search this blog

Tuesday, August 11, 2015

Finngolians


Of course, Mongolians never made it to Finland or even Northern Russia (Kargopol area) where these Russian samples are from. How did this crap get through peer review?

The proportions of admixture from ancestral EUR and EAS [European and East Eurasian, respectively] were estimated, and are shown in Table 2. CEU populations mostly originating from France and Germany had a small fraction (0.7 +/- 0.8%) of genetic material from EAS. People from Great Britain such as British (GBR) and Orcadian inherited 2.5%–3.8% from ancestral EAS. Finnish (FIN) and Russians inherited significantly more genetic material (>12%) from ancestral EAS, which is consistent with their historical record of admixture with Mongolian populations. Besides, Adygei from Caucasus inherited 3.2 +/- 1.0% from ancestral EAS.

Pengfei Qin et al., Quantitating and Dating Recent Gene Flow between European and East Asian Populations, Scientific Reports 5, 02 April 2015, Article number: 9500, doi:10.1038/srep09500

See also...

Finngolians #2

54 comments:

Aram said...

What means gene flow of EAS at 89 % in Sindhi???

http://www.nature.com/srep/2015/150329/srep09500/fig_tab/srep09500_T2.html

Balochi even higher.

Davidski said...

It probably means these people are incompetent.

Kristiina said...

I refer again to the admixture chart in the Turkic ancestry –related paper as it is relevant here:
http://biorxiv.org/content/biorxiv/suppl/2014/07/30/005850.DC1/005850-1.pdf

It gives us indications on the distribution and amount of Mongolic ancestry in Central Asia and in the west.
Central Asia (in descending order)
Kalmyks (c. 80%)
Kyrgyz, Kazakhs, Uygurs
Karakalpaks, Hazara
Bashkirs
Uzbekhs, Nogais
Turkmens
Tajiks
Kumyks (c. 20%)
All other groups have only trace amounts:
Kabardians, Adyghe
Yaghnobi
Ossetians
Balkars, Azeris, Turks
Pathans
Chuvashes
Komis, Mordvines
Gagauzes

blogspot said...

'which is consistent with their historical record of admixture with Mongolian populations'
this sounds incompetent too.

KatharĂ³s said...

What about Lipka Tatars also known as Lithuanian Tatars, Polish Tatars.
https://en.wikipedia.org/wiki/Lipka_Tatars

Kristiina said...

I forgot to add Tatars. They would come after Tajiks; and also Russians and Karelians have tiny amounts of Mongolic ancestry.

In any case, Mongolic ancestry does not explain ENA in North Eurasia.

Onur Dincer said...

I forgot to add Tatars. They would come after Tajiks; and also Russians and Karelians have tiny amounts of Mongolic ancestry.

In any case, Mongolic ancestry does not explain ENA in North Eurasia.


Kristiina, by "Mongolic ancestry" do you mean Mongoloid, in other words, East Eurasian ancestry? If that is the case, what explains ENA in North Eurasia other than East Eurasian ancestry? I know of no other source.

Aram said...

Thanks Kristiina

Kalmyks being the leaders is consistent with the fact that they are the only Mongolian (not Turkish) family speakers in Europe.

Davidski said...

Lipka Tatars didn't exist during the early Mongolian expansion. They came later and kept to themselves because they were Muslims. Most moved to Crimea and Turkey during the 17th Century. They never went to Finland or Northern Russia.

The East Eurasian admixture in Finns is not from Mongolians or Tatars. It's from North Eurasian Uralics, and it didn't arrive in Finland 1000 years ago during the early Mongolian expansions, but probably during the Bronze Age.

This paper should never have been published.

KatharĂ³s said...

I forgot to mention the Huns.

Davidski said...

I already explained who it was and when.

Onur Dincer said...

I wholly agree with Davidski.

Kristiina said...

Onur, By Mongolic ancestry I mean recent Mongol-related ancestry as seen in the above mentioned admixture chart. In no way did I mean ENA.

