search this blog

Saturday, March 31, 2018

Andronovo pastoralists brought steppe ancestry to South Asia (Narasimhan et al. 2018 preprint)


Over at bioRxiv at this LINK. Note that the Andronovo samples that are shown to be the best fit for the steppe ancestry in South Asians are labeled Steppe_MLBA_East (ie. Middle to Late Bronze Age eastern steppe). Below is the abstract and a couple of key quotes from the paper and its supp info PDF. Emphasis is mine:

The genetic formation of Central and South Asian populations has been unclear because of an absence of ancient DNA. To address this gap, we generated genome-wide data from 362 ancient individuals, including the first from eastern Iran, Turan (Uzbekistan, Turkmenistan, and Tajikistan), Bronze Age Kazakhstan, and South Asia. Our data reveal a complex set of genetic sources that ultimately combined to form the ancestry of South Asians today. We document a southward spread of genetic ancestry from the Eurasian Steppe, correlating with the archaeologically known expansion of pastoralist sites from the Steppe to Turan in the Middle Bronze Age (2300-1500 BCE). These Steppe communities mixed genetically with peoples of the Bactria Margiana Archaeological Complex (BMAC) whom they encountered in Turan (primarily descendants of earlier agriculturalists of Iran), but there is no evidence that the main BMAC population contributed genetically to later South Asians. Instead, Steppe communities integrated farther south throughout the 2nd millennium BCE, and we show that they mixed with a more southern population that we document at multiple sites as outlier individuals exhibiting a distinctive mixture of ancestry related to Iranian agriculturalists and South Asian hunter-gathers. We call this group Indus Periphery because they were found at sites in cultural contact with the Indus Valley Civilization (IVC) and along its northern fringe, and also because they were genetically similar to post-IVC groups in the Swat Valley of Pakistan. By co-analyzing ancient DNA and genomic data from diverse present-day South Asians, we show that Indus Periphery-related people are the single most important source of ancestry in South Asia — consistent with the idea that the Indus Periphery individuals are providing us with the first direct look at the ancestry of peoples of the IVC — and we develop a model for the formation of present-day South Asians in terms of the temporally and geographically proximate sources of Indus Periphery-related, Steppe, and local South Asian hunter-gatherer-related ancestry. Our results show how ancestry from the Steppe genetically linked Europe and South Asia in the Bronze Age, and identifies the populations that almost certainly were responsible for spreading Indo-European languages across much of Eurasia.

...

Third, between 3100-2200 BCE we observe an outlier at the BMAC site of Gonur, as well as two outliers from the eastern Iranian site of Shahr-i-Sokhta, all with an ancestry profile similar to 41 ancient individuals from northern Pakistan who lived approximately a millennium later in the isolated Swat region of the northern Indus Valley (1200-800 BCE). These individuals had between 14-42% of their ancestry related to the AASI and the rest related to early Iranian agriculturalists and West_Siberian_HG. Like contemporary and earlier samples from Iran/Turan we find no evidence of Steppe-pastoralist-related ancestry in these samples. In contrast to all other Iran/Turan samples, we find that these individuals also had negligible Anatolian agriculturalist-related admixture, suggesting that they might be migrants from a population further east along the cline of decreasing Anatolian agriculturalist ancestry. While we do not have access to any DNA directly sampled from the Indus Valley Civilization (IVC), based on (a) archaeological evidence of material culture exchange between the IVC and both BMAC to its north and Shahr-i-Sokhta to its east (27), (b) the similarity of these outlier individuals to post-IVC Swat Valley individuals described in the next section (27), (c) the presence of substantial AASI admixture in these samples suggesting that they are migrants from South Asia, and (d) the fact that these individuals fit as ancestral populations for present-day Indian groups in qpAdm modeling, we hypothesize that these outliers were recent migrants from the IVC. Without ancient DNA from individuals buried in IVC cultural contexts, we cannot rule out the possibility that the group represented by these outlier individuals, which we call Indus_Periphery, was limited to the northern fringe and not representative of the ancestry of the entire Indus Valley Civilization population. In fact, it was certainly the case that the peoples of the Indus Valley were genetically heterogeneous as we observe one of the Indus_Periphery individuals having ~42% AASI ancestry and the other two individuals having ~14-18% AASI ancestry (but always mixes of the same two proximal sources of AASI and Iranian agriculturalist-related ancestry). Nevertheless, these results show that Indus_Periphery were part of an important ancestry cline in the wider Indus region in the 3 rd millennium and early 2 nd millennium BCE. As we show in what follows, peoples related to this group had a pivotal role in the formation of subsequent populations in South Asia.

...

These results—leveraging our rich data from ancient samples closer in time to the Bronze Age—show that the group(s) that contributed Iranian agriculturalist-related ancestry to South Asia shared more genetic drift with the Iranian agriculturalist-related groups in our dataset that are temporally and geographically closest, compared to Caucasus HGs (CHG) or early Zagros related agriculturalists previously shown to be related to source populations for South Asians (11, 81). We are not only able to exclude these early farming and hunter-gathering groups, but also Copper and Bronze Age groups in western Iran (Seh_Gabi_C and Hajji_Firuz_C), and even in eastern Iran and Turan (Tepe_Hissar_C, Gioksiur_EN, and BMAC). Our detailed analyses in Text S3 indicate that what is driving the failure of these models is an excess of Anatolian agriculturalist-related ancestry in all of these groups, suggesting that the Iranian agriculturalist-related population that mixed into South Asia had less Anatolian agriculturalist-related ancestry than all of these. However, we find that mixtures using the Indus_Periphery sample (a pool of three outlier individuals from the BMAC site of Gonur and from Shahr-i-Sokhta), provides an excellent source population for the Iranian agriculturalist-related ancestry in South Asia when combined with any individuals in the Steppe_MLBA cluster (Srubnaya, Sintashta_MLBA, Steppe_MLBA_West or Steppe_MLBA_East).


Narasimhan et al, The Genomic Formation of South and Central Asia, Posted March 31, 2018, doi: https://doi.org/10.1101/292581

Update 12/04/2018: The dataset from the prerprint has been made available early at the Reich Lab website here. I've already started analyzing it. You can see the results in several new threads, for instance here, here and here.

See also...

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

Central Asia as the PIE urheimat? Forget it

Ancient herders from the Pontic-Caspian steppe crashed into India: no ifs or buts

868 comments:

1 – 200 of 868   Newer›   Newest»
Unknown said...

And there is proof of BMAC migration into South Asia. Also, there is no proof of steppe Mlba contributing to the main BMAC cluster while the influence of the Harappans on BMAC main cluster appears to be about 30 %.

There are only a few outliers around BMAC that show possible links with the steppe but their dates are 2100 - 1700 BC while the dates for steppe mlba are 2000 - 1400 BC.

All of this damning evidence makes the case proposed by this paper very weak.

Unknown said...

The 1st line should read "no proof of BMAC migration into South Asia"

Unknown said...

Further there is evidence of migration from BMAC on the steppe.

Davidski said...

@Jaydeepsinh Rathod

BMAC never expanded onto the steppe, and certainly never made it to Eastern Europe. At best, the descendants of BMAC probably made it onto the steppe during the Iron Age.

So quit making things up.

Davidski said...

Gotta say, I'm very proud of this little effort from almost a couple of years ago. I didn't exactly hit the nail on the head, but the fundamentals were there. Nice!

Through time AND space?

Jaydeep said...

I am not making things up. I am only saying whats been said in the paper. There are clearly steppe outliers with Turan admixture in the steppe_mlba stage. Why don't you read it carefully ?

Nirjhar007 said...

looks like L-657 IS again absent from steppe...it does not look like steppe in Origin ..

Nathan said...

It's been clear since Europeans started to look into South Asian pre-history, and more recently with the advent of population genetics, that Steppe herders invaded/colonized South Asia sometime around 4000 B.P. The denialists can't sweep away the stark reality of the genetic evidence.

Unknown said...

You should stop rambling nonsense and read the paper if you can.

old europe said...

Nathan

yes both genetics and archeology support this. I totally agree ( at least) with this part of the puzzle

mzp1 said...

Davidsku this is a poor poor effort from you and the papers authors.

These are clearly IRANIAN Scythians moving into Southern territory and clearly far far too late to be bringing Vedic culture into India. Lol. Come on man, have you even looked at Rigveda once. This little paper posits a migration into India around 1000BC, which is far too late. Infact it is quite close to Scythians in time and space.

The Aryan Migration Theory is already dead to me.

Sintashta is the most advanced steppe culture yet it built for the purpose of exporting metal to bmac. Clearly this was a bmac outpost, or otherwise was close to bmac.

Steppe culture settles down and learn complex metallurgy just to export to a pre-existing advanced metallurgical culture vs pre-existing advanced metallurgical culture creates outpost to utilise local deposits.

Davidski said...

Oops, there's an R1b-M269 in the Chalcolithic samples from Hajji Firuz, NW Iran.

Hajji_Firuz_C I2327 (5900-5500 BCE) R1b1a1a2a2

Hmm, in regards to their genome-wide ancestry, from the supp info...

Additional working models include an additional source of ancestry from a population related to hunter-gatherer populations from Karelia or West Siberia, but the proportion of this ancestry is small.

Page 114

Davidski said...

@mzp1

Don't be ridiculous. Scythians had a lot of East Asian ancestry.

So they can't explain much of the steppe input in South Asia. As per this paper, only certain Sintashta/Andronovo groups can, considering both genome-wide ancestry and Y-haplogroups.

mzp1 said...

Only Eastern Scythians had a lot of East Asian and we dont know when and where that admixture first took place. It is likely quite late given it is isolated to the East.

Matt said...

Right just beginning to read, salient bits:

"In the far eastern part of this cline (eastern Iran and Turan) we also detect admixture related to West_Siberian_HG, proving that North Eurasian admixture impacted Turan well before the spread of Yamnaya-related Steppe pastoralists (Steppe_EMBA)."

The great majority of individuals fall in a genetic cluster that is similar, albeit not identical, to the preceding groups in Turan in harboring a large proportion of early Iranian agriculturalist-related ancestry (~60% in the BMAC) with smaller components of Anatolian agriculturalist-related ancestry (~21%) and West_Siberian_HG-related ancestry (~13%)

"First, around ~2300 BCE in Turan, we observe two outliers at the BMAC site of Gonur with West_Siberian_HG-related ancestry of a type that we observe at multiple sites in Kazakhstan over the preceding and succeeding millennia. The most plausible explanation is that this ancestry is that of indigenous populations associated with the Kelteminar culture, the native hunter-gatherers of the region who covered a vast area of Central Asia before the BMAC.

Future ancient DNA data from Kelteminar contexts will make it possible to determine whether it is indeed the case that the genetic ancestry of Kelteminar people was similar to that of West_Siberia_HG. Importantly, in the 3rd millenium BCE we do not find any individuals with ancestry derived from Yamnaya-related Steppe pastoralists in Turan. Thus, Steppe_EMBA ancestry was not yet widespread across the region."


BMAC has EEF, not Levant_N? That will need some retesting...

Third, between 3100-2200 BCE we observe an outlier at the BMAC site of Gonur, as well as two outliers from the eastern Iranian site of Shahr-i-Sokhta, all with an ancestry profile similar to ancient individuals from northern Pakistan who lived approximately a millennium later in the isolated Swat region of the northern Indus Valley (1200-800 BCE).

These individuals had between 14-42% of their ancestry related to the AASI and the rest related to early Iranian agriculturalists and West_Siberian_HG. Like contemporary and earlier samples from Iran/Turan we find no evidence of Steppe-pastoralist-related ancestry in these samples.


Ergo, indirectly stated, Swat population (1200-800BCE) had no Steppe-pastoralist related ancestry?

EastPole said...

I have just started reading it but there seems to be a link between Central-Eastern Europe and India look at Tripolye/CWC/Sintashta/Andronovo and India trail:

I marked it in red here:

https://s9.postimg.org/t7iuhlba7/screenshot_353.png

and in green here:
https://s9.postimg.org/mttrejw67/screenshot_354.png

Nirjhar007 said...

''By 1500 BCE, there were numerous individuals in the Kazakh Steppe with East Asian-related admixture, the
same type of ancestry that was widespread by the Scythian period (34). This ancestry is hardly present in the
two primary ancestral populations of South Asia—ANI and ASI—suggesting that Steppe ancestry widespread
in South Asia derived from earlier southward movements.''

Wow ...

postneo said...

