search this blog

Monday, May 7, 2018

Protohistoric Swat Valley peoples in qpGraph #2


Three options. Just one passes muster; the one with Sintashta. Coincidence? I think not. Who still wants to claim that there's no Sintashta-related steppe stuff in these Iron Age SPGT South Asians? The relevant graph files are available here. Any ideas for better models?




Update 08/05/2018: The reason that I chose Dzharkutan1_BA, from what is now Uzbekistan, as the BMAC proxy in the above graphs was because it's geographically a proximate choice for SPGT. However, I've since discovered that Gonur1_BA, from what is now Turkmenistan, does a somewhat better job in these models. The additional graph files are available at the same link as above here.



See also...

Protohistoric Swat Valley peoples in qpGraph

The protohistoric Swat Valley "Indo-Aryans" might not be exactly what we think they are

266 comments:

«Oldest   ‹Older   201 – 266 of 266
mzp1 said...

RK,

After I destroyed your pathetic arguments on anthrogenica u came here. What did u think was going to happen differently this time?

South Asian Dna is mostly unchanged from the Neolithic.

Andronovo are clearly subordinate to Bmac people's. Settling down to farming where they can find land. Making and selling pottery to make a living. Sound like a destitute bunch they do. Then their bmac elites had to rescue them by riding them down to the Pontiac Caspian steppe. Lol and you think they were elite conquerors. You are delushional.

I did wonder though why Andromovo pottery increases inside late bmac settlements. Why would bmac procure those instead of using their own.

Because it was dirt cheap.

Just like your arguments

mzp1 said...

So what are the major ydna haplogroups in Andromovo?

And where are they found in South Asia?

postneo said...

@ryu
I somewhat understand this European fetish for blue eyed brahmans/vedic etc..
but the diff between brahmins and other castes is small ..pretty much follows a geographical cline like others. There are numerous clans more west eurasian shifted than brahmins especially in western india. the studies in these papers are superficial.

And now if more proximate older sources of Iran+EHG/steppe are food in South Asia suddenly the goal is shifted to CHG and EEF as being the de facto IE markers although for yamnaya the rules are different.

Has it ever occurred to any of you chuts, that CHG, EEF and Iran N and EHG are incredibly old components and would have spoken an incredibly large number of languages over time? Even Lazardis seems to have come under this spell of narrow correlation of language and ancient autosomal components.

Ryukendo K said...

@ Shaikorth

Hey thanks!

Looking at the list of pops its striking how high the CHG haplotype contribution is even to groups like icelandic (24%), Mordovian (25%) and Ukraine West (27%) who all follow just after Caucasus pops in peaking in segments shared with Kotias and even displace Turks(!).



@ All
I invite everyone to go to this anthrogenica thread to see who 'destroyed' whose arguments.

Unknown said...

@Chad
"I've gotta disagree here. CHG was long gone by the time the farmers got there. It is a non-factor. They didn't start farming or get involved in farming. We don't even know if they contributed anything to the farmers from Mesopotamia. With how high the European hunter is in the Chalcolithic samples, I'm going to guess it is next to none. Now, it could be that the Iranian group that mixed into Mesopotamians was closer to the ancestor of CHG on the tree than the Zagros samples, but that would be it. It sure as hell isn't CHG-proper in anything. I promise that.

As far as what is being proposed by some that are anti-steppe, is the expansion of the Geoksiur groups. They expanded everywhere, as they make up the majority of the ancestry of the Sarazm group, as well as Shahr_I_Sokhta, and it looks like even the Indus group. That ancestry is actually quite large compared to some rather paltry steppe numbers. I think they propose that IIr was from the Turan region in the Chalcolithic. Then, the expansion of Geoksiur groups into the Indus and surrounds became Indo-Aryan, while those that stayed behind became Indo-Iranian. The Scythians, or their Sintashta ancestors were "Iranianized" by BMAC and they then took this language and Turan ancestry up into Europe. They may see steppe people like Sintashta as something like some parts of the Germanic migration, where you might see their genes, but their overall cultural impact is negligible. They may also see the steppe as non-IE and with the non-IE substrate in Avestan coming from the steppe groups.

Just my two cents on what might be their line of thinking. I could very well be wrong though.. Just some interesting tidbits to follow..."

Would you expand on these thoughts (although you have already written a lot)?

@Ryukendo
"I notice the comment about CHG. I dunno, Haplotypes and chunk-based analyses have served us extremely well so far.

All its inferences have been more or less proven right with aDNA... it pointed to Srubnaya-Sintashta type contribution to S Asia when unlinked markers were pointing to the opposite...

I get that you have archaeological and anthropological reasons as to why CHG may be much reduced (agricultural spread and all) but the haplotypes are sorta just there and can't really be explained in any other way... Lets see what appears in the Maikop genomes."
What are you guys talking about?

@Shaikorth
"Yeah, EHG and Iran_N combining could maybe make a pseudo CHG haplotype which might appear in populations when the fit has insufficient sources. When haplotype methods were used earlier without Han to account for East Asian, Nogais were pseudo WHG. An interesting question is whether Iran_Chl could account for CHG haplotypes.

With sufficient sources a pseudo CHG shouldn't appear, the one population in Broushaki 2016 that behaves weirdly re: CHG and Iran_N is Lezgins (unlike other N-Caucasus samples like Chechens and Ossetians) they appear with lots of Euro HG and Iran_N).
http://science.sciencemag.org/highwire/filestream/681447/field_highwire_adjunct_files/3/Table_S24.xlsx

In any case, sufficiently covered steppe and ANE-related samples are now available so haplotype fits should see major improvements."
Interesting.

Earlier I had asked the following question:

"
Lazardis 2016 modeled CHG as Iran N + EHG + WHG : https://i.imgur.com/CDHMKoI.png

Can Iran N be replaced by any of the new samples (a steppe population?) to make the model be closer to actual CHG?

https://i.imgur.com/RPinzft.png

Any ideas?"

Can the location on a PCA be 'pinpointed' by inferring where the IranN like ancestry of CHG should lie? Can we see if it approaches any of the new Iranian Plateau + Central South Asian samples?

ryukendo kendow said...

@ Alberto

Just saw your comment, hmm this is a very valid observation actually!

"But also look at those 5 populations that they highlight:

Brahmin Nepal Z=3
Bhumihar Bihar Z=3.5
Brahmin Tiwari z=4.3
Brahmin_UP Z=4.1
Bhumihar_UP Z=2.6

Looking at the models, it seems that these 5 groups are autosomally almost identical and seem to come from one source population. The Brahmins from Nepal, Bihar and Tiwari are probably not locals. We could be looking at 5 populations that are essentially the same one, only recently split and migrated to other areas. Or maybe not. But do you know the history of those 5 groups highlighted? More importantly, does the person who wrote the text on the pre-print has a deep knowledge of the history of these groups or not even Wikipedia level knowledge of it?