It should be clear by now that several different ENA components have moved in Northern Eurasia since palaeolithic times. The most ancient ones are found in Ma-1 and Native Americans and another more recent ENA's were found in Mesolithic Swedes and the Karelian hunter gatherer. The Finnish ENA is related to these northern/arctic movements of people and their linguistic affiliation/identity remains undetermined.

Onur Dincer said...

Onur, By Mongolic ancestry I mean recent Mongol-related ancestry as seen in the above mentioned admixture chart. In no way did I mean ENA.

Which ADMIXTURE component do you mean specifically Kristiina?

Kristiina said...

Mongolic ancestry is marked with orange colour, but it is not dark orange that is found in Japanese. The Finnish one is violet and you cannot seriously claim that it is a component typically shared between groups speaking Uralic languages.

Anonymous said...

The paper didn't account for relative basal and WHG/ANE proportions, so it is completely worthless. I also laughed at this inaccuracy:

"CEU populations mostly originating from France and Germany"

Kristiina,

another more recent ENA's were found in Mesolithic Swedes

Are you referring to the presence of EDAR? ADMIXTURE doesn't detect any ENA in them, as far as I've seen.

Onur Dincer said...

Mongolic ancestry is marked with orange colour, but it is not dark orange that is found in Japanese. The Finnish one is violet and you cannot seriously claim that it is a component typically shared between groups speaking Uralic languages.

Kristiina, so you mean the light orange component seen in K=13 and K=14. I'd say that for West Eurasia and Central Asia that component is much more connected with Turkic ancestry than any Mongolic ancestry.

Shaikorth said...

@Onur

That ADMIXTURE run can't distinguish between the two, I've yet to see one that can. Buryats have the highest amounts of component in question, but Tuvans, Kalmyks and Mongols have almost as much. It's found all along both Turkic and Mongol migration routes.

Yakuts have their own component shared with Evenkis etc. instead of "orange" shared by Turks and Mongols because they seem to be assimilated Tungusics and Paleosiberians with little autosomal Turkic ancestry remaining. Similarly Gagauzes look rather indistinguishable from East Balkan peoples.

Kristiina said...

Onur, of course you can call it Turkic ancestry if you so wish. Moreover, the paper itself used the term Turkic. I used the word "Mongolic" because orange color peaks in Mongolic speaking groups.

Irrespective of the word we use, that component probably formed in South Siberia and Mongolia.

Onur Dincer said...

Kristiina, normally ADMIXTURE components are named after the population or area in which they peak. But because you used the word "ancestry" rather than "component", I misunderstood your statement and thought you were really referring to Mongolic ancestry rather than a component that peaks in Mongols. I am sorry for the misunderstanding on my part. We all should know that that component is connected overwhelmingly to Turkic ancestry in West Eurasia and Central Asia rather than to Mongolic ancestry.

Kristiina said...

Murglip, please take a look at this admixture run from Haak et al paper
http://biorxiv.org/content/biorxiv/suppl/2015/02/10/013433.DC1/013433-1.pdf

From K=3 to K=8, Motala HG's show c. 10 % of ENA, in the same way as Finns and Russians. At K=3 this component is yellow East Asian in all of them. At K=4, in Motala, it becomes pink Native American (NA) and in Finns yellow and pink, i.e. NA and Euarasian ENA. At K=5 and K=6,the same continues. At K=7, both Finns and Motala HG's show pink NA and red Beringian component. At K=8, the same continues. From K=9 upwards, Motala hg's become fully North European, while Finns retain a small amount of Arctic component up to K=20 which could mean that their admixture is more recent than in Motala.

Anonymous said...

Kristiina,

Murglip, please take a look at this admixture run from Haak et al paper
http://biorxiv.org/content/biorxiv/suppl/2015/02/10/013433.DC1/013433-1.pdf


Ah, thanks. I explain ADMIXTURE's error here in comment number 24 in this thread:

http://dienekes.blogspot.co.uk/2015/06/101-ancient-genomes-from-bronze-age.html

The error stems from ADMIXTURE mistaking Amerindians as pure East Asians, and therefore mistakes a large portion of the West Eurasian populations that contributed to Amerindians as also being East Eurasian. If you removed modern Amerindians from those ADMIXTURE runs (or entered ancient West Eurasians in calculators that recognised Amerindians as a fundamentally mixed population), the East Eurasian percentage of ancient West Eurasian HG's would be much lower. Even authors of papers frequently get confused by this phenomenon.

