iran_Chal, and no Siberian-related admixture. The latter is probably an indication of isolation, but the lack of Iran_Chal is more of a mystery.
Just a side-note, but I should mention that David's qpGraph topology involving only the Kalash matches this nMonte fit (PCA nMonte's fits of either 45% Iran_N + 45% Steppe_EMBA + 10% ASI, or 50% Iran_N + 40% ANE-rich Steppe_MLBA outliers + 10%, are basically identical to the trees for the Kalash, the ones in which only they have to be fit).
Now, looking at South India:
24.40% Iran_Neo/Meso + 1.15% MA1
32.5% Iran_Meso/Neo + 3.8% MA1
47.40% Iran_Meso/Neo + 8.25% MA1
The West Eurasian admixture is quite distinct from what one sees in southern Central Asia.
No Steppe_EMBA, and no Iran_Chal, only Iran_N (with some extra ANE, suggestive of an Iran_Hotu-like population).
(Sorry for the consecutive posts)ReplyDelete
With regard to ADMIXTURE, the Baltic-centered component tends to disappear in South India (with exception to Brahmins), but reaches around 25% among some North Indian Jatts.
By contrast, components indicative of ancient ancestry from the Iranian plateau, like Baloch-centered and Georgian-centered components, are everywhere in India (at the very least, they show a decent presence in South India).
Just something of interest, I think.
Also, the models above are not meant to be taken more seriously than qpGraph.
On the contrary, qpGraph involves a very powerful method, and I would take it very seriously.
Still, I would note that the co-fitting topologies don't gel with PCA and ADMIXTURE results, and I would also note the models with only a single Central/South Asian test population tend to match nMonte output.
Regardless, aDNA will give us the clarity that we want, and thankfully the data is coming very soon. All our questions will be answered.
But, it's still fun to speculate.
Sein: Ah, so If I'm not mistaken, you have in mind the first 2 dimensions of a West Eurasian PCA?ReplyDelete
That, but also clines within PCA on the South Asian populations alone, where IRC you usually have a single cline, and where there's deviation from that, it tends to be driven by either East Asian ancestry, Iran_N (among Balochi, Makrani, etc) or outlier drift effects (if all East Asian and Iran_N enriched are removed).
IRC just using these populations alone, not really a break from a single cline (so not like multi-dimensional structure is being obscured by other dimensions more prominent in West Eurasia wide PCA).
I've just scanned your comments for the word 'plague' and 'epidemic' and didn't find them, but maybe something similar is in there.ReplyDelete
A highly mobile population containing a virulent virus that the population already has some immunity to can have devastating effects on neighbours (I think it would have to be this way round and not an effect of trade amongst the neighbours carrying plague, but could be wrong).
Think Spaniards in the Americas and a 90% die off over three hundred years. Also note the Olalde 2017 paper talking of perhaps 75% or higher population replacement in Britain in the long term (short term 90+%).
This helps a lot when trying to take over someone else's land, though a good bit of violence always helps.
For a good reference on the presence of Yersinia pestis during expansion (which David Reich has alluded to) see Rasmussen 2015 (Early Divergent Strains of Yersinia pestis in Eurasia 5,000 Years Ago) although I'm guessing you already know.
Just bear it in mind as an alternative.
Ned (armchairprehistory.com - most of which is now wrong)