So you see, the statistical significance depends on the correctness of test and the knowledge about who they are testing.


Question to all the people familiar with Indian ethnography:
How diverged are the populations above actually? From what I know Brahmin and Bhumibar do not have a similar ethnic origin and Nepali Brahmin etc are actually local to the area for quite a long time, not so sure for the UP and so on though.

Looking at the list of Brahmins, is there a chance that we are actually getting a biased sample of recent diaspora of a single group?

Shaikorth said...

@Stefan Molyneux
Depends on whether any of the steppe samples is sufficiently Iran_N-like, which means they shouldn't have extra AASI which pulls towards that corner in a West Eurasian PCA, or ANE from the northern steppe and so on.

If you consider the age of Satsurblia CHG a likely temporal and proximal source would be an Upper Paleolithic Iranian. Alternatively this kind of ancestry was always dominant in Caucasus and EHG/WHG were later intrusions, Kotias apparently is slightly shifted towards Euro HG compared to Satsurblia.

Ryukendo K said...

@ Stefan

Do you know what haplotypes and crossover events are? You can go look up the wikipedia article on 'Meiosis' to understand this. Then look up the article on 'haplotype'.

Caucasian and European populations share a lot of chunks of their chromosomes with the genome Kotias from Jones et al 2015, while South Asians and West Asians share a lot with Iran_N ('Wezmeh Cave') from Broushaski et al 2016. The fact that so many chunks are shared with Kotias instead of Wezmeh Cave+EHG means that the "CHG" effect in Yamnaya and modern populations could not be due to combined contribution of Iran_N+EHG, is what I'm saying.

(Its very unlikely for you to share a haplotype with someone if you do not actually have some kind of ancestry shared with that person at some point. Unless both of you are part of a very inbred population, which kinda means everyone shares some ancestors, which results in the same thing.)


ryukendo kendow said...

^^ If you want to check out the data source yourself, look at Table S24 in Broushaski et al (not Martiniano et al, that was a misremembering).

Jaydeep said...

Ryukendo,

I left out EHG due to oversight. Nothing else. In any case, EHG/WSHG like ancestry is native to South and Central Asia unlike the ANF which is clearly intrusive.

DO go through the links I have given if you haven't already. It is far more probable that PIE arose in the region I specified.

You may recall that in the recent Narasimhan et al paper, the Haji_Firuz_C is taken as source for the southern admixture in Yamnaya. If you look at the proximal qpAdm modelling of various Chalcolithic groups of Iran & Central Asia (Supplementary text, pg 100 onwards), you can see that most of them can be modelled as admixed with the adjacent Chalcolithic groups. So there is clearly evidence of close genetic links across Chalcolithic Iran & Turan.

It is therefore quite probable that some ancestry originating from Central Asia did end up on the steppe via the Caucasus. As I showed via the links, there is also linguistic and archaeological evidence for such a migration.

The Yamnaya_Ozero sample can be even modelled as having 45 % Sarazm_EN.

At the other end, see how the non-ANF admixed Sarazm_EN and now Namazga_EN are so close to the West Eurasian ancestry in South Asians. Infact, if you looked at the figure S3.10 in the Supple Text of the Narasimhan et al paper, the Sarazm_EN shares significantly more alleles with the South Asian Birhor group than do the BMAC main cluster who already have AASI admixture.

All of this suggests, that the Chalcolithic Central Asians are genetically quite close to South Asian populations. Let us wait for Rakhigarhi aDNA. I am positive these links are going to be more solidified.

ryukendo kendow said...

@ Jaydeep

What will falsify your hypothesis?

What are the implications if the IVC is heavily Iran_N, AASI and contains only West_Siberia_N (not EHG, the two are readily distinguishable)?

Something else I want to mention. You are very against the idea that Steppic people could have invaded and made the very civilised people throughout West-Central Asia abandon their languages.

Now, there are multiple reasons you could believe this, some of them social and political, perhaps regarding the differences in population density and power.

But the fact of the matter is, by the Iran_Iron_Age sample ~25% of the ancestry of the Iranian plateau is replaced by Sintashta ancestry.

By the early Iron Age 30% of the SPGT ancestry has been replaced as well (though you believe this may be a native movement, however for this moment accept the premise that it may not be).

In the Armenia_MLBA (Armenia was dominated by powerful Middle Eastern states for a long time by then) Steppe ancestry increased from Armenia_Chl by ~20% as well.

This is less like a structured conquest, and more like a mass immigration and settlement (though friction and conflict surely did exist).

The Iron Age Iranian and Armenia_MLBA shows us that this did happen, it is possible for it to happen. Therefore I urge you to reevaluate the strength of this objection.

Shaikorth said...

The new paper also clarifies the findings of Moussa et al. where the Early Neolithic Baikal HG's in Shamanka II and Lokomotiv were found to be K*, C-M217 and R1a - that K* as many suspected turned out to be N-L666/old N1b. It's an old branch but the MRCA's are recent, somehow BHG's may have ended up in Vietnam during the Metal Ages. https://www.yfull.com/tree/N-L666/

ryukendo kendow said...

^^ Armenia_EBA, not Chl.

Seinundzeit said...

Yes!

Good stuff.

Preliminary thoughts:

1). If Namazga_CA is similar to Sarazm_Eneolithic, it won't be the right population to capture the West Eurasian stream of ancestry now specific to contemporary South Asia, and peaking in Shahr_I_Sokhta_BA2 (which is 80%-90% West Eurasian, of a kind that is highly unique and "Indian"-like).

Assuming similarity to Sarazm_Eneolithic, Namazga_CA will turn out to be far too "western" for South Asia's unique West Eurasian component.

And, if similar to Geokisur_Eneolithic, it will be an even poorer proxy.

Since this paper doesn't include the INP samples, or the other ancient Central Asians, Namazga_CA seems like a fairly decent proxy, especially in comparison to Iran_N.

Again, the best sample for this is really Shahr_I_Sokhta_BA2, since it only has 10%-20% AASI.

But David's qpGraph strategy is probably the best way forward, until we find ancient "Indian-like" samples missing the 10%-20% AASI of Shahr_I_Soktha_BA2.

His method of using Sarazm_Eneolithic and INP to infer a West Eurasian ancestral population for South Asia seems to work pretty well in qpGraph.

2). Indus_Periphery samples prefer neither EHG nor West_Siberian_N.

AG3 and MA1 still do the best job. Ditto for many of the ancient Central Asians in the Narasimhan et al. 2018 preprint.

3). I guess a Caucasian/West Asian PIE homeland is now the most parsimonious model. Perhaps EHG people spoke something with an affinity to Uralic? (Don't kill me; I'm just spit-balling)

4). That being said, and as RK has been bravely arguing, both this paper and Narasimhan et al. strongly suggest a Steppe_MLBA-related introduction of Indo-Iranian languages into Central Asia, West Asia, and South Asia.