Kristiina said...

Be it East Eurasian or West Eurasian, that "apparent ENA" is a component autochtonous to Siberia and it is still found in North Eurasia. Perhaps, it is ultimately neither West Eurasian or East Eurasian but Central Eurasian.

However, if you claim that that that "apparent ENA" is West Eurasian and not East Eurasian, then you have to accept that Arctic people are also West Eurasian for that part of their ancestry. I do not think that it is logical to say that the same component in Motala is West Eurasian and in Eskimos East Eurasian.

Anonymous said...

Kristiina,

Be it East Eurasian or West Eurasian, that "apparent ENA" is a component autochtonous to Siberia and it is still found in North Eurasia. Perhaps, it is ultimately neither West Eurasian or East Eurasian but Central Eurasian.

We know that modern Amerindians have more than 0% West Eurasian genetic material. ADMIXTURE identifies them as having 0% West Eurasian genetic material (at k=3). Therefore, it logically follows that some West Eurasian genetic material is falsely identified as East Eurasian genetic material in those particular ADMIXTURE runs. So it also logically follows that some (possibly all) of the 'East Eurasian' in West Eurasian HGs at k=3 is very likely actually West Eurasian.

However, if you claim that that that "apparent ENA" is West Eurasian and not East Eurasian, then you have to accept that Arctic people are also West Eurasian for that part of their ancestry. I do not think that it is logical to say that the same component in Motala is West Eurasian and in Eskimos East Eurasian.

When a component originating from an unadmixed population occurs in another population, then -- assuming it is not noise -- it is genuine admixture. If a component originates in an admixed population (such as essentially all Arctic populations), then that component appearing in other populations could mean admixture, or it could mean that the second population shares genetics with just one or two of the several ancestries that constitute the admixed population (which ADMIXTURE considers pure). One easy example is Kalash. Non-Kalash who have the Kalash component don't actually have Kalash ancestry, nor do they necessarily have all of the components that make up the Kalash (ANI, ASI etc.). A simple ANE connection (without either ANI or ASI) can often forge a connection. Unless ancient West Eurasians have components that originate from much purer East Eurasian populations (e.g. Han, Koreans, Japanese etc.), then I don't consider the minor, inconsistent appearance of 'East Eurasian' components to be meaningful. I'd like to see an ADMIXTURE run of ancient samples without admixed-but-inbred populations, so that messy components don't have an opportunity to show up.



However, if you claim that that that "apparent ENA" is West Eurasian and not East Eurasian, then you have to accept that Arctic people are also West Eurasian for that part of their ancestry. I do not think that it is logical to say that the same component in Motala is West Eurasian and in Eskimos East Eurasian.

Onur Dincer said...

Kristiina,

ENA is a catch-all term used to define all the non-West Eurasian (i.e., non-Caucasoid) part of the Eurasian ancestry. So West Eurasian ancestry is not included in ENA.

Kristiina said...

Murglip, I do not think that we can ever find pure populations.

"We know that modern Amerindians have more than 0% West Eurasian genetic material. ADMIXTURE identifies them as having 0% West Eurasian genetic material (at k=3). Therefore, it logically follows that some West Eurasian genetic material is falsely identified as East Eurasian genetic material in those particular ADMIXTURE runs".

It may be like that because that part is neither West Asian or East Asian.

How can you know that just that part in Amerindians that is falsely identified as East Eurasian is the same West Eurasian component detected in Motala. If there was a population in Siberia/Beringia that became Native Americans and was an admixture between East and West, as you wish, why this mixed population could not have moved to the West but must have remained in pure West Eurasian form in Fennoscandia?

Anonymous said...

Kristiina,

Murglip, I do not think that we can ever find pure populations.

You can easily find populations that are lacking particular admixtures, or for them to have admixture levels so small as to be negligible.

It may be like that because that part is neither West Asian or East Asian.