Something wrong with figure 1d the component colors. They don't match the legend or any of the other plots?
Has anyone asked the authors?

Swat and BMAC are mostly Iran neolithic with some Anatolian. How does it explain the huge purported influx of steppe MLBA to South Asia? They are directly in the path of where this influx should have Come from.

The admix plot 1d and ancestry proportion 2C contradict whats written.
There the south asian samples don't show MLBA affinity but are more west siberian.

Seems like the authors left some unrevised figures that are inconsistent with their narrative.

Davidski said...

@Nirjhar

Yeah, so the Andronovo people that contributed ancestry to Indians moved out of the Kazakh steppe by 15000 BCE. But that doesn't mean they were in India already around 2000 BCE.

And, obviously, most R1a in India is from them, not from Scythians.

Nirjhar007 said...

There is not a single L657 Dave....neither any R1a on those Early Swat samples.Did Aryan women Invade?...

Anonymous said...

@Davidski
Is this the mythological South Asia paper? Or another one by different people?
Either way, what a glorious paper.

Davidski said...

@Nirjhar

The Eneolithic R1a sample from Ukraine belongs to a lineage directly ancestral to L657.

Cpk said...

Reich calls Yamnaya Late Proto-Indo-European speakers. He probably has data on the early speakers.

Nirjhar007 said...

The Eneolithic R1a sample from Ukraine belongs to a lineage directly ancestral to L657.

So where are his descendants? vanished? or directly came to S Asia via airplane ? :)...

Davidski said...

@Nirjhar

So where are his descendants? vanished? or directly came to S Asia via airplane?

His paternal descendants, or those of his close relatives, were the Andronovo people and they moved to South Asia via the steppe.

When the bam files from this paper become available we might see some L657 in the data from Central Asia.

Davidski said...

@Cpk

I don't know, but I'd say that David Reich thinks Hajji_Firuz_C from NW Iran represents an early PIE community, because the earliest instance of R1b-M269 is in this sample set.

However, if so, this doesn't square with the consensus amongst historical linguists, who see the PIE homeland on the Eneolithic steppe, in Sredny Stog and/or Khvalynsk.

It'll be interesting to see how this plays out, and certainly Hittite data will be crucial.

Cpk said...

@Davidski He seems insistent on it. I hope he isn't doing it on political/emotional grounds. Just tell the truth.

Davidski said...

The data will do the talking for itself in the end, but speaking of data, some of the Y-haplogroup assignments in the paper's supp info spreadsheet look way off.

There could be quite a few mistakes there.

Matt said...

Ugh, think I might have misunderstood about Swat, will refrain from posting while still thinking for a bit.

@Sein, we were discussing whether Patterson's new model would change the proportions in South Asians of what they call here "Onge (AASI)-related" ancestry in South Asians.

To check, the models in the supplement - qpAdm and hierarchical - give Pathan/Kalash percentages of about 28% steppe_MLBA related, 67.5% Indus_diaspora related and 4.5% AASI. It's not clear how much AASI ancestry Indus_diaspora related has, but "we observe one of the Indus_Periphery individuals having ~42% AASI ancestry and the other two individuals having ~14-18% AASI ancestry", which should work out to an average of about 23% (average of two individuals at 14% and one at 42%).

So that would seem to give total AASI in Pathan/Kalash about 15.52% Indus_diaspora related (0.23*67) and 4.5% elsewhere, so about 20%. Pretty much the same as the f4 ratio with no real changes there.

Unknown said...

@Davidski
PIE in Hajji_Firuz_C would fit with the wanderworter "wine" in many language families of the area and with the early production of wine at Hajji Firuz Tepe.

Ariel said...

There are 27 Y-dna from South Asia and only one of them is R1a, from what I understand is from Saidu_Sharif Swat valley, one the "younger" (500-300BC) sample: I6891.

Ariel said...

Someone should ask Reich why Iran and not the Caucasus, I think the caucasus is far more likely given the data that we have on the "southern" CHG admixture in Yamna, if R1b was in Iran at very very low percentages, we should assume that there was R1b in the caucasus too. And don't we have a Kura araxes with R1b already?

Davidski said...

Let's wait until the Y-haplogroup assignments make more sense. There are two instances of R1b-M269 in the Iron Gates HG samples. These samples weren't R1b-M269 before.

old europe said...

I always suspected that the PIE urheimat had to be in a mountain region for reason that i don't have enough time here to explain. I thought it could be the alps but fine if it will be the caucasus and armenia georgia
Olympus really a resurrection night for you!!!

Olympus Mons said...

Yes... A glorious day for me. "My" Sulaveri deserve it. :)

Davidski said...

@Olympus Mons

You're banned from here.

And no, your ridiculous "thesis" was way off. No migration via Africa for starters. Now piss off.

@All

Please don't encourage this troll to post here.

Archaelog said...

Ha finally. Good work reporting so fast David!

Davidski said...

@Rob

As much as you wish, you can't pick this apart.

South Asians have considerable ancestry from the Pontic-Caspian steppe. This is obvious in their genome-wide DNA, Y-DNA and even mtDNA.

And as per the paper, the most plausible source of this ancestry are the Andronovo people.

Davidski said...

From the paper, page 12...

However, at least some (possibly all) of the Steppe pastoralist ancestry in South Asia owes its origins to southward pulses in the 2nd millennium BCE, as indeed we prove directly through our observation of this ancestry in the Swat Iron Age individuals dating to ~1000 BCE (discussed further below).

vAsiSTha said...

Eh, why does the Ganj Dareh 8000BC sample id I1949 have Y haplogroup R1?

Davidski said...

Holy shit, MA1 (Mal'ta boy) is also listed as R1b1a1a2.

Pft...

Chad said...

Looks like I was spot on with Andronovo, ANE presence, and Anatolian ancestry being widespread.

Davidski said...

From the paper, page 14...

Finally, we examined our Swat Valley time transect from 1200 BCE to 1 CE. While the earliest group of samples (SPGT) is genetically very similar to the Indus_Periphery samples from the sites of Gonur and Shahr-i-Sokhta, they also differ significantly in harboring Steppe_MLBA ancestry (~22%). This provides direct evidence for Steppe_MLBA ancestry being integrated into South Asian groups in the 2 nd millennium BCE, and is also consistent with the evidence of southward expansions of Steppe_MLBA groups through Turan at this time via outliers from the main BMAC cluster from 2000-1500 BCE. Later samples from the Swat time transect from the 1st millennium BCE had higher proportions of Steppe and AASI derived ancestry more similar to that found on the Indian Cline, showing that there was an increasing percolation of Steppe derived ancestry into the region and additional admixture with the ASI through time.

Archaelog said...

@All That M269 from Iran could be a massive red herring you know. M269 is a very old clade and there is P297 all over Russia which dates older or same as this. I would reserve judgement I am travelling and will give the paper a proper read as soon as I can

Alberto said...

Amazing set of samples!

But yes, something is wrong. Did they mix up all the data or whoever wrote the main text and conclusions didn't look at it?

Maybe the rushed the preprint because of the publishing date of the book?

Rob said...

Moreover, it seems odd to call it "Siberian HG" admixture in Turan
It is simply central Asian hunter-gatherers.

@ Dave
"Oops, there's an R1b-M269 in the Chalcolithic samples from Hajji Firuz, NW Iran."

Thats huge. 5000 BC. Looks Majkops could indeed be the link.
OF course, R1b could still be from anywhere around Black Sea, but it does show the 'nuanced links & networks' long-predating Yamnaya i had been suggesting. .

I remember all the "experts" on Anthrogenica pontificating about female-mediated this, lack of south-Caucasian that. I hope they can take the time to reflect on their need for less talk, less dogmatic chavanism and more learning.

"As much as you wish, you can't pick it apart"

Just offering friendly suggestions.

Samuel Andrews said...

Yeah, the Y DNA analysis has mistkes. One of the Iron Gates HGs is labelled to a super derievd R1b-U106 clade. So the R1b M269 result in Chalcolithic Iran is proabbly wrong.

@Rob,

Steppe ancestry in Greece, Iberia, Balkans, South Asia in 2nd millenium BC. Pretty obvious what's going on.

Alberto said...

Yes, at least the part of the excess ANE in SC Asia is clear. So one debate less!

"North Eurasian hunter-gatherer ancestry in Central Asia before the Yamnaya expansion.
We observe significant West Siberian hunter-gatherer related ancestry (also related to Ancient North Eurasians (ANE) and Eastern European HGs (EHG)) in individuals from Iran and Turan during the EN period. Of particular note is that the samples from Tajikistan from the site of Sarazm, directly radiocarbon dated to the mid-4th millennium BCE, have about 23% of their ancestry attributable to this source. This indicates that this ancestry cannot be attributed to admixture with Steppe pastoralist populations such as the Yamnaya and their successors, as these archaeological cultures are observed at least 700 years later. It is tempting to speculate that this ancestry is also characteristic of the hunter-gatherer Kelteminar culture, which archaeological evidence suggests was dominant in this region in this time, although we do not have any genetic data directly from this population.
"

Seinundzeit said...

Finally!!!! (Lol, sorry for the overdone punctuation, but I've been waiting to see this paper for so long)

In an hour, I have an emergency session at work (pricks can't even spare me a meeting on Easter Weekend), so I can't dive in now.

But tonight, I will study this thing, and provide questions and (constructive) criticism in a couple days....

Alberto said...

The West_Siberian_HG admixture was present in the presumed migrants from the IVC found in east Iran and BMAC, from 3100 and 2700 BCE respectively. So yes, the name given to this ancestry is a bit of a stretch. But that's just semantics.

Archaelog said...

@Rob lol What is your take on this?

vAsiSTha said...

Ganj dareh I1944 8200BC is R2a. I1949 8000BC is R1. These seem to be major ancestors to R2a & R1 in India.

Otherwise looks like Miss/Mrs. Aryas invaded enmasse into Punjab.

Ron said...

ok so ivc people were already a mix between iranian farmers and ASI (more iranian and less ASI). then steppe MLBA people mixed with these indus periphery people as they call it. this is something everyone knew all along (except for the MLBA part). Did these MLBA people impact southern india?

Rob said...

I guess the R1b could be some other kind , like that found in khvakysnk & Armenia EBA
Still , it completes a link
Shame we can’t play with raw data for a while

Dave -
Are you going to admit I wa correct ?

Davidski said...

@Kulkarni

Quit acting stupid and read the paper properly, or I'll start deleting your posts here.

There's no R1a-Z93 or even R1a in any of the ancient Iranian samples.

Shaikorth said...

Sein's models with Srubnaya_outlier being an important component around South Asia are also explained by the paper's findings of West Siberian HG (or related) ancestry in SC-Asia, these qpAdm fits for Srubnaya outlier and similar samples imply high WSHG ancestry:

0.084 Iron_Gates_HG 0.278 Ganj_Dareh_N 0.638 West_Siberia_N

or

0.193 Karelia_HG 0.321 Ganj_Dareh_N 0.486 West_Siberia_N

Davidski said...

Holy crap! From the paper, page 16...

Our analysis also provides an entirely new line of evidence for a linkage between Steppe ancestry and Indo-European culture. When we used qpAdm to test if a mixture of ANI and ASI is a fit to the data for all 140 Indian Cline groups, we found 10 groups with poor fits and a significantly elevated ratio of Steppe_MLBA- to Indus_Periphery-related ancestry compared to the expectation for the model (Z≥3). We found the strongest two signals in Brahmin_Tiwari (p=2×10 -5 ) and Brahmin_UP (p=4×10 -5 ), and more generally there was a striking enrichment of a Z≥3 signals in groups of traditionally priestly status in northern India (57% of groups with Z≥3 were Brahmins or Bhumihars even though these groups comprised only 11% of the 74 groups we analyzed in northern India). Although the enrichment for Steppe ancestry is not found in the southern Indian groups, the Steppe enrichment in the northern groups is striking as Brahmins and Bhumihars are among the traditional custodians of texts written in early Sanskrit. A possible explanation is that the influx of Steppe_MLBA ancestry into South Asia in the mid-2nd millennium BCE created a meta population of groups with different proportions of Steppe ancestry, with ones having relatively more Steppe ancestry having a central role in spreading early Vedic culture. Due to strong endogamy in South Asia—which has kept some groups isolated from their neighbors for thousands of years (35)—some of this substructure within Indian population still persists.

Arza said...