Indo-Hittite, probably West Asian; everything else, probably steppe. We'll see if further samples modify this picture.

With this being duly noted, the Narasimhan et al. preprint does also seem to suggest that this Indo-Iranization wasn't the explosive phenomenon we previously had in mind, due to the 50%-40% Yamnaya percentages we saw for Kalasha and Brahmins.

Rather, it probably involved some very complex dynamics, with gradual gene-flow and acculturation over a long period of time. Also, the "male-mediated elite conquest" conceptual scheme was probably wrong.

5). Botai samples! We'll see what effect these samples have for the modelling of the ancient Swat valley folk (and for Burusho, Khalils, etc).

For what it's worth, I think the Steppe_MLBA signal will go down for those aDNA samples, with Botai in the mix.

Also, I was wondering about the East Asian component in the ancient Swat valley; didn't know what to make of it. Now though, we know that the Botai folk had noticeable ENA admixture, so that aspect is explained.

6). They have Kohistani, Gujar, Yusufzai, Uthmankhel, and Tarkalani data! This must be a dream. This will be very, very fun...

Jaydeep said...

Ryukendo,

This is going to be my last post for sometime as I have to catch up on some work.

My hypothesis is based on the ancient Indian literature which talks of no Indo-European migration into India but talks of the banks of ancient extinct Saraswati in the region of Haryana & Western UP as the Vedic and possibly other IE groups and it also mentions groups that eventually migrate into Central Asia and come to establish several kingdoms among those 'mleccha' people. I also have a linguistic hypothesis and archaeological evidence showing how from Central Asia, the IE groups could have spread onto the steppe and the Near East.

Anthony & Ringe had argued recently how, wheeled vehicles originated among the Late PIE people in the 4th millenium BC. Well, wheeled vehicles also appear among the Harappans in the 4th millenium BC. Whats more, it is only the Indo-Iranians who have the entire wheeled vehicle vocabulary intact, something which would make no sense if these Indo-Iranians came from the steppe into Central & South Asia when already these Southern groups were suing wheeled vehicles for more than 2 millenia. How would Indo-Iranians manage to keep their entire vocabulary intact ? Makes no sense.

------------------------

Genetically, to falsify my hypothesis, you would need significant amounts of AASI in pre-3000 BC Indus civilization inhabitants. After 3000 BC we already know that AASI was also present across Eastern Iran & BMAC, so a genome from Northwestern South Asia dating to after 3000 BC may have significant AASI but it will not falsify my hypothesis.

I do not know how WSHG/EHG can be segregated as yet since this Willerslav paper argues for EHG in Namazga while the Reich team argues for WSHG across Central Asia.

On the other hand, they separate CHG ancestry from Iran_N while the Reich team does not. So these are things that need to be worked out.

But if pre-3000 BC samples from Indus civilization cannot be shown as admixture sources for the Chalcolithic Central Asians, it will dent my model. My model also predicts that there will already be ANF affinity in IVC samples before the steppe_mlba period.


Regarding Iron Age Iran, I think there is evidence that the Iranian language speakers probably came to present day Iran from the East. The East naturally has more EHG/steppe so it does not prove a steppe origin for the steppe_mlba signal.

As for SPGT, it clearly has linkages with BMAC which increases its ANF ancestry quite a bit. As for its steppe related ancestry, there is quite a bit of variation as per the PCA, with some SPGT samples being close to the Indus_Periphery while others much higher. But you have to find significant no of IVC genomes from an earlier era to show that the steppe-related ancestry in IVC was uniformly as low as the Indus_periphery. Already Indus_P samples do show variation in their WSHG ancestry as per the qpAdm models.

As for Armenia, we do have EHG related ancestry in Armenia_Chl. Where could that come from ? Since its y-dna is L1a, which is quite prominent only in South Asia, it clearly supports my argument of a east to west population movement from Central Asia in the Chalcolithic.

EastPole said...

@ryukendo kendow

“the IEs were in fact a very primitive culture, however they were extremely competitive and warlike“

How do you explain then that this very primitive culture became the source not only of some of the world’s richest and profoundest religions, but of some of its subtlest metaphysical philosophies.

Corded Ware and Sintashta cultures were not primitive. Believe me. People have been brainwashed by Pan-Germanic and Pan-Semitic propaganda over centuries. Great Civilizations of Middle East, of India, of China, of Western Europe etc. Nothing came from Eastern Europe (Hegel). Now it looks like everything came from Eastern Europe.

Shaikorth said...

@Sein
There's a haplotype comparison in Fig s31. (Yamnaya vs Botai and Botai vs more Okunevo/East Asian steppe). Burushos, interestingly enough, do not look more Botai shifted than Tajiks and do not look more CentraSteppe_EMBA-shifted than Pathans. It might be prudent to look at Tibet as a possible source of Burusho ENA, they have a preference for Tharus and Yi in some of the Broushaki fits.

Jaydeep said...

Sein,

Why do you think that the West Eurasian in South Asia is quite different from what it is among Central Asians ? I am inclined to think that they shared a common origin . Perhaps it is possible that there was gene flow between Central & South Asia during the mid-holocene but that there were already ancient groups of West Eurasian origins in both the regions which had diverged during the late Pliestocene, kind of like the early split of CHG.

This theory is supported by the mtdna U7 paper which had come out last year.

Ryukendo K said...

@ Jaydeep

"But if pre-3000 BC samples from Indus civilization cannot be shown as admixture sources for the Chalcolithic Central Asians, it will dent my model. My model also predicts that there will already be ANF affinity in IVC samples before the steppe_mlba period."

Let's hold you to this standard. Its likely that the Rakigarhi samples will not be pre 3000BC, and it will may well be many years before we get any sample so old. But it is still a courageous thing to place a stake through the ground.

"Regarding Iron Age Iran, I think there is evidence that the Iranian language speakers probably came to present day Iran from the East. The East naturally has more EHG/steppe so it does not prove a steppe origin for the steppe_mlba signal."

"As for SPGT, it clearly has linkages with BMAC which increases its ANF ancestry quite a bit. As for its steppe related ancestry, there is quite a bit of variation as per the PCA, with some SPGT samples being close to the Indus_Periphery while others much higher. But you have to find significant no of IVC genomes from an earlier era to show that the steppe-related ancestry in IVC was uniformly as low as the Indus_periphery. Already Indus_P samples do show variation in their WSHG ancestry as per the qpAdm models.

As for Armenia, we do have EHG related ancestry in Armenia_Chl. Where could that come from ? Since its y-dna is L1a, which is quite prominent only in South Asia, it clearly supports my argument of a east to west population movement from Central Asia in the Chalcolithic."