ADMIXTURE demonstrably makes an error. I can explain it in any number of ways. For example, assuming West Eurasian is blue at k=3, and East Eurasian is yellow ay k=3. MA-1 is predominantly blue, Amerindians are completely yellow. Amerindians are part MA-1. Therefore, ADMIXTURE makes a mistake. There is no other conclusion to draw than that ADMIXTURE is wrong, or that Amerindians do not derive any of their ancestry from an MA-1-like population.

How can you know that just that part in Amerindians that is falsely identified as East Eurasian is the same West Eurasian component detected in Motala.

Easy: remove Amerindians. And I don't 'know'; I strongly suspect, based on logic.

If there was a population in Siberia/Beringia that became Native Americans and was an admixture between East and West, as you wish, why this mixed population could not have moved to the West but must have remained in pure West Eurasian form in Fennoscandia?

Because I've not seen strong evidence that SHGs are East Eurasian admixed. They have EDAR, and there's a vague reference to formal statistics (not supported by other papers, even from the same team) in the same paper as the EDAR was reported in. However, I've seen no signal in ADMIXTURE that suggests input from a bona fide East Asian source, just an obviously spurious 'Amerindian' signal, and a few other sporadically-occurring components that originate in Siberian populations that obviously have ANE-admixture in them.

Either Samara or Karelia (I don't remember offhand) and an Iron Age Hungarian have East Eurasian uniparental markers in addition to small, but clear and consistent East Eurasian ADMIXTURE signals, so the East Eurasian in them is fairly obviously real. SHGs, on the other hand, tend to have less Eastern (and other) noise than even La-Brana.

I'm indifferent to whether SHGs have East Eurasian ancestry, btw, so I'm not sure why you're insinuating bias on my part. I'm explaining an ADMIXTURE phenomenon that obviously confuses many people, including researchers.

Kristiina said...

Murglip, thank you for a good comment! I am a linguist so I do not really claim that I thoroughly understand how admixture runs or other calculations work, so I appreciate a response like the one you gave me.

I readily agree that the signal in Motala is very weak, but it is true that EDAR is intriguing.

As for that issue about unadmixed populations, I suppose that it is also a question of time perspective. By unadmixed populations you probably meant populations that have been isolated and mating among themselves for, let say, 500-1000 years. When I said that there are no unadmixed populations, I meant a longer time perspective of thousands of years. I thought that there are not any/many populations who would have remained unadmixed for many thousands of years (perhaps Onge). I suppose that from 10 000 years' perspective all of us are mixed.

”ADMIXTURE demonstrably makes an error. I can explain it in any number of ways. For example, assuming West Eurasian is blue at k=3, and East Eurasian is yellow ay k=3. MA-1 is predominantly blue, Amerindians are completely yellow. Amerindians are part MA-1. Therefore, ADMIXTURE makes a mistake.”

When I look at the admixture run in Haak et al, I see that at K=3, Northern Amerinds are 30% West Eurasian. Is it only because their recent European admixture has not been masked?

In any case, I find Northeurasian ANE/ENA an intriguing question and I would be happy, for example, to see ANE percentages of Arctic populations and an estimation of the West Eurasian part of it.

Anonymous said...

Kristiina,

Murglip, thank you for a good comment! I am a linguist so I do not really claim that I thoroughly understand how admixture runs or other calculations work, so I appreciate a response like the one you gave me.

No problem. I studied all available ADMIXTURE runs that contained Motala after it was revealed that they had EDAR. I couldn't really find any Eastern component that distinguished them from other European hunter gatherers (who, as far as we know, lacked EDAR). In fact, at certain k levels, ADMIXTURE identifies Scandinavian HGs as the most unadmixed HGs. That is, it gives them their own component, which they score 100% in, while other HGs are less than 100% of anything at the same k-level. Unless they are more inbred than other HGs, it seems weird that ADMIXTURE would give them their own component if they were significantly admixed. I'd like to see some further exploration of f3 and f4 statistics regarding the issue, but we'll probably have to wait for another Reich lab paper for that.