Fig. 1 Overview of ancient DNA data
E Y chromosome haplogroups
Page 21
https://s6.postimg.org/yxrwid8jl/y_haplo.png

vAsiSTha said...

@davidski

"Although the enrichment for Steppe ancestry is not found in the southern Indian groups, the Steppe enrichment in the northern groups is striking as Brahmins and Bhumihars are among the traditional custodians of texts written in early Sanskrit."

No major conclusion from this. Already existing Brahmins would be high status people and preferred mates.

mzp1 said...

@davidski,

I suppose they compare brahmins to local populations but do they have comparison with different (pakistani) punjabi castes and gujjars? These are the main people living in the area that was once IVC and was, according to you, invaded by Aryans.

I say this because knowing the Steppe component in average Pakistanis would give us a hint to the level of population replacement required by the invasion theory.

EastPole said...

Dave, how do you interpret this:

Page 18
“Fig. 3 The Genomic Origins of Indians. (A) We used qpAdm to model four groups that are
545 representative of major sources of South Asian ancestry over the last few thousand years
546 (Punjabi.DG, Mala.DG, SPGT, and Butkara_IA) as mixtures of Onge, an Iran/Turan-related
547 population, and a Steppe-related group, and report the minimum p-value (highlighting cases at
548 p>0.01). The only working models involve a combination of Indus_Periphery and a
549 Steppe_MLBA group (note that the Steppe_MLBA_West group includes a subset
550 Sintashta_MLBA and Srubnaya).”

My first thought is that it tells us that Indo-Iranians most likely originated as a mix of Balto-Slavs from CWC who on their trek east admixed with some steppe folk and then mixed with the population of Indus Valley Culture.
So they confirm what we have been speculating about for long time.

Shaikorth said...

The authors are suggesting Austroasiatics are relatively old in S-Asia, perhaps even the oldest major language group if Dravidian is related to Iranian Neolithic and not native.

"The Juang are an Austroasiatic speaking group in India which our PCA analyses in Data S4 show have a low proportion of West Eurasian ancestry. We were unable to model Juang as a mix of ASI and a source that was a clade with Nicobarese (isolated Austroasiatic speakers from the Nicobar Islands). However qpGraph obtains an excellent fit by adding a substantial component of AASI ancestry to Juang (Figure 3C). In other words, the Juang have too much AASI- related ancestry relative to ancient Iranian agriculturalists to be a simple two-way mixture of a Nicobarese-related population and ASI. These results suggest that Austroasiatic speaking groups were in peninsular India at a time when there were still populations that had little if any Iranian agriculturalist-related admixture."

vAsiSTha said...

@ davidski

Plz help explain to a noob.

Out of 35 pakistan male samples from periods ranging 1300bc to 40BC, only 1 has R1a downstream 400BC.

How & when did R1a get into Punjab? Whats your hypothesis?

Rob said...

Another interesting possible link: J2b Iran Chalc - Armenia BA - Croatia MBA

Davidski said...

@All

Hold your horses in regards to that R1b-M269 in Hajji_Firuz_C.

That sample is the only one from that group that hasn't been carbon dated. And apparently he's an outlier.

Shaikorth said...

@Rob
What even is Parpola's precise model? I'm not an expert on his stuff, just superficially read some papers. Someone else mentioned @Anthrogenica that a relative lack of BMAC ancestry in South Asia supports his model because he said BMAC and Indo-Aryans were hostile, that was actually news to me.

Alaron said...

I2312_d K2a E1b1a1a1c2b1 Belt_Cave_Mesolithic_LC 12000-8000 BCE Iran

E1b1b from Mesolithic Iran. WOW. Probably can explain the high Basal Eurasian there.

Rob said...

@ Shaik

No I was just joshing mate. It was more in relation to other end of the steppe - IRC you were supporting his/ Anthony's model.

BTW what do you make of the relevance of Zevakinsky BA ?

vAsiSTha said...

Just to add:

Bhumihars were classified as shudras by Britishers, though they started a movement to get themselves identified as Brahmins.

Can we at least conclude that R2a in India is from Iran neolithic?

Rob said...

Where are the Botai samples ??

Davidski said...

@Alaron

E1b1b from Mesolithic Iran. WOW. Probably can explain the high Basal Eurasian there.

MA1 is shown to belong to R1b-M269, while one of the Swat samples belongs to a young R1b-U106 subclade, and so on, and so forth.

So I'd wait for the revised spreadsheet.

Shaikorth said...

@Rob

The oldest Zevakinsky? I'd say it looks like a mix of Dali_EBA (West Siberian HG + Iran_N "pseudo ANE/EHG/steppe") and a western steppe invader.
When it comes to the IE origin model, looks like the Yamnaya/steppe fit will work for surviving I-E languages if Reich is right about Iran being the ultimate origin and Hittite branching out earlier. If M269 is from Iran and never went to steppe things might get complicated but in light of that MA-1 call every R1b (and probably the rest when you're at it) in the study should be re-examined.

Rob said...

@ Shaikorth

I was asking about the Zevakinsky with some ENA in Fig 2; and its relevance for F-U.

Davidski said...

@Rob

I'm pretty sure South Asian R1a is more closely related to European R1a, than to anything from Neolithic Siberia.

And obviously Andronovo_SE is a mixture of Andronovo_NW plus various groups along the way from Eastern Europe to the Hindu Kush.

So the equal C14 dates aren't important, because obviously Andronovo moved from NW to SE, not the other way around.

Rob said...

BTW Other people have seen R1b in Afantova., although calls are conflicting as it also could be Q. I think the R1b tree is awaiting an upgrade and new hg set.

Chad said...

I told Dr. Narasimhan about the issues with the Y-DNA table. I'm sure they'll look it over again.

Shaikorth said...

@Rob
The Zevakinsky_LBA samples then. Y-DNA R1a/R1b/Q1a and 3000 years old in eastern Kazakhstan, these look connected to eastern Scythians.

Bronze said...

hahahahahahaha this is awesome, all of you talking shit about India being invaded when we didnt even have any ancient dna from the region, now youre proven wrong and backtracking.

There definitely was no aryan invasion into India, and EastPole you need mental help there was absolutely not any balto-slavic migration into South asia ever. Thats confirmed, R1a is not from Europe.

Chad said...

^^ someone didn't actually read the paper.

Rob said...

@ Dave

Obviously Z93 is closer and younger than some side branch Siberian R1a
But we have seen its age drifting up slowly. Obviously it makes a big difference to interpretation as to when it actually appears.
And I have no issue with steppe admixture in SA, but their epic image in Fig 4A is not c/w an Andronovo model. This is basics.

a said...

This will be an excellent opportunity to parse the J2b samples with the R1b sample from Iran(is native or has steppe) since the R1b-M269-Z2103 and J2b's are buried in the same burial mound, only 6 samples from Hajji Firuz were plotted on the PCA and 6 samples in the admix comparison,
Hajji Firuz Iran I4243 I1b .. Hajji_Firuz_BA 2465-2286 calBCE (3875±25 BP, PSUAMS-2113) Iran
I2328 K1b1a CT Hajji_Firuz_C 6013-5898 calBCE (7075±30 BP, PSUAMS-2345) Iran
I4241 K1a3a J2b Hajji_Firuz_C 6016-5899 calBCE (7080±30 BP, PSUAMS-2163) Iran
I4349 U1a4 J2b Hajji_Firuz_C 5887-5724 calBCE (6915±40 BP, PSUAMS-2126) Iran
I2327 K1a17a R1b1a1a2a2 Hajji_Firuz_C 5900-5500 BCE Iran
I2323 K1a20 .. Hajji_Firuz_C 6060-5851 calBCE (7090±50 BP, Poz-81115) Iran
I4351 HV9 .. Hajji_Firuz_C 6056-5894 calBCE (7100±45 BP, PSUAMS-2151) Iran

Arza said...

It looks like they've finally split GAC into GAC_Poland and GAC_Ukraine (Globular_Amphora_Ukraine is used in the stats).

Chad said...

Hajji Firuz is listed as Z2103 in the Supplementary Table. Can't be correct as this is 1500-2000 years too early. None of these look as outliers though. They are all within 10-15% of the same amount of Anatolian and Iranian ancestry. They are going to go through the Y results carefully again.

Bronze said...

@Chad Rohlfsen

I did read it, the conclusion does not fit the data. And again nothing in this paper suggests an invasion. Even if this elevated ”steppe” ancestry is actually ultimately from the yamnaya in the pontic steppe, then the age period which they believe it arrived should leave clear traces of an invasion if that was what happened, obviously it was never any kind of invasion, and also quite possibly, steppe-related component could have existed in south-central asia long before any bronze age migration.

Shaikorth said...

The graph @ figure 3 evidently combines Lazaridis 2016 models for East Asians and ANE with later ones: first EA are modeled as 94% Onge-like and 6% slightly BMAC-like Western population (with Basal!) and then EHG is modeled as 94% "ANE/HG" with remaining 6% coming from the previous mix.
Would this work with MA-1/Afontova Gora/West-Siberan HG in place of EHG though?

Carlos Aramayo said...

@ Davidski,

The pre-print article should be properly refered to as Narasimhan et al. 2018, not Vagheesh et al. 2018 as you did.

Davidski said...

@Carlos

The pre-print article should be properly refered to as Narasimhan et al. 2018, not Vagheesh et al. 2018 as you did.

Well there you go, everyone makes mistakes. Fixed now.

Matt said...

Things I'll really need to dig to understand:

a) Why models of Steppe_MLBA+Indus periphery+Onge work at all when models of Steppe_MLBA+Iran_N+Onge previously did not work for Lazaridis, when the text emphasises that Indus periphery can be modelled by an Iran_N+Onge.

b) Is it actually worth them doing an extensive model using Steppe_MLBA+Indus periphery+Onge at all, when the Indus periphery is 3 samples who are highly variable in AASI ancestry?

It seems like if further samples for Indus proper ended up being in AASI (as seems reasonably possible) then there would end up be a need for some of the "far Turan" ancestry and not just Steppe_MLBA.

Sure there must be places in the text where these answers are quite clear. Maybe in the supplements.

Also what do y'all think about the absence of CHG and Levant_N from the modeling?

Together with the Fig 2a with lots of odd features, e.g. LBK_EN->Baden_LCA->Tripolye cline(?), a Steppe_EMBA cline from Yamnaya to Afanasievo(?), migration edge from BMAC->Yamnaya(?). Particularly what's everyone's opinion on that (BMAC->Yamnaya, wtf)?

Davidski: Holy crap! From the paper, page 16...

Yeah, Reich also talks about the Brahmin thing in his book.

The triangle Fig 3 in this paper kind of shows this, but doesn't really indicate which of the populations are which; who actually are the blue dots with a high Steppe pastoralist to Indus Periphery ratio but a high level of South Asian HG ancestry, blue squares, yellow squares?

So I took the admixture proportions from the hierarchical model in the supplement table 3 and made a few graphics with population labels: https://imgur.com/a/vba4x

Chamar_UP (North Indian) look unusually positioned in terms of combining levels of steppe ancestry with AASI, while the Panta_Kapu population looks outlying in terms of having relatively low levels of AASI for a population with their level of Indus_Peripheral.

capra internetensis said...

@Alaron

That's not E1b1b in Belt Cave but E1b1a, specifically an Atlantic West African subclade of E-M2. So actually this is nothing to do with Basal Eurasian, but instead Black Civilizations of the Ice Age. Or, my guess would be, wrong.

The men from the Neolithic Lake Urmia site with good coverage are J2b, the R1b-Z2013 sample has okay coverage (0.64) so might be right, but it is awfully old and a lot of these Y haplogroup calls are out to lunch.

Matt said...

From the full list of "Z-score for fit in the hierarchical model (positive implies more Steppe_MLBA-related ancestry than expected, negative more Indus_Peripheral-related)" where scores >=3 / <=-3:

Brahmin_Tiwari : 4.3, Brahmin_UP: 4.1, Chamar_UP: 4, Bhumihar_Bihar: 3.5, Adiyan: 3.5, Sikh_Jatt: 3.4, Palliyar: 3.2, Kolcha: 3.1, Brahmin_Nepal: 3

Gaud_Karnataka: -3, Reddy_Telangana: -3, Muthuraja: -3, Naidu :-3.2, Yadav_Pondicherry: -3.7, Nadar: -3.7, Coorghi: -4.1, Kallar: -4.3, Vysya: -4.4, Panta_Kapu: -5.8

AWood said...