Unfortunately, these statements can only hold because you obviate the distinction between EHG, WSHG and ANE. They are very different and cannot possibly stand in for each other (as David has already shown, it is not possible to model the SPGT as differing from previous samples by increased WSHG only.

In the Willerslev paper, you already see the extremely different PCA positions of Botai (likely similar to WSHG), MA-1 and EHG samples.

(As Sein points out, and its true that the most negative f3 statistics for Eastern Iranian Neolithics and in the Indus Periphery is for ANE and not even for WSHG, i.e, there is no evidence for EHG, or even WSHG in India and surrounds, and I shouldn't have conflated the two.)

In fact, part of the reason why SPGT cannot be that significant in mainline Caste IArs is because they have too much ANE or WSHG ancestry and too little EHG.

Sintashta, Armenia_Chl (which is different from Armenia_EBA--what I meant to mention) and SPGT all have EHG ancestry and cannot be modeled purely as deriving from local West Eurasian substrates.

When you come back from ur work, I'd like to hear if you accept an absence of EHG in SC Asia (and presence of only WSHG and/or ANE) as another thing that will falsify your theory. This position will probably be evidenced much earlier. Theoretically it would, but I'm open to arguments as to why it won't.

Anonymous said...

@Davidski
Another thing that you don't understand is that no one has ever claimed that Steppe_MLBA was the Proto-Indo-European population. How could it be, since it dates to the Middle to Late Bronze Age? It's impossible.

PIE as the parent of extant IE languages (the way the term was coined) dates to MLBA and there are no extant IE speakers without Steppe_MLBA admixture.

I guess many (most?) misunderstandings in PIE discussions stem from mixing up PIE and PIH.

Jijnasu said...

@jaydeep
This tradition of 'founding' of Kingdoms in central asia is very doubtful. Other versions of the yayati legend and the geneological lists identify the druhyus with the people of Gandhara and the anus with the other kingdoms of the panjab. A fair statement is that there is neither any memory of any migration to or from India

Matt said...

@Ryu: Re: CHG, I just wanted to comment and put my 2c on it, my stance is pretty much the same as the last time I discussed with Chad (and we both agreed not to discuss further w/ each other pending further data);

Possible CHG represents some stream from outside North or South Caucasus (even likely, given languages from present day CHG heavy people probably were not in Caucasus before Bronze Age and are probably the results of later migrations there that replaced earlier people to a degree).

However no good reason to label this stream as "Iranian farmers" when they likely were NW Iran at most, and no more likely to be NW Iran than Armenia / Eastern Anatolia. (This seems to me like an odd extension of the terminology in Reich's opus where CHG and Armenia_Chl appear to be labeled "Iranian farmers" for some reason).

CHG ancestry signals in haplotypes, Fst, probably f2 being a pseudo-effect of ancestry from Ganj_Dareh like people (West Iran, about Central on N-S axis) + Barcin_N + EHG ancestry I still give absolutely no confidence to.

Seinundzeit said...

Shaikorth,

Ah, I just started reading the supplementary materials!

I think you are absolutely correct on the Tibetan angle.

Interestingly, some aspects of Burusho architecture are of Tibetan origin, so the cultural linkages are certainly there.

Jaydeep,

The West Eurasian ancestry of Indus_Periphery (which ranges from 90%-80% in Shahr_I_Sokhta_BA2 to 65%-50% in Shahr_I_Soktha_BA3. Also, Shahr_I_Soktha_BA1 is at 60% for this sort of ancestry, with the rest of its genetic heritage involving an even mix of Iran_Chal and Iran_N) is much more "eastern" than even Sarazm_Eneolithic.

By "eastern", I mean less WHG-related ancestry, less "Basal Eurasian"-related ancestry, and more MA1/AG3-related ancestry. Also, when it comes to something like PCA or ADMIXTURE, it's far more "Indian"-like in comparison to Sarazm_Eneolithic. Essentially, this means that it shares a great deal more genetic drift with contemporary South Asians than Iranian plateau and Central Asian Neolithics do.

For what it's worth, one can probably model Sarazm_Eneolithic as a mix of Iran_N and the West Eurasian stream of ancestry seen with Indus_Periphery.

I'm guessing Namazga_CA is similar, but I should read the supplementary materials before making any further remarks.

Matt said...

@Anthro_Survey: "I haven't had a chance to go through it properly yet, but I, too was a bit confused about the Glazkovo-BHG distinction. So, which one is plotted in the paper's PCA and ADMIXTURE? LN/BA, right? In fact, I don't see any figs directly analyzing BHG. From what Alex says, I take it that Glazkovo_BA has less pristine ENA ancestry?"

LNBA; but as Shaikorth says, the other paper does have more proper coverage of Baikal_HG, who are about 89% Glazkovo_BA analysis in that supplement, rest North Eurasian (though they do not analyse with West_Siberian_N samples that only Reich lab has access to, so choose to use MA-1 for model).

Ultimately, I'd like to see some formal stat tests of whether Baikal_HG do have any ANE relatedness themselves as well, compare vs Xiongnu / recent East Asians, which would affect matters. These are not present in the paper though.

Ryukendo K said...

@ Matt

I'm interested to hear your opinion about the Baikal HG being separated from Han by 15kya.

About as separated as CHG is from Iran_N.

This is very weird to me, where could they have been hiding? They could not have been hiding in Siberia (since it was occuppied by MA-1 @ 23kya). That said, since they are so deeply diverged from the major component of ancestry in modern East Asians (Pleistocene separation) its possible for it to display 'exotic' effects like being more ENA than any ENA population alive today despite being post-Neolithic (temporally speaking) and approaching metal Ages in East Asia.

In PCA space the closer we approach to Northeast Siberia the worse fits become for all populations, not possible to get >10% even in Yakut, much less Yukagir and Nenets. The Baikal HG will be very illuminating in this respect.

Ryukendo K said...

Hmm actually Afontova Gora was found at 16 kya, so Baikal HGs must have rushed into Siberia almost immediately after him (if that occupation of Siberia was what resulted in the separation between Baikal_HG and mainline East Asian ancestry).

Shaikorth said...

@RK
The poor fits in Global25 are probably due to recent drift in various Siberian and Native American populations creating their own dimensions. If that's the case BHG won't help too much. For example when I try to fit Clovis now in scaled global25, the components are of the right type but the distances are way off. Adding MA-1 or Evenks etc. doesn't improve things.
"distance%=37.9579 / distance=0.379579"
Clovis

AfontovaGora3 52.6
Ulchi 43.0
Han 4.4

Ryukendo K said...

@ Shaikorth

Hmm could you explain something to me? Wouldn't dimensions explaining a lot of variance across more points be well-represented in the first 25 dimensions and dimensions that explain less (i.e. accounting for only the distance between Yakut and everyone else) be exiled into some higher dimension, say dimension 30, and therefore just be pruned from Global25 and just disappear from the dataset?