As for that issue about unadmixed populations, I suppose that it is also a question of time perspective. By unadmixed populations you probably meant populations that have been isolated and mating among themselves for, let say, 500-1000 years. When I said that there are no unadmixed populations, I meant a longer time perspective of thousands of years. I thought that there are not any/many populations who would have remained unadmixed for many thousands of years (perhaps Onge). I suppose that from 10 000 years' perspective all of us are mixed.

That actually neatly brings us back to this current paper (Pengfei Qin et al.). It makes a mess of its estimates for many reasons. Dienekes actually produced the same results as this paper three years ago. Basal Eurasian was unknown at the time. Now we know that it plays a major part in producing these kinds of statistics. We also know that the French (who are used as an ancestral European population in the paper) are themselves completely part of this pan-European phenomenon, so these statistics-relative-to-the-French are really meaningless, and don't even necessarily show East Asian admixture anyway. Where actual East Asian admixture is present (e.g. Russians), the percentage is exaggerated, because their score is a combination of actual East Asian admixture and HG ancestry relative to the French.

At the East Asian end, their 'European' scores may be related to ANE, or may be related to so as yet unknown internal East Eurasian phenomenon, maybe mirroring the basal/HG distinction among West Eurasians.

When I look at the admixture run in Haak et al, I see that at K=3, Northern Amerinds are 30% West Eurasian. Is it only because their recent European admixture has not been masked?

I think so. These were some of the first tribes to encounter Europeans colonists. I doubt they reflect key regional differences in pre-Columbian America. They contrast significantly with populations like Karitiana.

Alberto said...

Off topic: The full title of the talk about the Ancient Anatolians is:

Genome-wide data on 34 ancient Anatolians identifies the founding population of the European Neolithic. I. Lazaridis, D. Fernandes, N. Rohland, S. Mallick, K. Stewardson, S. Alpaslan, N. Patterson, R. Pinhasi*, D. Reich*.

So the rumours that they all look like EEF (like the Barcin sample) seem to be correct, but I guess it's restricted to early Neolithic samples (or up to Middle Neolithic).

http://www.ashg.org/2015meeting/pages/sessionlisting.shtml

Davidski said...

I've heard stuff about a good number of Neolithic samples from Anatolia now, some from deep in Anatolia, but not eastern Anatolia. They're all basically like Barcin.

So it'll be interesting to get word of what eastern Anatolian, or perhaps even better, Armenian and Iranian, Neolithic samples look like.

Hopefully someone's working on that even as we speak.

Alberto said...

Yes, it's also interesting that R.Pinhasi is also listed within the authors. It's great if the two teams are sharing data or even working together. The Pinhasi site has listed samples from Armenia and Caucasus from the Eneolithic, so probably it won't be long before we here from those ones too.

CroMagnon said...

I don't think this will answer the question of founding population for EEF. Without mesolithic data from anatolia, this study merely affirms that anatolia was in the path of the neolithic advance from further South east

Kristiina said...

The fact that Amerind (and Beringian populations) appear only East Eurasian in admixture runs could be related to their lack of the so called Basal and Near Eastern genetic material. Ust Ishim who was an ancient Central Eurasian was astonishingly far from modern near Easterners and close to all other non-African populations (http://s28.postimg.org/wuj44klpp/Kostenki14_Ust_Ishim_MA1ls.png).

With this I mean that there were very old Eurasians moving in northern Eurasia who reached Europe and who act as a bridge between modern Europeans and Native Americans, but Native Americans do not cluster with West Eurasians, because they lack Basal and all Neolithic near Eastern admixture which is very important in all modern West Eurasians.

If I have correctly understood the functioning of admixture runs, it is only natural that Native Americans appear unadmixed in world comparisons because they separated from all others so long ago (20 000-30 000 years ago) that they must by force form their own cluster.

By the way, does Malta1 or Ust Ishim have any Basal?

Karl_K said...

"By the way, does Malta1 or Ust Ishim have any Basal?"

No. But, Kostenki14 from (modern) Russia ~37,000 years ago does have it, or something similar, which is more basal than all other out of Africa modern humans.

It seems that most people are just dismissing this as a branch that completely died off, which is probably 99.9% true. It appears that maybe several smaller expansions out of North-East Africa or the Middle East were not so successful in the long run.

CroMagnon said...