R1b1a1a2a2 Hajji_Firuz_C 5900-5500 BCE Iran

This translates to Z2103. I find it hard to imagine that Z2103 can be part of an 8000 year old culture, I've never seen a calculation suggesting such a timeframe. I suspect it's a much younger sample.

EastPole said...

@Bronze
“There definitely was no aryan invasion into India, and EastPole you need mental help there was absolutely not any balto-slavic migration into South asia ever. Thats confirmed, R1a is not from Europe.”

They say that the only working models for Indians involve a combination of Indus_Periphery and a Steppe_MLBA group like Sintashta lub Srubna which as we know are derived from CWC which originated from Balto-Slavic steppe population according to Mittnik et al. 2018:

https://s9.postimg.org/sgzb3ky1r/screenshot_355.png

They also show a link between Central-Eastern Europe and India, look at
Tripolye/CWC/Sintashta/Andronovo and India trail:

I marked it in red here:

https://s9.postimg.org/t7iuhlba7/screenshot_353.png

and in green here:

https://s9.postimg.org/mttrejw67/screenshot_354.png

My opinion is that this confirms the link between Balto-Slavs and Indo-Iranians which is well known from linguistics and anthropology.

Notice the directions of the arrows.

vAsiSTha said...

89 male samples so far between 8000BC and 40BC from Iran, BMAC, Swat and only 1 has R1a that too 400BC. Hows this possible?

Davidski said...

@Kulkarni

Well, there will probably be more R1a samples in South Central Asia when the spreadsheet is fixed.

But don't expect any before Andronovo_SE (Steppe_MLBA_East) moves in.

Arza said...

@ Matt
BMAC->Yamnaya

It's Seh_Gabi_C. Two waves from the south, "Iran related" (CHG?) to Khvalynsk and this to Yamnaya.

There are other wtf's. E.g. total marginalization of CWC.

a said...

AWood said...
R1b1a1a2a2 Hajji_Firuz_C 5900-5500 BCE Iran

------This translates to Z2103. I find it hard to imagine that Z2103 can be part of an 8000 year old culture, I've never seen a calculation suggesting such a timeframe. I suspect it's a much younger sample.-----

R1b1a1a2a2 Hajji_Firuz_C 5900-5500 BCE Iran is about 1km from F38, Hasanlu IVb, 971-832 BC sample. That sample(F38) is way downstream from from L584* (Lezgin, Russia, Dagestanskiye_Ogni_in_Dagestan)



https://umap.openstreetmap.fr/en/map/ancient-human-dna_41837#11/37.0691/45.6695

https://mapper.acme.com/?ll=36.9944,45.4744&z=15&t=M&marker0=36.9944,45.4744,Hajji%20Firuz%20Tepe

Unknown said...

It seems like iranian agriculturalists had a major impact on south asia. Steppe people mixed with BMAC people who are of mostly iranian origin and this mixed group went on to mix with another hybrid group who were iranian agriculturalist/ASI mixed.

Arza said...

Fig 2. B

Khvalynsk is EHG + Iran related (CHG).
Yamnaya has additional shot of Iran_C (R1b guy).

Experiment with G25 conducted by me and Matt in two different ways showed that source of autosomal ancestry for both CWC and Beakers is CWC_early like (shifted away from Yamnaya towards European HG).

Yet, they are now completely twisting the story to make PIE out of Chalcolithic Iranians. Because of R1b. That's why they are hiding CHG behind the "Iran related" label.

Reich indeed is very diplomatic... now he doesn't want to upset Western Europeans. That's why they marginalized CWC here.

And as this paper is about Indo-Germanen they've put a straight line from Steppe_EMBA to Beakers to "German", bypassing any connection to CWC. To show who real Indo-Germanen are.

At this point, since the whole "diplomacy" was revealed, I don't trust that they didn't flush in the toilet any Mycenaean sample with the wrong kind of Y-haplogroup as medieval contamination (given the tensions between Greece and Macedonia).

But luckily they are not aware, that the more they will alienate CWC, the more obvious in the end will be, that Yamnaya together with R1b was not IE.

Chad said...

Hajji Firuz R1b calls:

0-I2327 R1b 0 1 0-I2327 R2a 1 0 0-I2327 E1b1a1a1c2b2a 2 0 0-I2327 E1b1a1a1c2c1a 4 0 0-I2327 E1b1a1a1c2c1b 9 0 0-I2327 E1b1a1a1c2c2a 6 0 0-I2327 E1b1a1a1c2c3c 12 0 0-I2327 C1b1a1a1a1a1a 1 0 0-I2327 H1a1d2b3a1 3 0 0-I2327 H1a1d2c1a1 2 0 0-I2327 H1a1d2c1b1 1 0 0-I2327 I1a3a1a1 1 0 0-I2327 I2a1a2a2 1 0 0-I2327 I2a2a1b1 1 0 0-I2327 N1c1a 1 0 0-I2327 O1a1a 3 0 0-I2327 O1b1a 1 0 0-I2327 O2a2b 2 0 0-I2327 M1b1b 1 0 0-I2327 Q1a1 1 0 0-I2327 Q1a2 4 0 0-I2327 R1b1 0 2 0-I2327 E1b1a1a1c2c3a1 2 0 0-I2327 E1b1a1a1c2c3a2 3 0 0-I2327 E1b1a1a1c2c3b1 6 0 0-I2327 H1a1d2b3a1a 2 0 0-I2327 I1a3a1a1c 1 0 0-I2327 I2a2a1b1a 1 0 0-I2327 I2a2a1b1b 1 0 0-I2327 T1a2b1a1 1 0 0-I2327 T1a3b2a1 1 0 0-I2327 O1a1a1 1 0 0-I2327 O1b1a1 7 0 0-I2327 O1b1a2 1 0 0-I2327 O1b2a1 1 0 0-I2327 O2a1c2 2 0 0-I2327 O2a2b2 2 0 0-I2327 Q1a1b 1 0 0-I2327 Q1a2a 1 0 0-I2327 Q1a2c 1 0 0-I2327 R1b1a 0 4 0-I2327 R1b1c 2 0 0-I2327 R2a3a 2 0 0-I2327 E1b1a1a1c2c3a2a 2 0 0-I2327 H1a1d2b3a1a1 1 0 0-I2327 I2a2a1a1a2 1 0 0-I2327 I2a2a1b2a3 1 0 0-I2327 N1c1a1a 1 0 0-I2327 O1a1a1a 1 0 0-I2327 O1b1a1a 1 0 0-I2327 O1b1a2a 2 0 0-I2327 O1b2a1a 1 0 0-I2327 O2a1c1a 1 0 0-I2327 O2a2a1a 3 0 0-I2327 O2a2b2a 2 0 0-I2327 Q1a2a1 1 0 0-I2327 Q1a2b1 2 0 0-I2327 R1b1a2 0 1 0-I2327 T1a1a1b2b2 1 0 0-I2327 N1c1a1a2 1 0 0-I2327 O1a1a1b2 1 0 0-I2327 O1b1a1a1 1 0 0-I2327 O1b2a1a1 2 0 0-I2327 O2a1c1a1 1 0 0-I2327 O2a1c1a5 1 0 0-I2327 O2a1c1b1 1 0 0-I2327 O2a2a1a2 1 0 0-I2327 O2a2b1a1 1 0 0-I2327 O2a2b1a2 1 0 0-I2327 O2a2b2a1 1 0 0-I2327 Q1a2a1b 3 0 0-I2327 R1a1a1b 1 0 0-I2327 R1b1a1a 0 5 0-I2327 R1b1a2b 1 0 0-I2327 R2a3a2b 1 0 0-I2327 T1a1a1a1b1a 1 0 0-I2327 N1c1a1a1a 1 0 0-I2327 O1b1a1a1b 2 0 0-I2327 O2a1c1b1a 1 0 0-I2327 O2a2a1a2a 1 0 0-I2327 O2a2b1a1b 1 0 0-I2327 O2a2b1a2a 2 0 0-I2327 O2a2b2a1a 1 0 0-I2327 Q1a2a1a2 1 0 0-I2327 Q1a2a1b2 1 0 0-I2327 Q1a2a1c1 2 0 0-I2327 R1b1a1a2 0 9 0-I2327 N1c1a1a1a1 1 0 0-I2327 N1c1a1a2a1 1 0 0-I2327 O1b1a1a1a1 1 0 0-I2327 O1b1a1a1a2 2 0 0-I2327 O1b1a1a1b2 1 0 0-I2327 O1b2a1a1a1 1 0 0-I2327 O2a1c1b1a1 1 0 0-I2327 O2a2a1a2a1 1 0 0-I2327 O2a2b1a1a6 1 0 0-I2327 O2a2b1a2a1 2 0 0-I2327 Q1a2a1a1c 1 0 0-I2327 R2a3a2b2b 1 0 0-I2327 T1a1a1b2b2b1a 1 0 0-I2327 O1b1a1a1a1a 1 0 0-I2327 O1b1a1a1b1b 1 0 0-I2327 O1b2a1a2a1a 1 0 0-I2327 O2a2a1a2a1a 1 0 0-I2327 O2a2b1a1a3a 1 0 0-I2327 O2a2b1a1a6b 1 0 0-I2327 O2a2b1a2a1d 1 0 0-I2327 R1a1a1b1a2 1 0 0-I2327 R1a1a1b1a3 1 0 0-I2327 R1b1a1a2a1 1 0 0-I2327 R1b1a1a2a2 0 3 0-I2327 R2a3a2b2b1 1 0 0-I2327 N1c1a1a1a1a1 1 0 0-I2327 N1c1a1a2a1a1 1 0 0-I2327 O1b1a1a1a1a1 1 0 0-I2327 O1b1a1a1b1a1 1 0 0-I2327 O2a2b1a1a6b1 1 0 0-I2327 O2a2b1a1a6b3 1 0 0-I2327 O2a2b1a2a1a1 2 0 0-I2327 R1a1a1b1a3a 1 0 0-I2327 R1a1a1b2a1b 1 0 0-I2327 R1b1a1a2a1a 3 0 0-I2327 R2a3a2b2a1a 1 0 0-I2327 R2a3a2b2b1a 1 0 0-I2327 R2a3a2b2c2b 1 0 0-I2327 O2a2b1a2a1a1a 1 0 0-I2327 R1b1a1a2a1a1 1 0 0-I2327 R1b1a1a2a1a2 1 0 0-I2327 T1a1a1b2b2b1a1a1 1 0 0-I2327 R1a1a1b2a2b1a 1 0 0-I2327 R1b1a1a2a1a1d 1 0 0-I2327 R1b1a1a2a1a1e 1 0 0-I2327 R1b1a1a2a1a2c 1 0 0-I2327 O2a2b1a2a1a3b2b 2 0 0-I2327 R1b1a1a2a1a1c3 1 0 0-I2327 R1b1a1a2a1a1g2 1 0 0-I2327 R1b1a1a2a1a2a5 1 0 0-I2327 R1b1a1a2a1a2b1 1 0 0-I2327 R1b1a1a2a2c1a1 1 0 0-I2327 T1a1a1b2b2b1a1a1b2 1 0 0-I2327 O2a2b1a2a1a3b2b1 4 0 0-I2327 O2a2b1a2a1a3b2b2 2 0 0-I2327 R1b1a1a2a1a2b1a 1 0 0-I2327 R1b1a1a2a1a2b1b 1 0 0-I2327 R1b1a1a2a1a2b3a 1 0 0-I2327 R1b1a1a2a1a2c1b 1 0 0-I2327 R1b1a1a2a1a2c1g 1 0 0-I2327 R1b1a1a2a1a2c1i 1 0 0-I2327 R1b1a1a2a1a2c1j 1 0 0-I2327 R1b1a1a2a1a2c1k 1 0 0-I2327 R1b1a2a1a1b1a1a 1 0 0-I2327 R1b1a1a2a1a1c2a1d 1 0 0-I2327 R1b1a1a2a1a1c2b1b 1 0 0-I2327 R1b1a1a2a1a1c2b2a 1 0 0-I2327 R1b1a1a2a1a1c2b2b 1 0 0-I2327 R1b1a1a2a1a2b1c1b 1 0 0-I2327 R1b1a1a2a1a2c1b2a 1 0 0-I2327 R1b1a1a2a1a2c1c1b 1 0 0-I2327 R1b1a1a2a1a2c1g3c 1 0 0-I2327 R1b1a1a2a1a1c2b1a1 1 0 0-I2327 R1b1a1a2a1a2a1a1a1 1 0 0-I2327 R1b1a1a2a1a2a1b1a1 1 0 0-I2327 R1b1a1a2a1a2b1c1b4 1 0 0-I2327 R1b1a1a2a1a2c1b1b3 1 0 0-I2327 R1b1a1a2a1a2c1e2b1 1 0 0-I2327 R1b1a1a2a1a2c1g1a1 1 0 0-I2327 R1b1a1a2a1a2c1g1b1 2 0 0-I2327 R1b1a1a2a1a2c1g2a1 1 0 0-I2327 R1b1a1a2a1a2c1k1a2 1 0

Chad said...