In the lower dimensions populations that have a lot of drift private to themselves will not be special, no?

This is how I expect the math of PCA to treat genetic variation, is something wrong w my understanding?

Shaikorth said...

@RK
That depend on how a) inbred and b) numerous the samples are. In Global25 we see, for instance, two dimensions which apparently reflect modern Siberian internal variation.

https://drive.google.com/file/d/1k6z0j1Rt0zSHwQqpJ2OO_xcxX9UX0Zbz/view

ryukendo kendow said...

@ Shaik

I see, so my understanding is correct but the issue is there are just too many Siberian populations and they are all too inbred so they account for a lot of the covariance?

Shaikorth said...

That's what it looks like, if correct BHG won't cause major distance improvements in nMonte when fitting Native Americans, more drifted Siberians etc.

Davidski said...

Intra-Siberian drift and variation in the Global25 will help track more accurately recent population expansions, like that of the Uralics.

If you want to model deep ancestry with Upper Paleolithic genomes, there are better methods.

ryukendo kendow said...

@ Alberto

Alberto, I uncovered some information that may tell us about the origins of Brahmins as well.

Reich et al sampled ~260 different caste populations in the study The promise of discovering population-specific disease-associated genes in South Asia

Of these, they found about 81 different groups with founder effects stronger than Ashkenazi jews and Finns, with some extreme examples like Gujjar expanding to >1 million people from just 100 persons or less in the last 100 generations.

They also have a table of cross-caste significantly elevated IBD sharing (Table S4), identified using the criteria "Groups with more than one match for high shared IBD across groups (greater than 3 times the IBD score of CEU and ~1/3 the founder event strength of Ashkenazi Jews)" (sounds reasonable) and among Brahmin groups they are:

1. Brahmin_Catholic_Goa, Brahmin_Catholic_Kumta, Brahmin_Catholic_Mangalore
2. Brahmin_Nepal, Brahmin_Uttarakhand
3. Brahmin_Karnataka, Havyaka_Brahmin

All others are non-Brahmin groups with Dravidians and especially Austroasiatics overrepresented.

So on the face of it I'm not sure that the objection you raised may apply here, it seems most Brahmin group underwent their own founder separately event except for a few cases where divisions of the community occurred.

zardos said...

@Rob: Both in Greece and Anatolia we know that there were non-IE from the start of recorded history, unlike other parts of Europe settled en masse by steppe related people. If Hittites would have been local and not of steppe origin, how PIE made it to the steppe would be much more difficult to explain than the other way around. The whole PIE culture and society has a profile which doesn't fit into the Anatolian context of that time. I remember when reading various books about the Hittite/Anatolian branch origins in the 1990's not a single author came up with a local origin theory. And so far, I heard no convincing one.

The case is even much more obvious for South Asia, where there is no way, even with the widest stretch of imagination, to explain the Indo-Aryans with a local origin. This was clear before the genetic results, but now a local scenario is not just highly unlikely, it is completely impossible and scientifically falsified.

Those talking about "primitive" PIE and Indo-Iranians should turn their head to Sintashta and its innovations. Look at Arkaim and say again that they were primitive for their time, even in the Eurasian context, for Eurasian standards of their time, that is absurd. They were close to a mass production of spoked wheel chariots and metal weapons.
I remember a comparison of weapon quality in Europe, which goes in the same direction, with the Swords Iron Age people from the North made were of much higher quality. You don't make high quality metal objects and such innovations in a small scale, culturally primitive social environment. That doesn't fit.
Just because they didn't build large temples, because they were no priest state culture, no theocratic society like many of the early farmer societies, doesn't mean they were primitive.

Talking about that, the position of priests and philosophy might be as much or even more of an IVC cultural element of the Iranian Neolithic side of ancestry in Indians as it is Aryan. Obviously, being vegetarian and worshipping cows is nothing from the original Aryans, but a rather strange development which happened in situ.
It was just possible because of the elite position the mixed Indo-Aryan/Iranian farmer upper class was able to secure. It fits better into a model of a leisure class which could protect itself from "the hard life" of the commoners. So I doublt it was brought to India, but it developed there.

There is the debate about warrior castes vs. priests in much of the older literature. I guess the priests gained more importance because of later conquests, because they could seclude themselves better from foreign influences than the warrior castes and worked as the completely conservative element keeping old traditions alive (or creating new ones which appeared old), which the warriors weren't.

Now I know some people might say that I don't know enough about this or that, and they are right, but the details you can add don't change the big picture.

ryukendo kendow said...

@ Santosh

You should check out "The promise of discovering population-specific disease-associated genes in South Asia", you might go gaga over it, which includes a very large number of Dravidian groups and their cross-caste IBD connections in table S4. This would be very interesting in teasing out the history of various castes and tribes, and whether or not it matches with the interconnections given by oral history.

Seinundzeit said...

In all fairness to Global_25, I think it does a very good job of modelling deep ancestry, even if the fits are poor.

Reference populations:

AfontovaGora3
WHG
IVC_West_Eurasian_Ghost
Natufian
DA45 (ancient East Asian sample)
AASI_Ghost

West_Siberia_N:

87.8% AfontovaGora3
10% DA45 (ancient East Asian sample)
2.2% WHG

distance=4.8111

EHG:

70.4% AfontovaGora3
28.8% WHG
0.8% Natufian

distance=6.2331

Ganj_Dareh:

92.2% IVC_West_Eurasian_Ghost
7.4% Natufian
0.4% DA45 (ancient East Asian sample)

distance=9.8721

CHG:

84.8% IVC_West_Eurasian_Ghost
8.6% Natufian
6.6% WHG

distance=18.2741

Sarazm_Eneolithic:

86.4% IVC_West_Eurasian_Ghost
10.2% AfontovaGora3
3.4% Natufian

distance=6.4899

Levant_N:

88.4% Natufian
8.6% IVC_West_Eurasian_Ghost
3% WHG

distance=10.8998

Barcin_N:

69.4% Natufian
15.6% WHG
15% IVC_West_Eurasian_Ghost

distance=19.4257

DA28:

85.8% DA45 (ancient East Asian sample)
14.2% AfontovaGora3

distance=19.8956

Pretty cool, and very sensible.

Rob said...

@ Zardos
I think you’re confusing me with RK who said PIEs were primitive
No need to waste time and space with strawmen about the Anatolian theory or barnyard factoids like “Anatolia was diverse and well populated” to explain away your lack of a coherent synthesis
Keep it for your ring of steppe - amateurs

Shaikorth said...