I dont think that how people take it.
The analysis of K14 showed clear links to later -including modern - europeans (unlike Oase). Moreover, what exactly his 'basalness" was was disputed by the Haak/Reich team vis-a-viz Willerslev's .

Anonymous said...

Kristiina,

The fact that Amerind (and Beringian populations) appear only East Eurasian in admixture runs could be related to their lack of the so called Basal and Near Eastern genetic material.

This is what several researchers have stated, but it makes no sense, since MA-1, Loschbour, Motala etc. all lack basal admixture, yet ADMIXTURE considers them predominantly West Eurasian at k=3.

but Native Americans do not cluster with West Eurasians, because they lack Basal and all Neolithic near Eastern admixture which is very important in all modern West Eurasians.

In PCAs, Amerindians plot between West Eurasian HGs and modern East Asians (a proxy for the ancient East Asians whose genomes we currently lack), in proportion to the contribution of those two populations to Amerindian ancestry.

If I have correctly understood the functioning of admixture runs, it is only natural that Native Americans appear unadmixed in world comparisons because they separated from all others so long ago (20 000-30 000 years ago) that they must by force form their own cluster.

An ancestral population (MA-1) and a proxy for another ancestral population (modern east Asians) are both present in those runs, so ADMIXTURE has all the information it needs to represent the ancestry of Amerindians correctly. For some reason, it chooses not to.

Also, some ADMIXTURE runs show Amerindians to be mixed at k=3 (https://docs.google.com/spreadsheets/d/1FTCC6UyauUeBhPrA9aHzr7DLEnVq5q-wnTsfpe2a9Jg/edit?pli=1#gid=0), which agree much better with what formal statistics inform us to be the case.


Karl_K said...

@CroMagnon

"I dont think that how people take it.
The analysis of K14 showed clear links to later -including modern - europeans (unlike Oase). Moreover, what exactly his 'basalness" was was disputed by the Haak/Reich team vis-a-viz Willerslev's ."

Ok. This is all true. And I especially have a high level of belief that if Reich publically stated that this genome was likely contaminated with more modern DNA, then he isn't just talking shit.

Most scientists are very careful about dismissing another person's work openly, unless they have a very good scientific reason to be skeptical of the results. In this case, I'm sure Reich has directly observed what contamination of genomes does to the statistics.

However, if we take it the available genome at face value, then although this ancient man was more related to Western Eurasians, he also had some kind of admixture from a very basal, but not African, population. This component does not show up in later populations. So, it must have died out.

The only alternative is that it later formed a portion of the Early Farmer population which expanded from the Middle East and Anatolia.

Kristiina said...

In that Excel file Athabaskans are 68.9% European, which should mean that the studied Native American groups are unmasked. In that case, I do not know how reliable the results are.

In the admixture run of Haak et al, Ma1 seems to be c. 40% East Eurasian at K=3, but half of it seems to be ASI-related.

Let's imagine a situation in which Ma1 is mixed between a more European group and a Central Eurasian group, while Native Americans are a mixture between a Central Eurasian group (without any European input) and an East-Asian/(Palaeo Melanesian) group. In that case, Ma1 would logically choose the same Western Eurasian branch as populations that carry WHG such as Loschbour and Motala while Native Americans would choose the East Asian branch. In the admixture trees a population can only choose one branch on which to sit.

Of course, many of you have explained that Ma1 was not admixed but I do not know if that holds any more. In any case, I think that in the absence of any Central Asian or East Asian ancient autosomal data, we cannot be sure of that.

I still think that because Native Americans have formed c. 20 kya ago and have been separated from all others for a longer time than any other population in the world, it is natural that they form their own clusters and do not share clusters with modern Europeans.

Unknown said...

@Davidski
"I've heard stuff about a good number of Neolithic samples from Anatolia now, some from deep in Anatolia, but not eastern Anatolia. They're all basically like Barcin."

Do you know whether any of the new Anatolian samples from Pinhasi and Reich are from SE Anatolia--Upper Euphrates/PPNB context or are they all from Central and Western Anatolia?

Davidski said...

I don't know where Reich's samples are from, but I've only heard stuff about Western and Central Anatolian samples.