0-I2327 R1b1a1a2a1a1c2b2b1a 2 0 0-I2327 R1b1a1a2a1a2c1e2b3b 1 0 0-I2327 R1b1a1a2a1a2c1e2b4a 1 0 0-I2327 R1b1a1a2a1a2c1g1b1a 2 0 0-I2327 R1b1a1a2a1a2c1k1a1a 1 0 0-I2327 R1b1a1a2a1a2b1c1b3a1 1 0 0-I2327 R1b1a1a2a1a2c1e2b2a1 1 0 0-I2327 R1b1a1a2a1a2c1e2b3a1 1 0 0-I2327 R1b1a1a2a1a2c1g1b1a1 2 0 0-I2327 R1b1a1a2a1a1c2b2a1b1a 1 0 0-I2327 R1b1a1a2a1a2c1a1a1a1a 1 0 0-I2327 R1b1a1a2a1a2c1a1a1a1e 2 0 0-I2327 R1b1a1a2a1a1c2b2a1b1a1 1 0 0-I2327 R1b1a1a2a1a2c1a1a1a1a1 1 0 0-I2327 R1b1a1a2a1a1c2b2a1b1a1b1 1 0 0-I2327 R1b1a1a2a1a2c1c1b1a2a1b1 1 0 0-I2327 R1b1a1a2a1a1c2b2a1b1a4a1a 1 0 0-I2327 R1b1a1a2a1a1c2b2b1a1a1e3a1 1 0 0-I2327 R1b1a1a2a2 | R-CTS1078 R-CTS1078

Davidski said...

My eyes are hurting after staying up all night here.

What's the verdict?

a said...

In terms of viticulture there are only two contenders. Georgia and Iran.
Etymology
Latin vinum, "wine" or "(grape) vine", itself derived from the Proto-Indo-European stem *win-o- (cf. Armenian: գինի, gini; Ancient Greek: οἶνος oinos; Aeolic Greek: ϝοῖνος woinos; Hittite: wiyana; Lycian: oino)
. Armenian gini, Latin vinum, Ancient Greek οἶνος, Russian вино [vʲɪˈno]), Kartvelian (e.g. Georgian ღვინო [ɣvinɔ]), and Semitic (*wayn; Hebrew יין [jaiin]), Proto-Kartvelian *ɣwino-,[48] which is either a borrowing from Proto-Indo-European[48][49][50][51][52][53] or the lexeme was specifically borrowed from Proto-Armenian *ɣʷeinyo-, whence Armenian gini.[54][55][56][57][48] An alternate hypothesis by Fähnrich supposes *ɣwino- a native Kartvelian word derived from the verbal root *ɣun- ('to bend').[58] See *ɣwino- for more. All these theories place the origin of the word in the same geographical location, Trans-Caucasia, that has been established based on archeological and biomolecular studies as the origin of viticulture.

Georgia from 6000 BC,[21][22][23] Iran from 5000 BC,[6] and Sicily from 4000 BC.[8]

https://www.penn.museum/blog/collection/125th-anniversary-object-of-the-day/7000-year-old-wine-jar-object-of-the-day-24/

a said...

R-Z2103Z2103/CTS1078 * Y4371/Z8128/M12149 * S20902/Z8130+7 SNPsformed 6100 ybp, TMRCA 5700 ybpinfo

R1b-Z2103---CTS1078
Hajji Firuz R1b will be linked with
Yamnaya and Northern Caucasus, and probably Maykop.
I0126, Kutuluk III, Kutuluk River, Samara (Russia), 2867-2484 calBCE
(AA53803)
RISE548, Temrta IV (Russia), Yamnaya_Kalmykia
RISE555, Stalingrad Quarry (Russia), 2857-2497 calBCE (AAR-20358)
RISE547, Temrta IV (Russia), 2887-2634 calBCE (GrA-58960)

capra internetensis said...

@Chad

If I'm reading that right there are 9 + calls for M269 and 3 + for Z2103, and it doesn't look like there are a lot of bad calls. There *was* a Bronze Age intrusive burial in the same site, but it was a genetic outlier. Could be legit.

Davidski said...

I'll have a lot more to say about this after I run these samples myself.

Arza said...

@ a
-wiń, -winąć - to wrap, to wind
Basically any kind of creeper can be called like that, so...
"which is either a borrowing from Proto-Indo-European".

Chad said...

Could be. Only 5-10% more off than yfull is thought to be. I've asked if they can get a C14 on it to be sure.

Samuel Andrews said...

@Bronze,
"and also quite possibly, steppe-related component could have existed in south-central asia long before any bronze age migration."

This study didn't do a good job sampling south Asia. It focused more on central Asia & iran. It includes no samples from South Asia predating the Bronze age. I'm disappointed by that because it leaves room for people to say "Well, maybe the Steppe ancestry in South Asia is older than the Bronze age."

It's literally impossible for Steppe ancestry to have been present in South Asia before the Bronze age. In the chalcolithic, bronze age Steppe ancestry was unique to eastern Europe. It can't be confused with ANE or EHG ancestry.

This study shows Steppe ancestry doesn't appear until the 2nd millenium BC when Andronovo rich in R1a Z93 arrived.

We have confirmation of an "invasion" of Steppe R1a Z93 people into Central Asia in the 2nd millenium BC. The existence of Steppe ancestry in Pakistan in the 2nd millenium BC obvisouly came with the same people (Andrnovo).

Andornovo cultural group many archeaologist consider to be proto-Indo iranians and who from ancient DNA we know carried essentially 100% R1a Z93.

Considering all of that, I don't see how anyone would argue against AIT. If not AIT, then at least AMT (Aryan migration Theory).

Rob said...

OK hopefully we can get raw data to quantify and qualify everything , and see if steppe = late PIE survives

Anonymous said...

This paper is very badly done, and it's a giant collaboration. At least it's the preprint, so it's acceptable.
But we have EN samples labelled as Chalcolithic and vice-versa, this is ridiculous. Where are the Neolithic and Bronze Age IVC and ASI as well? I hope they appear later.
Also, so many yDNA errors.
It looks very amateurish, for sure.

Now, a good surprise was finding the West_Siberia_HG, an EHG with more ANE shift. This can be, in fact, a proxy for the population who created the CHG, as there were instances of contact between the regions. Also, would it have anything with the Uralic question?

Another solved mystery was the Steppe_EMBA vs Steppe_MLBA as the population who went down under. Steppe_EMBA was the favourite candidate by far, but it turns out Steppe_MLBA dictated the expansions (not a surprise for some, considering R1a). And that, in fact then, we can classify these expansions as Europeans moving, instead of being shy about that.

I wish I could say more, but many of the samples are from the Iron Age, and therefore quite a disappointment, because they only tell the end results, not the process.
But at least it looks like that the far east Iranics of Turan looked like much more like present day Iranians than with the current peoples of the place. Same thing for the Indians, which looked like much more like Balochis for instance than with current Indians.

Or am I misinterpreting something?

Samuel Andrews said...

Someone needs to tell Reich and his team that based on Y DNA it is very unlikely Yamnaya represents PIE or LPIE.

Samuel Andrews said...

Chad, Rob, others.

Remeber, based on mtDNA, I argued the Steppe ancestry in Southcentral Asians had EEF admixture like Andronovo did.

Anonymous said...

@Samuel Andrews
Nonsense. The picture forming is like that:
1. Samara Eneolithic being EHG + CHG
2. Ukraine Eneolithic being Samara Eneolithic + GAC + Hunter Gatherer
3. Yamnaya being Samara Eneolithic + CHG
4. Corded Ware being Ukraine Eneolithic + Yamnaya + Hunter Gatherer
5. Bell Beaker being Yamnaya + GAC(and GAC-like) + Hunter Gatherer

All R1a and R1b. There's absolutely no incongruence here.

Arza said...

if steppe = late PIE survives

It won't survive, because there was no other "early PIE".

Chad said...

The Siberian Hunter didn't make CHG. It is something different. The hunter side of CHG is actually more WHG shifted than the hunter in Iran_N. CHG can actually be modeled as Iran_N, Boncuklu, and EHG. Which makes much more sense.

Jijnasu said...

So Basically south asian prehistory -
1. AASI like people inhibit mesolithic SA
2. In the 4th millenium people with iran neolithic like ancestry settled in the NW of the subcontinent mix with AASI to form ASI
2. Sometime in the late 3rd millenium rice farmers from SE asia migrate into eastern India mixing with AASI HGs and later with ASI
3.Sometime between 2000 & 1500 BCE Indo-Aryans move southwards from Kazakhstan into the Indian subcontinent intermixing with local populations

Anonymous said...

@Chad Rohlfsen
I see then, what a shame, because current samples still don't fit CHG well.

Samuel Andrews said...

The dozens of new Sintashta, Andronovo samples from central Asia belong mostly to the same handful of mtDNA lineages already published data does. Makes me think they come from a small founding female population.

The data from Central Asia & Iran, confirms what had looked like Iran Neo-related mtDNA lineages for a long time based on modern mtDNA....

J1d, J1b3, W3, W6, I1, HV13

Samuel Andrews said...

@namedguest,

Corded Ware is of 70% of Yamnaya-like decent. If they have any excess UkraineHG or WHG (beyond what is in MN farmers), it is small.

Anonymous said...

@Samuel Andrews
"Corded Ware is of 70% of Yamnaya-like decent."
And 30% not Yamnaya.

"If they have any excess UkraineHG"
Please, see this here: http://eurogenes.blogspot.com/2018/03/was-ukraineeneolithic-i6561-proto-indo.html

"or WHG"
Not WHG, probably Narva.

"it is small."
30% isn't.

Vara said...

@Nirjhar

"Second, samples from three sites from the southern and eastern end of the Steppe dated to 1600-1500 BCE (Dashti-kozy, Taldysay and Kyzlbulak) show evidence of significant admixture from Iranian agriculturalist-related populations, demonstrating northward gene flow from Turan into the Steppe at the same time as there was southward movement of Steppe_MLBA ancestry through Turan and into South Asia."

Probably this is south to north movement is the start of the formation of East Iranians which makes sense with since Indo-Iranians were familiar with camels while Andronovo only got their camels around 1500 BCE with this movement.

"However, at least some (possibly all) of the Steppe pastoralist ancestry in South Asia owes its origins to southward pulses in the 2nd millennium BCE, as indeed we prove directly through our observation of this ancestry in the Swat Iron Age individuals dating to ~1000 BCE"

Indo-Aryans made it to the Indus in the BA, 1000BCE is too late. Though to be fair they do not have Swat BA samples.

PS. FIG. 2C shows no steppe ancestry in South Asia?

PPS. I hope this doesn't bring up R1a vs R1b and who is more Indo-European crap back.

Tone said...

I take a slight issue with the Fig. 4 map inferring that Andronovo were "Yamnaya pastoralists" spreading east out of the Volga region after 3,000 b.c.

The Andronova people were later in time, and significantly different genetically from Yamnaya. Of course, Andronovo people were clearly related to Yamnaya (50% to 75%) but they really are a product of a mixing of Yamnaya and a "Globular Amphora type" people; a mixing that occurred in Central Europe. Their starting point for Andronovo was not really the Volga steppe as part of the Yamnaya horizon, but actually Western Ukraine and Poland. I think the Fig 4 map should show this.

Off topic but I'm starting to wonder if the proto-Indo-European language was a hybrid, and not born until the Yamnaya mixed with farmer folk they encountered in Central Europe. I wonder if these hybrid peoples: --- Andronovo, Corded Ware, Bell Beaker -- all with roughly the same amount of EHG/CHG/EEF/WHG --- if these peoples were the first Indo-Europeans, and they exploded out of Central Europe after 2800 bc. and not the steppe per se, although they did move into the steppe, replacing their Yamnya ancestors.