@Ryu

IBD comparisons work well in explaining recent population events in Siberia as well. Nganasans, Evenks and Yukaghirs are mostly Baikal-HG (minor differences according to Damgaard et al, with Nganasans having some western mixture but that just translates to slightly more general European IBD sharing), and have symmetrical IBD relationships outside their immediate neighbourhood. So their private population history (genetically if not linguistically) happened in isolation, after BHG lived and expanded.
Nganasan vs Yukaghir IBD sharing for demonstration:
https://c.radikal.ru/c19/1801/e7/9e2d9d44ace0.png

I would expect a comparison between something like Pulliyar and Irula would reveal a similar picture in South Asia.

ryukendo kendow said...

@ Zardos

I said "PIEs" were primitive, not Sintashta and BB and so on. Technically even CW used were still stuck in Neolithic technology.

Chad said...

RK,

Fine structure and all that is difficult when it comes to aDNA. Too much damage, and a sample size of two is too insufficient to really say anything. Archaeology says they were gone. aDNA is showing the same. Armenia is looking too northern to even be a source going from Khavalynsk to Yamnaya. EBA, being more like Iran ChL is a much better fit.

zardos said...

@Rob: I know you didn't say that, I just didn't differentiated in my post, sorry for that.

@RK: You have to put PIE into context. And the context is not LBA, but Neolithic and Chalcolithic times and not Mesopotamia, but Europe North of the Alps. In that context they brought a lot of innovations, which were more advanced.
Obviously, they were no high culture, but primitive they were not, unless partially if being compared with the most advanced cultures of their time. Which doesn't make sense for their context, because they established first in Europe North of the Alps. And it seems that PIE were used to metal technology already, regardless of whether all their tribes used it when they expanded in various directions. So one could even ask whether some Corded Ware tribals on the move were actually below the standards achieved in their homeland. But the other question is whether they completely forgot the knowledge their ancestors had.
It reminds me of some Germanic tribes which didn't build ships for generations, but once they had to, with their old knowledge and probably some old transregional connections, they were able to build a fleet in their ship building tradition almost instantaneously - Vandals come to mind.


Davidski said...

@Chad

How is it that Abkhazians from the northwest Caucasus show such an extremely close relationship to CHG and not to any Iran_ChL population in all analyses, and also to the non-EHG part of Yamnaya?

How did this "faux" CHG-related affinity form exactly in the northwest Caucasus in terms of genome-wide and uniparental markers?

Archaelog said...

PIE speakers definitely had advanced metallurgy, at least from the Yamnaya stage. But they were pastoral nomads, "primitive" in that sense. Also chariots which caused something of a revolution in warfare was probably invented by MLBA Indo-Europeans and the Indo-Aryans definitely made use of it, according the the texts.

I don't think though that steppe incursions were the primary reason for the so called Bronze Age collapse in many regions. They just capitalized on that collapse.

zardos said...

@Rob: You have no better concept but just raising doubt. Which is ok insofar, as that we still don't know for sure. But for excluding a steppe origin of Hittites, more than a handful of genomes of some Anatolians in a wider Hittite cultural context is needed.
Just look at the Hittite laws and customs, their pantheon and approach to life. That is clearly different from the local traditions South of the Caucasus. And yes it is important that Anatolia was never fully Indo-European and a pre-IE population of Hattians is the predecessor in the region. That can't be ignored.
There is no general rule for a language shift. It can be demic diffusion on a large scale, it was most of the time, but it doesn't have to be. The more advanced the administrative structure is, the more populated and productive an area, the more likely it was a top down transmission.
The Mitanni were just another example in the Near East, which impact was even smaller and therefore not able to cause the shift. But the principle might have been largely the same.






EastPole said...

@ryukendo kendow

“I said "PIEs" were primitive, not Sintashta and BB and so on. Technically even CW used were still stuck in Neolithic technology.”

CWC evolved by assimilating highly advanced Neolithic cultures like Tripolye, TRB, GAC. No need to assume that the resulting culture was primitive. How do you explain the expansion and influences it exercised on the development of languages and religions of India, Iran, Greece and Western Europe. I don’t think a primitive culture would be able to do it.
If we investigate common elements in poetry and religion of Vedic Aryans, Greeks and Slavs which no doubt had to be derived from CWC we can conclude that poetry and religion of CWC was quite advanced.
These observations are based on real languages, real poetry and elements of real religions which survived.

On the other hand nothing certain is known about primitive PIE before CWC.

Shaikorth said...

In case someone wants to genotype "The first horse herders and the impact of early Bronze Age steppe expansions into Asia" samples.

Y-DNA and mtDNA
https://zenodo.org/record/1219431
ENA accession for genomic data: ERP107300 and PRJEB26349

Rob said...

@ Zardos

“You have no better concept but just raising doubt. Which is ok insofar, as that we still don't know for sure. But for excluding a steppe origin of Hittites, “

Lol no. just amassing the evidence
Death is something best served well done
I’ll leave the half baked theories to you steppe girls

a said...

Hittites had 3 man spoked chariots. However-Hittite language- lack all[complete] the specific names for parts of a wagon.Proto-Afroasiatics-Natufians-and or Basal Eurasians[other than advanced Sumerians a-non Afro-Asiatic based language] had no word for wheel/[just like modern day tribal groups found in various parts of the world]while spreading EEF component into Europe. Sintashta made spoked wheels-different altogether from earlier solid wheels. Now with these new EEF samples from "Hittites" I eagerly await an explanation how the proto-Indo Aryans -Sanskrit-derived the words for wagon/wheel?

mzp1 said...

Yamna, CWC et all are all extremely primitive compared to IVC.

Indo European is the most advanced culture in Eurasia.

IVC is the most advanced civilization in Eurasia for it's time.

Sintasta levels of trade, metallurgy and urbanisation can only be derived from South Central Asia. There is no development Yamna->CWC->Sintashta that can lead us to Sintashta social organization. But we can get 'worker' populations that way.

postneo said...

@Ryu
UP is a recent well known source for recent migration of brahmins to surrounding areas in the last 1000 years, e.g kanyakubja brahmins etc. Some through invitation the others would due to exodus. there are a huge number of endogamous units with Brahmanas numbering in the hundreds in north india alone. Only a small subset are vedic chanters.

Anonymous said...

Chetan said...
" PIE speakers definitely had advanced metallurgy, at least from the Yamnaya stage. "

no, the Indo-European language contained only the metallurgical vocabulary of the primitive level of the Copper age associated with BCMP. Contained only notions copper and gold, and one thing the notion of forging metal, but not casting. The word axe comes from the concept of the stone. He did not know the vocabulary of the Bronze age, bronze, silver, had no idea metal casting, but Yamnaya as used for machining metal casting metal.

Indo-European language corresponds to the stage of Dereivka and CWC.


Sanuj said...