Anonymous said...

Kristiina,

In that Excel file Athabaskans are 68.9% European, which should mean that the studied Native American groups are unmasked. In that case, I do not know how reliable the results are.

Karitiana don't need to be masked, because they are unadmixed. Also, you can look at k=4 to see whether populations are 100% Amerindian or not. Populations that are 100% Amerindian at k=4 are still West-East Eurasian admixed at k=3.



andrew said...

In fairness to the authors of the paper in the OP, I think that it is quite likely that this is an English as a second language issue rather than a deep flaw in the paper.

It is easy for a non-Native speaker to confuse Mongolian which is a proper ethnonym, and Mongoloid, which is a description of a phenotype, and further for a non-Native speaker to confuse the Mongoloid phenotype with the EAS genetic component. These fairly subtle mistakes in translating their ideas to the English language (the surname of the lead author suggests a Chinese authorship) seem more probable to me than a genuine mistaken belief that the Mongolian empire of the 13th-14th centuries actually extended to Northeast Russia and Finland, unless you have access to the full text of the paper and the body text is actually historically inaccurate in this respect.

Also, it has been my experience that Chinese authors are not a sensitive to fine distinctions like that between the phenotype usually associated with people who have a particular genotype and the genotype itself, and between the type example of a phenotype and the phenotype that the type example illustrates, as Americans are.

I suspect that some of the extreme care that is used by Americans in discussing these distinctions in academia flows from the very prickly business that Americans deal with on a daily basis in fields where one has to talk about race-like subjects in a country where race is an extremely sensitive matter and certain ways of talking about race are taboo in educated academic company.

While the ethnic issues between the Chinese Han majority and the minority populations of China are sensitive political issues there, the way that these delicate issues are discussed in China flows not primarily from resort to academia sourced scientific and technical precision that makes fine distinctions and erects taboos were being imprecise, but instead is managed as a subset of the Chinese communist party propaganda double speak that pervades all aspects of Chinese life.

Also, Chinese scholars just plain tend to be somewhat less accomplished in their English language skills relative to their scholarly ability because unlike scholars in European countries, Japan, and former British colonies, they aren't generally taught English as a second language starting in elementary school by reasonable competent instructors. Mandarin is such a widely spoken language (including many expatriot communities), and foreign travel is sufficiently uncommon for the Chinese, that the Chinese, like the Americans whose language is also very widely spoken and who also don't overwhelmingly engage in foreign travel, don't prioritize good quality foreign language learning at a young age in their educational system.

andrew said...

On second thought, after reading the body text, I reach a somewhat different conclusion. While the historical context is shallow and not very well thought out, put in more as window dressing than as genuine analysis, it is also the case that if you read the body text carefully, it never actually states that Finns and Russians have a "historical record of admixture with Mongolian populations.", in any place but the quoted language of the abstract, although the abstract writer can be forgiven for trying to connect the dots because the body text doesn't discuss in any meaningful way where the Finns and Russians actually do get their EAS from.

So, it actually looks as if this is more of a case of statistics nerds who have only shallow world history and anthropology knowledge being just careful enough not to say anything untrue in the body text and then having the group assign the task of writing the abstract to the uber-nerd who did most of the quant work rather than to the team member who supplied what thin historical and anthropological gloss there is in the paper and having him screw it up. Because the error is so narrow, the simplest way for it to be corrected would have simply been for them to strike the second half of the offending sentence beginning with the word "which" and leave the rest of the paper unchanged and not actually wrong.

The paper's historically and contextual analysis is still dreadfully shallow, and the paper really doesn't even make any meaningful conclusions about their data other than reporting it descriptively. Honestly, I've seen better analysis in amateur blog posts. But, when push comes to shove, its probably better to publish the undigested statistical analysis and data description is unimpressive an effort as it is, than it is to deprive the discipline of that labor (under the theory that the best is the enemy of the good).

capra internetensis said...

@andrew

They dated the East Asian admixture in Russians and Finns to ~1300 and 1900 years ago, and in the previous paragraph dated the Mongolian expansion to the 13th century. So yeah, I'm pretty sure they didn't literally mean Mongols.