Disclaimer: I'm an amateur and I don't know crap. But it's fun to speculate.

Arza said...

@ Jijnasu

Power of Eurogenes spreadsheets unleashed.

https://s6.postimg.org/5aylcxt29/so_basically_south_asian_prehistory.png

Archaelog said...

So there is Z2103 in Sintashta present with the R1a samples. That suggests the existence of a Poltavka-Corded Ware hybrid culture in which the language (Pre Indo-Iranian) could have come from either of the two groups.

My bet is that Poltavka spoke Pre-Indo Iranian and they introduced the language to the Corded Ware folk who shared their territory (like the model I linked to before). Fits in neatly with the proposed Greco-Aryan theory (if Proto Greeks are assumed to have been Z2103)

Jijnasu said...

@Arza
Nice work

AWood said...

@Vara

There is already R1b in both Celtic and Germanic warrior graves so that debate is already settled. All that can really be decided is if the ancestor language was inherited from Middle Eastern people (the J2, J1, G, E1b crew), through exogamy, or the local development of the language spoken by their ancestors.

Rob said...

@ AWood
I agree
By 100 BC, most L51 had been indo europeanised

Samuel Andrews said...

This paper mentions a "Anatolia_Mesolithic." Looks like it is similar to Anatolia Neolithic.....

It is mentioned in Fig. S3.4

In that figure are the most negative F3-stats for Chalcolithic samples from Seh_Gabi Iran. ANatolia_Mesolithic is near the top for most negative alongside with Romania_EN & Balkans_N.

Samuel Andrews said...

The claim Reich & Razib have been making a lot that within Mesolithic west Eurasia were populations as different from each other as modern Europeans & east Asians seems like an exaggeration to me.

EHG was very similar to WHG. It was basically of 50% WHG origin. We don't know the deep ancestry of Mesolithic-early Neolithic Middle Easterners yet. But, we can be confident they all had significant common ancestry in the last 20,000 or 30,000 years unlike modern Europeans & east Asians.

Having no common ancestry in the last 40,000 years & being as different from each other as modern Europeans & east Asians are are two different things.

Someone of 50% European & 50% African ancestry will technically be more different from Europeans than East Asians are. But that is a totally different situation than being decended from an isolated Eurasian group like Papuans are. In the type of analysis Reich is basing the claim on, the 50% very divergent African ancestry would mask the recent common ancestry that person would share with Europeans.

Anatolia_Neolithic, Iran_Neolithic, and Natufians may seem really different from each other because each have ancestry from different really divergent groups.

Davidski said...

Those three Indus_diaspora samples are awesome.

I don't doubt for a second that they really are migrants from the Indus Valley Civilization. I'm sure that at least some of the authors have a very good idea what IVC samples from India and/or Pakistan look like by now, so this classification is unlikely to be based on someone's wild imagination.

Also, they really do stick out, by totally lacking the Anatolian farmer signal and clearly showing indigenous South Asian Onge-related ancestry. There must have been quite a bit of South Asian ancestry in Harappans, although the levels no doubt varied, like among this little diaspora group.

Well, there goes the idea that Harappans were unadmixed ANI, in other words West Eurasians lacking South Asian ancestry.

Arza said...

@ Rob
But not without a hassle.

Due to sound-shifts case system collapsed:
http://i.imgur.com/2MsxnQu.png

And some non-IE features were acquired, like the use of articles:
https://en.wikipedia.org/wiki/Article_(grammar)#/media/File:EuropeArticleLanguages.png

Or vigesimal counting:
http://www.academia.edu/2353028/Old_Danish_vigesimal_counting_A_comparison_with_Basque
the best known western case being perhaps that of Old French. In the Celtic languages, vigesimal counting is found in the Gaelic – Modern Irish, Scottish Gaelic, and Manx – as well as the Brit-tonic branch – Welsh, Cornish, and Breton – while literary Old Irish is decimal and the situation in Gaulish virtually unknown.

Chad said...

That's Boncuklu. For some reason, they have it labeled Mesolithic.

a said...

So with this latest batch of samples; still needing some tweaking, we have
R1b-Z2103 samples from:
1)Yamnaya culture
2)Afansievo culture
3)Poltavka culture
4)Potapovka was contemporary with the Sintashta culture
5)Sintashta culture
6)Vucedol culture
7)Bell Beaker culture

Archaelog said...

@a And L51 from the Bell Beaker territories. L51 was Z2103's sister clade, from the steppe obviously even though it hasn't been detected there yet. Where else could it be from?

Seinundzeit said...

Okay, so I just skimmed the paper (have yet to tackle the supplementary material). Will read in more detail tomorrow, but preliminary thoughts...

1. Lots of mistakes throughout the preprint, but I'm sure they'll fix this (if I have time tomorrow, I'll leave a comment on all those errors at biorxiv).

2. And yes, we were right about ANE/Srubnaya_outlier! (Thanks for the shoutout, Shaikorth)

I agree with Alberto; calling this sort of ancestry "West_Siberian_HG" does seem debatable, as it's everywhere in South Central Asia and apparently North India (see below). I guess Chad and Rob were right about the meaning of those Srubnaya_outlier percentages we see from Tajikistan to Tamil Nadu (it's a "native" component, representing heritage from Central Asian foragers).

3. Also, if I'm reading this correctly (and as Alberto also noted), the Indus_diaspora samples are three-way mixtures between Iran_N-related, West_Siberian_HG-related, and Onge-related ancestries.

So, this explains why Steppe_EMBA + Iran_N was the only setup which worked. Basically, the Steppe_EMBA signal was an "illusion", due to actual Steppe_MLBA, in conjunction with some West_Siberian_HG which was already combined with Iran_N and Iran_Chl.

4. Matt, it really would make more sense for them to model all these populations as Iran_N, Iran_Chl, West_Siberian_HG, and Steppe_MLBA. In fact, their setup strikes me as being somewhat convoluted.

Also, on the AASI question, I would note that the Indus_diaspora percentages eat up Iran_N, Iran_Chl, West_Siberian_HG, and AASI. There is probably great variability in the streams which these mixed samples are capturing (for different populations).

In other words, Indus_diaspora may only contain as little as 10% AASI, or it may contain as much as 50% AASI, but the percentages will stay pretty close to what they are for these Soyth Central Asians, as "Indus_diaspora" is the only place for their Iran_N, Iran_Chl, and West_Siberian_HG ancestries to go.

And considering the eastern shift of West_Siberian_HG, an excess of West_Siberian_HG for Kalash or Pashtuns will lead to an increase in Onge (in comparison to Indus_diaspora), since that's the only eastern reference. So, I'm still betting on 10%-15% AASI.

If 20% AASI for the Kalasha and the HGDP Pashtuns, then northwestern Iranians have to be at 5% AASI, and Pamiri highlanders have to be at 10% AASI, which strikes me as being implausible.

Of course, just because it is implausible, doesn't mean it's impossible! I have an open mind on this. Anyway, we'll find out soon enough, once David adds these samples to Global_25.

5. Speaking of which, David, I can't wait to see these samples in Global_25! Exciting times...

Nirjhar007 said...

Vara,

It is shame they didn't include Mature Harappan samples, their main argument is based on those outliers, but stay ready for surprises when they come :)...

samples from three sites from the southern and eastern end of the Steppe dated to 1600-1500 BCE (Dashti-kozy, Taldysay and Kyzlbulak) show evidence of significant admixture from Iranian agriculturalist-related populations, demonstrating northward gene flow from Turan into the Steppe
Interestingly from this period we see Fedorovo culture with cremation ...

Seinundzeit said...

Also, just something to throw out there...

In South Central Asia today, the Burusho probably have the highest levels of West_Siberian_HG (as inferred from Srubnaya_outlier percentages), so I wonder if Burushaski has a relationship with the languages spoken by the Kelteminar people...

(just spit-balling)

Alberto said...

@Matt

a) Why models of Steppe_MLBA+Indus periphery+Onge work at all when models of Steppe_MLBA+Iran_N+Onge previously did not work for Lazaridis, when the text emphasises that Indus periphery can be modelled by an Iran_N+Onge.

That's because Indus_periphery is Iran_N+Onge+West_Siberian_HG.

b) Is it actually worth them doing an extensive model using Steppe_MLBA+Indus periphery+Onge at all, when the Indus periphery is 3 samples who are highly variable in AASI ancestry?

This is a good question, with 2 different aspects. The first one is the very deceiving wording of the text. They say that the only models that work for SPGT (Early Swat Valley samples) involve Onge+Indus_periphery+Steppe_MLBA. But what they really mean when you look at he data is that the only steppe group that fits into those SPGT samples is Steppe_MLBA, and the older steppe populations are rejected. Not quite the same thing.

They didn't care to test any model for SPGT as Onge+Indus_periphery+Other SC Asian samples. So we don't know if that works, and if it does so better or worse than Steppe_MLBA.

The second thing is indeed about basing the models in those 3 variable samples. I don't doubt that they are almost surely from the IVC, but it's more doubtful that they represent the whole population or even the average. What is more worrisome, though, is that the models can't tell if or how much Steppe_MLBA is in SPGT except by using the average of those 3 samples as a baseline. As an example, if instead of the average of the 3 they used the sample that is closer to the SPGT samples (Gonur2_BA), the Steppe_MLBA admixture would go down from 22% to maybe half. Which doesn't inspire much confidence in the power to detect Steppe_MLBA admixture.

Rather contradictory is the fact that SPGT's West Eurasian side is much less steppe shifted than the modern Indian cline (Fig. S4.2).

So when you try to combine the archaeology of those sites (we discussed the paper about the Udegram burial site here already) and check the Y-DNA to try to verify or falsify the steppe admixture, you end up with a bit of a WTF feeling.

Don't want to be too hard with the main text (this is a preprint, and would never be published in its current state, after all), but their venturing into linguistic territory makes it even that bit more... embarrassing?

Anthro Survey said...

@Sein

Well, I guess now we know what's up. Central_Asia was, in a sense, analogous to Neolithic Western Europe as expected.
Like I'd also speculated, there was admixture with these ANE-like folks along the way, too. Just LOOK at those Sintashta samples! Wow.

Anthro Survey said...

@Matt

"BMAC has EEF, not Levant_N? That will need some retesting..."

That makes perfect sense, though. We wouldn't expect a Circum-Arabian Natufian/Harifian component far north of the southern Levant before the Akkadian era.

The North Mesopotamian culture accounting for the introduction of wheat cultivation, rectangular mudbrick architecture, lighter skin/Gracile Mediterranean traits and the signature West-Ward shift to Greater Iran was probably a mix of ANF-like and CHG-like, leaning towards the former. Maybe like Tepecik.

Seinundzeit said...

Anthro Survey,

Absolutely; those samples are incredible.

Lots of gene-flow, in all directions.

I'm really looking forward to working with these samples, and I am pretty excited to see what David comes up with.

Personally, I think we'll all get a better idea with Global_25. That data is quite sensitive to recent drift and has tended to guide us on the right path (I mean, Steppe_MLBA and Iran_N + Srubnaya_outlier was the truth of the situation, as was a lack of Iran_Chl in India, and many other things).

Anthro Survey said...

@Rob

Not so fast. ;-) We both expected this native "pseudo-steppe" forager component to be an inflator of steppe ancestry in C, SC and S Asians.

Now, assuming there is no mistake, thoughts on the R-m269 in that LN/Chalcolithic Hajji Firuz sample?

Maybe this m269 was originally connected to CHG(note: I do emphasize CHG, not Iran_N), after all. MAYBE. Basically, I'm implying here that Hajji Firuz is Iran_N+North Mesopotamia. N.Meso was potentially ANF+CHG(source of R1b?).

Anonymous said...

Jezus Christ, that Y-DNA list is a mess. This is becoming a bit of a nuisance, Mathieson called one of the Tiefbrunn CW R1b while all others called R1a. This list has flaws too.

I mean, this is supposed to be Top of the Bill laboratory work.

postneo said...

The paper has a very large number of ancient samples from Iran and Central Asia and a few from swat. Yet it does not model ancient vs modern populations from Iran or discuss anything about the origin of Iranian languages. Is Iranian not IE? It’s more obsessed by South Asia despite having very few samples from There

Anonymous said...

Unless off course the R1b-M269 and especially the R1b1a1a2 in MA1 are an April Fools joke. In that case it's a brillant joke

postneo said...