"Brahmin Nepal Z=3
Bhumihar Bihar Z=3.5
Brahmin Tiwari z=4.3
Brahmin_UP Z=4.1
Bhumihar_UP Z=2.6"

Bhumihar_UP and Bihar are practically the same people, these are bordering states, and these people live between the two. They are also not classified as proper Brahmins, and were infact classified as Shudra by british initially, but are more of a landed working class.
Brahmin Tiwari, should be close to Brahmin UP, as Tiwaris are basically one class of Brahmins from UP. Brahmin Nepal would be a little different, but they still have relations with Brahmins of UP/Bihar.
Also,
"Brahmin_Catholic_Goa, Brahmin_Catholic_Kumta, Brahmin_Catholic_Mangalore"
They are converted Brahmins who were converted during the Portuguese inquisition of Goa. They may have admixture from Portugal as we can't be sure about intermixing. I would trade them for actual Brahmins from Konkan or Maharashtra.

@Jaydeep

The paper says this explicitly now,
"...Yamnaya pastoralists, whose IE language may have evolved under the influence of a Caucasian language, possibly from the Maykop culture."

We also had the India specific mtDNA found in Maykop.
http://eurogenes.blogspot.com/2015/10/mitochondrial-dna-from-maykop-wolfgang.html
This was brushed away earlier but now that Maykop is becoming a serious source of IE in Yamnaya, this becomes serious, and needs an explanation of how Indian mtDNA reached Maykop and with what sort of ancestry.

zardos said...

The crucial point is not whether or not PIE had Caucasian influences, which seems to be a sure thing, but which part was constitutive and which contributed. Since we deal with a strictly patriarchal culture and societal system, the patrilineages are key.

Shaikorth said...

Links to the genome data:

https://www.ebi.ac.uk/ena/data/view/PRJEB25389
https://www.ebi.ac.uk/ena/data/view/PRJEB26349

Samuel Andrews said...

Btw....

The previously published Scythian genomes aren't representative of all Scythians or IA Eurasian Iranians. They have much more Asian ancestry than many of the new IA Iranian genomes. The historical Wusun Iranians, whom the ancient Chinese knew, appear to be like 80% Andronovo.

Alberto said...

Thanks @RK for looking that up (about the Brahmin groups), and thanks @Sanuj for the further comments. I think we'll get a more accurate picture of these things in the final version.

Now on to the new data...

Couldn't look at most of it yet, but that French-like DA111 (R1b-P312 mtDNA: H6a1a) is an early Celt from Hallstatt (Lovosice 2; Czech Republic) C. 800 BCE.

Top 40 closest populations in G25: https://imgur.com/a/VtQucmV

Matt said...

Ryukendo: I'm interested to hear your opinion about the Baikal HG being separated from Han by 15kya. About as separated as CHG is from Iran_N.

Pedantically, divergence time for Shamanka_EN and Han is only about 8kya at the actual time of Shamanka_EN. Also where are models with split of Iran_N and CHG? I can only find models with split of KK1 and Yamnaya_main, and EHG and West Siberian.

I like seeing use of these momi models, not sure how to take some of the date stuff tbh. I mean EHG probably did not split from ANE but rather has 20-30% Villabruna admixture? So how does this affect estimation of split time?

Like, assuming EHG/MA1 are admixed, how much is true divergence time suppressed/inflated? You sort of don't have to think about that with whatever qpAdm is doing, relations to populations off the graph won't be fitted which is a problem, but it's all abstract f2 drift units. But the question becomes salient here, where the models sort of inevitably reify things as actual chronology.

Unlike with qpGraph, I'm also not sure what the way is to test goodness of fit with these models. Or if there even is a test of the topology; in qpGraph, you'd reject all the models where Kotias and EHG/ANE diverge after Shamanka/Han splits off, but here not so. Would have been good to see qpGraph tests of same models.

All of which is a roundabout way of saying I have no way to usefully answer your main question here!

(Also, have to comment, you are probably a better man than I to continue to engage in good faith despite the ad hominins without real hypotheses, and the "Indo European was from the IVC, because IVC advanced! IE most advanced!" level of argumentation.)

Alberto: Couldn't look at most of it yet, but that French-like DA111 (R1b-P312 mtDNA: H6a1a) is an early Celt from Hallstatt (Lovosice 2; Czech Republic) C. 800 BCE.

Interesting that this sample is akin to Beaker Southern France average; IRC found Beaker Southern France average actually seems closer to modern day English in G25 than Beaker Britain average. Not sure about how Welsh / England_Roman / England_IA (all seem roughly interchangeable on the Steppe-Neolithic axis) lie on that continuum.

Though probably the more salient model considering trying to model expansion of Celts would be England_IA as England_LBA+DA111 (or ideally more samples), rather than Beaker, since England_LBA had already had some shift towards the Neolithics, either from geneflow with continent or local absorption (so considering early Beaker Britain and DA111 would probably inflate DA111).

Unknown said...

@Shaikorth
"Depends on whether any of the steppe samples is sufficiently Iran_N-like, which means they shouldn't have extra AASI which pulls towards that corner in a West Eurasian PCA, or ANE from the northern steppe and so on.

If you consider the age of Satsurblia CHG a likely temporal and proximal source would be an Upper Paleolithic Iranian. Alternatively this kind of ancestry was always dominant in Caucasus and EHG/WHG were later intrusions, Kotias apparently is slightly shifted towards Euro HG compared to Satsurblia."
Interesting, thanks for responding! With the lack of samples from Iran and South Central Asia, is it possible to create 'theoretical' populations that might have existed at certain time frames? For example, you mentioned that the IranN like sample would have to lack AASI. Presumably, many of the new samples would have admixed with ASI/AASI at a specific time - would it be possible to create theoretical populations without ASI/AASI admixture (i.e. solely west eurasian)? I think the new papers lend credence to ASI/AASI admixture spreading north as time progressed. Presumably many of the new samples from Iran and Central Asia would have had ancestors without ASI/AASI admixture.

@Ryukendo
"
Do you know what haplotypes and crossover events are? You can go look up the wikipedia article on 'Meiosis' to understand this. Then look up the article on 'haplotype'.

Caucasian and European populations share a lot of chunks of their chromosomes with the genome Kotias from Jones et al 2015, while South Asians and West Asians share a lot with Iran_N ('Wezmeh Cave') from Broushaski et al 2016. The fact that so many chunks are shared with Kotias instead of Wezmeh Cave+EHG means that the "CHG" effect in Yamnaya and modern populations could not be due to combined contribution of Iran_N+EHG, is what I'm saying.

(Its very unlikely for you to share a haplotype with someone if you do not actually have some kind of ancestry shared with that person at some point. Unless both of you are part of a very inbred population, which kinda means everyone shares some ancestors, which results in the same thing.)"