There was no shortage of Asian nomads running around invading everyone in the first millennium, but I can't think of any that ended up in Finland. Unless it was the proto-Finns themselves, with more European ancestry, mixing with the proto-Saami who had more (but still minority) East Eurasian? Dubious.

Kristiina said...

For the sake of curiosity, I checked the admixture results of Anzick and Kennewick man, as they are surely free of European admixture. In the Kennewick man paper, at the upper level, Kennewick man does not show any CEU, and Anzick maybe 1%. The main components are the following: Anzick c. 35% Central American, 25% Quechua, 15% Northern Amerind, 8% Karitiana, 5% Cabecar, 5% Pima, 3%Surui. Kennewick man is not very different but he is more north American with c. 40% Northern Amerind and, unlike Anzick, he has 5% Koryak and 5% Nganasan. Surprisingly, he seems to be more ”Caucasoid” because he is more Arctic. Quite funny indeed.

Then, let’s have a look at the Anzick paper: Anzick is 75% Central/South Amerind, excluding Karitiana, 10% Arizona and the rest was Karitiana and Beringia. The only non-American fragment is a tiny bit of South Asian. At K=3, Anzick is perhaps 1% West Eurasian and 4% East Asian. At K=9, West Eurasian becomes Papuan, and East Asian Beringian. AT K=11, West Eurasian becomes South Asian.

To sum up, I am not convinced with the evidence we currently have that Native Americans are in any way European, let alone West Asian. I would say that their deep roots are in Central Asia and East Asia, and the North Eurasian ENA stems from that source.


Nick Patterson (Broad) said...

@Cromagnon wrote

< I don't think this will answer the question of founding population
< for EEF. Without mesolithic data from anatolia, this study merely
< affirms that anatolia was in the path of the neolithic advance
< from further South east


That's kinda right but there likely will never be a definitive
answer. We derive an ancient EEF population which itself is going
to be an approximation, probably with internal structure that we
may not be able to detect with present data. This "EEF" formed
(approximately) from other ancient populations, probably more than
one, and so it goes back in time.

The final answer will never be reached.
~

Kristiina said...

Capra Internetensis, I also took note of the admixture dates. If I try to match it with the history, I would suggest that the proposed admixture happened when the Corded Ware related cultures / the subsequent Bronze Age cultures merged with the inland cultures which may have happened over a longer period of time. Corded ware people and the subsequent costal populations were probably like Balts while the inland people were more Saami like, and it is known that Saamis have a higher percentage of Siberian (15-20%) than Finns while Balts lack it. Finns are probably a result of this merger of costal and inland cultures.

Unknown said...

If you limit your Native American samples in admixture, they lose their own component. That's when they split up.

CroMagnon said...

Dear Nick
Thanks for reply
Perhaps to rephrase my statement :
I would be curious what Anatolia looked like before the neolithic / Pottery period. Did it have this "Basal eurasian" already, or was it merely the southeastern -most "outpost" of an essentially "WHG" population?

I think this answer is achievable

This will moreover tell us about the broader picture : should anatolia be considered as
part of the original "Neolithic core" ?

CroMagnon said...

@ murglip

"Scandinavian HGs as the most unadmixed HGs"

Unless I've musunderstood what you're getting at; I find that categorically untrue. Aside from what we already know about the archaeology of post-glacial europe, look at their mtDNA profle-which speaks of Eastern and western european contributions. This is "admixed" , right ?

Anonymous said...

CroMagnon,

Unless I've musunderstood what you're getting at; I find that categorically untrue.

Yep, you misunderstood, but I'm not sure how. The full sentence is:

"In fact, at certain k levels, ADMIXTURE identifies Scandinavian HGs as the most unadmixed HGs."

And in the very next sentence I explain further:

"That is, it gives them their own component, which they score 100% in, while other HGs are less than 100% of anything at the same k-level"

CroMagnon said...

Got you

Fanty said...

Doesnt some study claim the scandinavian HGs beeing all incestoid clones originating from a extremely tiny group? Isnt it totaly normal then, to form a cluster? Doesnt families create their own cluster in admixture if you use more than 1?