I raised this earlier, but in fig 1d the yellow onge show up in Siberia neolithic and also many outlier individuals in steppe mlba. Indus periphery formation date is earlier than steppe mlba. Perhaps thats the reason for the fits of MLBA for South Asia.

Clearly theres much in the figures that differs from the text.

EastPole said...

@Chetan
„So there is Z2103 in Sintashta present with the R1a samples. That suggests the existence of a Poltavka-Corded Ware hybrid culture in which the language (Pre Indo-Iranian) could have come from either of the two groups.

My bet is that Poltavka spoke Pre-Indo Iranian and they introduced the language to the Corded Ware folk who shared their territory (like the model I linked to before). Fits in neatly with the proposed Greco-Aryan theory (if Proto Greeks are assumed to have been Z2103)”

No you are wrong. We have a following set of equations which need solving:

CWC + steppe = Sintashta

Sintashta + Poltavka = Andronovo_East (Pre-Indo-Iranians?)

What was the language of CWC, Sintashta and Poltavka?

In my opinion:
CWC is linked with Balto-Slavic, Sintashta probably too. What was the language of Poltavka?

old europe said...

Happy easter to everybody!

@all and Tone

Since I'm not really an expert in genetics I want to ask if the role of EEF in the steppe ethnogenesis Andronovo Sintashta stuff, according to these results, is confirmed and if so it is more or less strong than previously expected?

Rob said...

@ Anthro

"@Rob

Not so fast. ;-) "

What's not so fast ?

"Now, assuming there is no mistake, thoughts on the R-m269 in that LN/Chalcolithic Hajji Firuz sample?"

The origin of shaft-holed axes could trace to northern Iran, which then spread to Majkop. After a pause, it became part of the Yamnaya & Vucedol status set (Ivanova) due to expansion of groups of men (these Z2103, perhaps along with J2b2).

Curiously, CWC R1a-M417 doesn't seem to be part of that link. Instead, they had their own techniques based on old Carpato-Balkan borrowings.


But, if correct, it means we have been under-dating haplogroups. L23 is currently dated to 4400 BC, several hundred years later than this derived group.
So what does this mean for Z93 ? Possibly the same - a late Neolithic/ Chalc expansion marker and not solely due to Andronovo.

Bronze said...

@Samuel Andrews. Wrong you dont know that steppe ancestry originally came from eastern europe.

Andronovo is way to late for proto indo-european and no signs of any invasion at all, no aryan invasion of central asia, no invasion of south asia. And also as has been said previously, Andronovo is a bad autosomal fit for the ancestry of modern south asians.

old europe said...

Do not want to troll anybody but in my humble opinion all these arrows, genes are not yet quite understandable. As for relation between genetics and language just think about that:

1) Romans changed the language of western europe with a genetic contribution equal to zero
2) Greek terms in western language percolated with a genetic contribution equal to zero
3) I could be wrong but the arabs changed the language of large parts of africa and asia with a genetic contribution that is equal to zero
4) In south america spanish and portoguese languages imposed themselves with a genetic contribution that is equal to zero
5) Norman invaders changed the old english language in nearly 2/3 of the vocabulary with a genetic contribution that was less than 1 per cent.
6) Sub saharian africa speaks european languages with a genetic contribution that is equal to zero

All these examples should make us more humble when talking about genes and language shift.

To me the IE can be solved with this procedure:

Examine all the indoeuropean cultures
1)Detect which cultural traits are common to the majority of them ( all of you already know what I mean especially my obsession I admit for the cremation rite)
2)Detect where this cultural traits popped up first
3) Than examine if genetics can support the most likely place of origin.



Aram said...

A0-T,DE and E1a in Pakistan. OMG!!
Humanity is from South Asia?
A and BT from Gonur tepe Turkmenistan.
Either this is a revolution either it is a complete mess.

vAsiSTha said...

any thoughts about H2 found in 6200BC anatolia(Turkey); H3 found in 3500BC Iran and its relation between H2, H3 in India?

Matt said...

Arza:There are other wtf's. E.g. total marginalization of CWC.

Yeah, I was thinking that was marginal because of the relatively small offsets CWC_Early_Baltic and Yamnaya in the big picture, but I can see that you could argue that the way the early CWC samples vary is actually important in terms of understanding the Steppe_MLBA and so matter to all their models.

The lack of use of CHG in the modelling (I think they may it in some outgroups at some point) is the most eyebrow raising bit for me though.

From what we've seen in the previous lit and Davidski's trees, CHG seems like a perfect "tracer dye" to distinguish steppe ancestry from these other sets between combinations of EHG, West Siberian, Anatolian, and Iran_N related.

CHG related dimensions/drift peaking in the Caucasus emerge readily and quickly in every form of analysis, and steppe EMBA have positions/values, etc. on which Steppe EMBA and certain modern people always have a unique position.

Anthro Survey: That makes perfect sense, though. We wouldn't expect a Circum-Arabian Natufian/Harifian component far north of the southern Levant before the Akkadian era.

Good argument, though I'll note in the paper their reasons for selecting Anatolians seem down to their concerns about sample size, coverage, etc:

p106 "An important caveat is that we do not consider here Levantine agriculturalists who were closely related to those of Anatolia with some uncertainty as the direction of gene flow between Anatolia and the Levant [Laz16].

Relatedly, the distribution of Anatolian/Levantine/Iranian-Neolithic related populations in the ancient Near East is only sparsely known[Cite Laz/Broushaki/Boncuklu paper], with an important lacuna in Mesopotamia. Our results do not imply that the shift related to populations sampled in northwestern Anatolia [Mathieson] implies admixture from that area, but we use this set because of its large sample size and high quality as representatives of Neolithic populations of the western Near East."


(I vaguely remember, at some point when we were discussing Iran_Chl and whether it was CHG like, drawing clines Iran_Chl dots intermediate Iran_N and Anatolia_N on PCA though, rather than as CHG related, so it kind of makes me smile).

I think there's some kind of thing going on in the paper as well where they seem to be trying to equate Anatolian ancestry with the development of crop agriculture and Iranian ancestry with early pastoralism. That kind of cleavage would seem pretty confused to me though.

Aram said...

A question to experts.
Is it possible that we don't see R1a in SA because of cremation practice? How widespread was cremation Swat valley?

old europe said...

On the opposite side we have

immigrants from non english speaking countries make up 80 per cent of us population. They did not alter a single bit the language landscape of north america.

In the high middle ages hundred of thousand of Goth, Visigoth, Ostrogoth, Avars, Unns, and every kind of germanic tribes entered western europe without ( except for england) altering a single bit the language landscape ( in the sense that western europe remained romance speaking)

In the last decades millions and millions of people entered europe from asia and africa without altering a single bit the language landscape of the continent.


When we will recognize that language has more to do with culture and religion than with genes it will be always too late.

old europe said...

In this blog and all over the world we speak and communicate with english.

How much english ancestry do we have in the world?

Aram
Good question! cremation rite can alter the data.

Kristiina said...


@ Aram "Humanity is from South Asia?"

I tried to google for Out of India maps and I found this:
Figure 1, http://rstb.royalsocietypublishing.org/content/367/1590/770

You just add an arrow going from Pakistan to North Africa and redirect the arrow going from Arabia to Africa to go from Pakistan to Africa and voilà.

However, this may not be the ultimate origin of humans, but am I right we were recently told about very early dates for human presence in India.

Matt said...

Alberto:
That's because Indus_periphery is Iran_N+Onge+West_Siberian_HG.

This seems quite likely to me given everything, though for whatever reason not in their conclusion (maybe there's some contraindication there I haven't stumbled on).

(Sidenote: If you had even 10% West_Siberian_HG in Indus_periphery, and it could be more or less, and Steppe_MLBA is about 37.5% EHG, then a pop like the Kalash modelled with 68% I_P and 28% Steppe_MLBA would by implication have 6.8% West_Siberian_HG and 10.5% EHG. Pop like Brahmin_Vaidik via their modeling would have 5.8% EHG and 5.8% West Siberian_HG.).

The second thing is indeed about basing the models in those 3 variable samples. I don't doubt that they are almost surely from the IVC, but it's more doubtful that they represent the whole population or even the average.

Conclusion on page 173 supplement: The fact, that the ratio of Iranian farmer-related to South Asian HG related ancestry in the Indus Periphery, is the perfect fit for the ancestry in the SPGT along with additional steppe pastoralist ancestry, suggests that the Indus Periphery samples, which are anomalous in their ancestry type when compared with all the other samples we have from Iran and Turan in them harboring significantly higher proportion of South Asian HG related ancestry, suggests that these indiviudals are likely to be migrants from South Asia.

I mean, yeah, not that confidence inspiring to me at least that three samples with variable ancestry happen to be perfect in their average; I guess you can make the statistical argument that averages tend to be representative, but with three samples!

Don't want to be too hard with the main text (this is a preprint, and would never be published in its current state, after all)

Do think the language may need some cleaning and clarifying up at least; it's a huge paper and perhaps it would have been better for them to try and eat the (Indian) elephant one bite at a time. Several papers over the last year could've been a better outcome (no doubt not possible for one reason or another). Maybe kind of just starting with the basics with understanding levels of basal Eurasian in each group, more analysis breaking down early populations into ghosts, then moving to later ones.

There's a bit of something going on here with timing to try and fit with Reich's opus and probably multiple revisions to deal with the fraught collaboration with Indian researchers.

@Sein:Personally, I think we'll all get a better idea with Global_25

Yeah, this could be true. I do really want to understand what they've done here though, because it's clearly a lot of work that's gone into this any to developing methods to deal with the complexities of possible paths.

Also, on the AASI question, I would note that the Indus_diaspora percentages eat up Iran_N, Iran_Chl, West_Siberian_HG, and AASI. There is probably great variability in the streams which these mixed samples are capturing (for different populations).

I'd suspect this is not an issue for Kalash and populations with a similar ancestral balance net drift because the hierarchical model fit seems close to best for these samples (Z:-0.3; Pathan is Z:0.1), which seems to imply that the levels of AASI vs "Other Indus periphery ancestry" are close to perfect as an ancestor for them (for whatever I_P actually has) with a little extra AASI added on, but I'd definitely agree I'd welcome a more transparent form of analysis if it's possible to really clear it up.

vAsiSTha said...

R1b1a1a2a1a from Turan/current afghanistan 2650BC.

Archaelog said...

@EastPole Well since Polatavka shows continuity with Yamnaya in the Volga-Ural region, archaeologically and genetically, I daresay they spoke a language directly descended from late PIE! My opinion is that it is this language which became Proto Indo-Iranian by the time of Sintashta.

The case of CW (Abashevo) is much less clear. CWCs were spread over a large region and it is not at all necessary that they spoke the same language everywhere. They could have spoken an IE dialect, but first we would need to know the language of the Ukraine region in the Eneolithic since that is where CW appears to originate from.

At this stage, I'm open to all possibilities.

Rob said...

MATT / Alberto
Have you noted which andronovo they have used in their analysis ?(a subset or all grouped)

Anonymous said...

Someone on Anthrogenica already noted that Reich uses a proxy for IVC, and that may point to a deteriorating relation between his team and the Indian team. Possibly that could mean that they rushed to publish this preprint.

Matt said...

@All,
If anyone is interested, I've used the values from the hierarchical model in the supplement table 3, with the scaled G25 data to generate a couple of simulated samples for Indus_Periphery and AASI -

Sims here: https://pastebin.com/1FJTYFZ2 (slightly earlier version of the post put them in the comment, screwing up the page formats ;) ).

Few graphics including them https://imgur.com/a/4pUpI

Caveat, note that particularly for the Indus_Periphery sim, there's probably some uncertainty, and this is relatively quick and dirty. For the regression, there were only 11 populations from the Indian cline overlapping between G25 and the supplement which plus a Steppe_MLBA average.

I think they're worth a try if anyone wants to have a play around though (note the Indus_ Diaspora-related_Sim seems to peak dimension 3 in a very unique position opposed to WHG and Siberians so I doubt it will be modeled by anything else too well, but whether this is an artefact of the method or real I can't tell).

Matt said...

Re: my Indus_Periphery sim, on reflection the position looks a bit more like AASI plus something like Iran_N exaggerated against WHG and to a lesser degree Levant_N. (That is, exaggerated on an WHG->Iran_N cline and to a lesser degree Levant_N->Iran_N cline. Expected?)

vAsiSTha said...

Swat is derived from vedic river Suvastu, and the name is carried by indian community of Kayastha through last name Srivastav.

Davidski said...