I think I have a hazy understanding of what your getting at, I doubt I 100% follow though. A few questions if you don't mind:
1. CHG can be modeled as having part of its ancestry from IranN like populations. While CHG has a specific haplotype, would the IranN like portion of CHG's haplotype be a sub-set of the modal haplotype in the IranN like population? It seems like IranN was expanding everywhere - couldn't expansion lead to a loss of haplotype diversity.

2. While the CHG haplotype spread by steppe pastoralists is clear, what is the origin of CHG itself? If CHG is a sub-set of 'IranN' haplotype that became 'fixed' in a population migrating out of Iran, shouldn't we able to now more closely model the probable sources of CHG's ancestry with any of the newly available samples?

@Seinundzeit
"In all fairness to Global_25, I think it does a very good job of modelling deep ancestry, even if the fits are poor.

Reference populations:

CHG:

84.8% IVC_West_Eurasian_Ghost
8.6% Natufian
6.6% WHG

distance=18.2741

Pretty cool, and very sensible."

What is IVC_West_Eurasian_Ghost? And surely there must be better models of CHG than the one you produce (don't mean as insult, I have to go somewhere and don't have time to be diplomatic). Is there any other IranN like pop that could be the source of CHGs ancestry? A distance of 18.2741 can be improved no with different samples?

Seinundzeit said...

@ Stefan Molyneux

I don't think you understand what I did there.

You see, the West Eurasian component unique to INP (again, 80%-90% for Shahr_I_Sokhta_BA2, 75%-85% for Gonur_BA2, and 50%-65% for Shahr_I_Soktha_BA3. And again, Shahr_I_Soktha_BA1 is at 60% for this sort of ancestry, but with 40% Iran_N and Iran_Chl) is at the extreme end of the "Mesolithic/Neolithic Near Eastern cline". For now, Natufians seem to be at the other end.

So, in theory, if one is to model other ancient Near Easterners using deeply diverged components, it is quite sensible to make use of Natufians and an IVC_West_Eurasian_Ghost for the uniquely "Near Eastern" portion, with WHG and ANE also present for the purposes of capturing any extra "Western Crown Eurasian" affinity.

For CHG, the model is suggesting that (in terms of very deep/basic ancestral streams) the Caucasus hunter-gatherers are around 80%-85% "eastern Near Easterner", but with a bit of a skew towards the western Near East and Europe.

Anthro Survey said...

@Matt

To which quantification do you refer in the Namazga/Anatolia paper? I do see ADMIXTURE output in the supplement and it does appear that Shamanka_LNBA are, indeed, about 90% Baikal_HG(Shamanka_EN). I also see that Shamanka_EN is used a sister clade to Han/as ENA proxy in modeling of Botai and Okunevo, but don't see any where HG is the object of modeling.

@Ryukendo

Check out the ADMIXTURE in the 137 genomes supplement. Even though they haven't used a proper agriculturalist population like the Han, we can probably infer that Xiognu(the ENA-rich sample) and modern Turkic populations are more shifted in this direction compared to Altai_IA and neolithic/bronze age Baikal populations. See K11 where ENA-rich ancestry begin to differentiate. The former mainly get a lot of blue, while the latter get the military green hue. Xionitic populations may have played an important role in medieval Turkic ethnogenesis, after all.

Anthro Survey said...

@Seinundzeit

Wow, so SiS2 actually has considerably MORE West Eurasian DEEP ancestry than Sarazm does in its non-ENA? It must have been a relatively early admixture event, then, and/or involving a different subset of ANE-rich HGs(unlike that for, say, GonurBA1_outlier) because it's apparently being consumed by the blue component here.

https://drive.google.com/file/d/1B_7mi39alC5TflmzAqu_fiALM3xRuOXQ/view

I wish that paper included the full ADMIXTURE output in the supplement so that this could be more conveniently visualized at lower Ks.

Good work and thanks for sharing this.

Unknown said...

@Seinundzeit
"I don't think you understand what I did there."

Guilty as charged!

"You see, the West Eurasian component unique to INP (again, 80%-90% for Shahr_I_Sokhta_BA2, 75%-85% for Gonur_BA2, and 50%-65% for Shahr_I_Soktha_BA3. And again, Shahr_I_Soktha_BA1 is at 60% for this sort of ancestry, but with 40% Iran_N and Iran_Chl) is at the extreme end of the "Mesolithic/Neolithic Near Eastern cline". For now, Natufians seem to be at the other end.

So, in theory, if one is to model other ancient Near Easterners using deeply diverged components, it is quite sensible to make use of Natufians and an IVC_West_Eurasian_Ghost for the uniquely "Near Eastern" portion, with WHG and ANE also present for the purposes of capturing any extra "Western Crown Eurasian" affinity."

I see... https://i.imgur.com/RugUOKG.jpg

"For CHG, the model is suggesting that (in terms of very deep/basic ancestral streams) the Caucasus hunter-gatherers are around 80%-85% "eastern Near Easterner", but with a bit of a skew towards the western Near East and Europe."

What can 'eastern Near Easterner' be? How east? Davidski said the current samples are too east.

ryukendo kendow said...

@ Stefan

From what you wrote, you probably have a somewhat shaky grasp of the concepts and the misunderstanding is at a level where it would probably take a high school biology course to correct (I'm not being insulting or anything, sorry if I come across in that way, just pointing out what I see). I recommend you download a high school biology textbook and read the section on genetics and heredity, as well as evo-bio and phylogenetics (looking at the index at the back may speed things up) and then come back to the "haplotype" and "meiosis" wiki. Or Khanacademy may help too.

Anthro Survey said...

@Ryukendo

Scratch what I wrote above!
After closer inspection, I can't make this conclusion from the ADMIXTURE graph at all(but it does appear to be the case that Xionitics were more closely related to ENA agriculturalists from DA45's monte analysis).
I missed a couple of things at K11 and also, for some reason, treated Glazkovo as Baikal_HG.

Davidski said...

@Rob

The only variation that this model allows, apart from different Steppe_MLBA groups instead of Srubnaya, is Tepecik_Ciftlik_N for Anatolia_BA.

That's it. Everything else is a fail.

But, of course, I welcome you or anyone else to prove me wrong and come up with a model that works better than mine.

Jaydeep said...

Sanuj,

Indeed. You may be interested to know that about 2 years ago there was an study of dog aDNA which included a sample from a Corded ware site in Germany.

Guess what, the dog dna had admixture from Indian/Iranian dog and wolf.

https://www.nature.com/articles/ncomms16082

We also see that the Armenia Chalcolithic samples have EHG which is missing in the Near East otherwise at that early stage. And its y-dna is L1a which is most common in South Asia.

No one is going to remember these facts because it does not suit their beliefs and narratives. But we ought to as this is part of accumulative evidence that is coming slowly but surely.

«Oldest ‹Older   201 – 266 of 266   Newer› Newest»