@All

I updated the post with a couple of key quotes from the paper and supp info.

Let me just reiterate, and this time with the correct terminology: those three Indus_Periphery samples are awesome!

Matt said...

Combining those G25 sims I created upthread (from the hierarchical model proportions with regression) to make an their ANI and ASI they state:

https://pastebin.com/1N4t6mNd

(ANI= 28 Steppe_MLBA:72 Indus_Periphery; ASI=61 AASI:39 Indus Periphery)

Does seem like with that ANI sim, Balochi/Makrani become likely to be basically Iran Iron Age+ANI, while Kalash/Pathans become ANI+Tajik, with both taking fairly substantial non-ANI ancestry... (Present day Iranian groups probably would take some level of ANI+Iran IA as well).

With these I_P and AASI sims however, I think an ANI Sim that is slightly more Steppe_MLBA rich at 37:63 may make more sense to fit clines:
https://pastebin.com/MTdyUd5q

Anonymous said...

Proszę przeczytać „R1b-M343 (xP312 xU106) Y-STR Report”
Pana Dirk Struve Dec 4, 2017.
Można tam znaleźć dobre szacunki co do R1B-Z2103.
Z2103 ma ponad 7000 lat.

MaxT said...

@Davidski

Will you be adding these samples to your basal-rich K7 spreadsheet?

Davidski said...

@MaxT

Yes, I will add the best quality samples from this paper to the K7 spreadsheet, and certainly the Indus_Periphery individuals.

Matt said...

Slight update on those Global25 simulations, using Onge in the regression as a stand in for 100% AASI(Onge-like) seems to produce qualitiative improvements in all the simulations. Although the result for AASI is still significantly better than Onge in all the right dimensions.

Pastebin: https://pastebin.com/feVZKjmP
Graphics: https://imgur.com/a/ezP05

Definitely use these instead if you are going to use any of these sims I've made in G25 models as stand-ins awaiting real samples.

Unknown said...

I think it is ridiculous for this team (and D. Reich), to think PIE began with some Iranrian farmer population. The team itself states that Indo-EUROPEAN language was brought to South Central Asia ~2300BC. If this is related the Indian branch of Indo-Iranian, then the Iranian branch was also brought by this or related migration. So the Iranian branch expanded into what? Was there other offshoots of PIE in Iran that were wiped out with incoming Indo-Iranians?

capra internetensis said...

@Nirjhar

As I recall you've mentioned before skeletal/dental evidence from Mehrgarh showing a population change between the Early Neolithic population, which was closer to Ganga Mesolithic and Maharashtra Bronze Age populations, and the Chalcolithic (4500-3800 BC) population, which was closer to Harappans. The estimated admixture date for the Indus Periphery samples (Zagros Neolithic mixing with AASI) was 4700-3000 BC, so that could fit pretty well, especially considering LD dates tend to be on the low side.

The Aryans were missing in action on the bio-anthroplogical side too, IIRC.

Nirjhar007 said...

Yes, I will add the best quality samples from this paper to the K7 spreadsheet, and certainly the Indus_Periphery individuals.

Today?.

Nirjhar007 said...


Capra,
Yes, around that period a movement is detectable, but we must see samples from India proper before arranging things in suitable manner.

Anonymous said...

1) This is really great that they found some IVC samples, IA, and BMAC samples.

2) Their conclusion clashes with previous analyses. This was AN IA migration into South Asia, but not "THE IA migration.

This migration date is way too close to 1400 BC, by when we know the Mitanni elite had already lost its IA language, except for a few words, names and deity names.

@ Davdiski, do you remember the comparison you did for me between North Indian Brahmins and South Indian ones? You noted that the NI samples were shifted towards Andronovo, while the SI samples were shifted toward Yamnaya.

Is it possible to do the sample comparison with South Indian Brahmins? I am strongly inclined to believe that this is a second pulse of migration, not the first.

Jijnasu said...

@aniasi
they haven't mentioned the date of the migration. All we know is that they migrated south from the kazakh steppe between 2100 BCE and 1700 BCE and that they were in South Asia by 1000 BCE. I believe they will be publishing admixture dates based on the samples from swat in the final paper.

Jijnasu said...

I wonder why they decided to ignore origins of iranic peoples in the paper.

Matt said...

Another G25 simulations; this time taking the Indus_Periphery simulation and AASI simulation from the previous run and making another simulation with -23% AASI, to try and get at a hypothetical "proto-Indus Valley" West Eurasian population position: https://pastebin.com/v4t0Riab

Seems this would be basically West Eurasian based on position when I test by reprocessing it's G25 scores through PCA with a bunch of other West Eurasian populations: https://imgur.com/a/TYA8M

Looking at this, it does seem to me that this would basically accord with the IVC population, net of AASI, being something like a slight exaggeration on the Levant->Iran_N cline, plus a bit of extra AG3 related ancestry 7-10%. (And IVC itself would probably turn out to involve some ancestry from Iran_N...).

(And so the idea that a "pseudo-steppe" of ANE+Iran_N actually could remove *all* steppe contribution to South Asia is probably well wrong... but the idea that this existed in Northern South Asia/Southern Central Asia and would reduce the amount of steppe ancestry required for models of present day people, to levels well below what both Lazaridis 2016 and prior qpGraph modeling found, was probably right. Testing with the real data will be the proof in the pudding!)

(Similar graphics with more labeling for the same kind of reprocessing at the end of the graphics from my last post: https://imgur.com/a/ezP05)

Alberto said...

@Matt

Thanks for the graphs and simulations. I used the Indus_periphery and AASI ones and they work well with modern populations:

https://docs.google.com/spreadsheets/d/1qUuHwk_It5nhX76FRICmVULd7ZuI1ZbGYuav0BSUN_c/edit?usp=sharing

Did you try simulations using the data from Fig. 2C? Looks rather straight forward, though not sure how the resulting ghosts would work. If I have time I'll try myself, but not today.

(As a side note, for running samples fast against different targets anyone might find Xmix useful. I'm happy to run them myself when I have the time, but I don't always do. And since I wrote the script for this and it seems to work for me... Link)

MaxT said...

Intersting, steppe ancestry in South Asians is now "split" between Steppe_MLBA and other half going into Indus_Prersueues (which has Western_Siberian_HG) ancestry. Kalash were 50% 'EMBA' in previous studies, now they are 28% 'MLBA', with other half of steppe going into Indus_Prersueues, being eaten up with Western_Siberian HG.

MaxT said...

Steppe_EMBA affinity in C/S Asia in previous study can now be explained by presence of Western_Siberian ancestry in the region before the appearance of MLBA steppe pastorals.

Matt said...

@Alberto, nice. I was sort of hoping you'd use your script to run a mass nMonte with them ;). I really need to download that and use it. Did you use the second set of sims I created, where I used Onge in the regression? I think those should work better, though these fits are already pretty much as expected (and look consistent with the hierarchical model as they should be). I may download your script and have a go at a big run, though no time to today.

Some slight variations in populations taking Andronovo+Indus_Diaspora_sim+AASI_sim vs Srubnaya_Outlier+Iran_N+Onge seems to come out in the mix. The Iran_Chl and low Indus_Periphery_sim in the Kho_Singanali seems the think that was most unexpected to me given their generally Kalash like impression.

Fig 2c, is that the complex qpGraph? There aren't many populations modelled on it, unless there's something like that in the supplements. I'm a bit wary about estimating up from small proportions using only one or two populations. Any noise or slight complexities in population history will become very amplified.

Alberto said...

@Matt

Yes, I used your latest simulations. And here with the proto-IVC one instead or I_P:

https://docs.google.com/spreadsheets/d/17ADpC3YGdP-1ZMT_CT-MpTfalQyYEIlmkhkDAyeoexE/edit?usp=sharing

No, Fig.2C (unless I'm reading wrong the number) is the qpAdm models using the 6 main ancient populations involved, and represented in coloured bars.

MaxT said...

from the paper, on ANE-rich Central Asian forgers before Yamnaya:

(1)North Eurasian hunter-gatherer ancestry in Central Asia before the Yamnaya expansion. We observe significant West Siberian hunter-gatherer related ancestry (also related to Ancient North Eurasians (ANE) and Eastern European HGs (EHG)

"It is tempting to speculate that this ancestry is also characteristic of the hunter-gatherer Kelteminar culture, which archaeological evidence suggests was dominant in this region in this time, although we do not have any genetic data directly from this population."

Wonder if their next paper will be on Kelteminar.

MaxT said...

It seems like ANE-rich forgers have always been on the move in Eurasia since Holocene, probably has something to do with Late Glacial climate warming? We see them everywhere, mixing with various populations, they are most nomadic folks in pre-history so far.

Anonymous said...

How dire it is that the 3 West_Siberian_HG found were girls. They could be a key for the origins of the R2 yDNA.

Anonymous said...

The paper also showed a substantial amount of Europe_MN in the entire Andronovo horizon. That solves Mallory's "cloud 2":

https://www.proto-indo-european.ru/ie-cradle/_pdf/clouds-over-ie-homelands-nallory.pdf

old europe said...

epoch2013

thank you. that's very good news for me.

old europe said...

epoch

could you quantify in percentage this substantial amount of Middle neolithic ancestry in andronovo?

Anonymous said...

@old europe

From the paper:

"All previously reported samples fall into a subcluster we call Steppe_MLBA_West that harbors ~26% Europe_MN ancestry and ~74% Steppe_EMBA ancestry."

old europe said...

epoch2013

thank you. In the post Time and Space ( i think it was may 2017) on this blog someone commented that that the most EuropeMN percentage was concentrated in elite burials. That is also very very telling.

Rob said...

How dire it is that the 3 West_Siberian_HG found were girls. They could be a key for the origins of the R2 yDNA.”

I bet R1a is also from Central Asia / south Siberia

Bronze said...

Yeah Rob youre probably right, And considering how old Mal’ta boy is with R1b, R1a as well could likely have been in south asia since neolithic times, or at least in south central asia.

Samuel Andrews said...

y DNA R2 already existed in Neolithic Iran without West Siberian HG admixture. Hats off to everyone who detected excess ANE in Srubnaya outlier & Andronovo outlier. I first thought the idea any mostly ANE population existed up to 5,000 years ago was crazy.

vAsiSTha said...

From the preprint
"Second, between 2100-1700 BCE, we observe BMAC outliers from three sites with steppe_EMBA ancestry in the admixed form typically carried by the later Middle to Late Bronze Age Steppe groups (Steppe_MLBA). This documents a southward movement of Steppe ancestry through this region that only began to have a major impact around the turn of the 2nd millennium BCE."

However, no R1a or R1b present in the 9 odd male samples of BMAC 2100-1700BCE.

"Second, samples from three sites from the southern and eastern end of the Steppe dated to 1600-1500 BCE (Dashti-kozy, Taldysay and Kyzlbulak) show evidence of significant admixture from Iranian agriculturalist-related populations, demonstrating northward gene flow from Turan into
the Steppe at the same time as there was southward movement of Steppe_MLBA ancestry through Turan and into South Asia. These findings are consistent with evidence of a high degree of human mobility both to the north and south along the Inner Asian Mountain Corridor"

Out of the 3 steppe male samples 1600-1500bc, one has J2a which is also present in 3 samples from 2100-1700bc BMAC.

Looks like the BMAC male ancestry travelled north.

Davidski said...

@Bronze

Mal’ta boy doesn't belong to R1b. It's a mistake. Many of the Y-haplogroup calls in this paper appear to be mistakes. We're waiting for a revised version to come out.

And no, it's not very plausible that R1a was already present in South or even Central Asia during the Neolithic, and certainly not R1a-Z93, which is what really matters in the context of the Indo-Aryan migrations.

Try to be more accurate and realistic when commenting here. This will help to improve the quality of the discussions and save me from correcting you or even deleting your comments.

And that goes to everyone who wishes to post here.

Vara said...

"We're waiting for a revised version to come out."

When is that coming?

As of now we have 3 non-Steppe R1b; Kura-Araxes, Hajji Firuz and Darra i Kur. Maybe we can add BMAC (I4315) and SPGT (I8998) to that list.

Davidski said...

@Vara

My guess is that a revised version with a corrected spreadsheet will come out within a week or so.

Meantime, I wouldn't get my hopes up that most of those R1b calls in samples outside of the steppe will survive the corrections.

Apparently, the genotype data might be available within a day.

«Oldest ‹Older   1 – 200 of 868   Newer› Newest»