search this blog

Friday, December 1, 2017

Descendants of ancient European (fair?) maidens in Central Asia's highlands


Several South Central Asian populations have a reputation for producing individuals who look surprisingly European, even the lighter shade sort of European from Eastern and Northern Europe. This is especially true of the Pamiri Tajiks, and that's unlikely to be a coincidence, because these people probably do harbor a lot of ancient Eastern European ancestry.

My own estimates, using various ancestry modeling methods, suggest that Pamiri Tajiks derive ~50% of their genome-wide genetic ancestry from populations closely related to, and probably derived from, Eneolithic/Early Bronze Age pastoralists from the Pontic-Caspian steppe of Eastern Europe, such as the Sredny Stog and Yamnaya peoples. Below is a simple Admixture graph using the mostly Yamnaya-derived Iron Age Sarmatians from Pokrovka, Russia, in far Eastern Europe, to illustrate the point. Note that Sarmatians were East Iranic-speakers, which is what Pamiri Tajiks are. The relevant graph file is available here.


But, some of you might retort, this is all just statistical smoke and mirrors, and what it really shows is that these so called Europeans came from Central Asia or even India.

Not so, because my models can't be twisted any which way, and they have strong support from uniparental marker data.

Many South Central Asian groups, and especially Indo-European-speakers, like the Tajiks, show moderate to high frequencies of two Y-chromosome haplogroups typical of Bronze Age Eastern Europeans: R1a-M417 and R1b-M269. This is old news to the regular visitors here and its implications are obvious, so if you still think that these haplogroups expanded from South Central Asia to Eastern Europe, rather than the other way around, then please update yourself (for some pointers, see here and here).

And now, courtesy of Peng et al. 2017, we also have a much better understanding of ancient European influence on the maternal gene pool of Pamiri groups (see here). The paper doesn't specifically cover the topic of European admixture in South Central Asia, but it nevertheless demonstrates it unequivocally.

Below are a couple of phylogenetic trees from the paper featuring a wide range of mitochondrial DNA (mtDNA) sequences shared between Europeans and Central and South Asians; quite a few of these lineages are rooted in Eastern Europe, as shown by both modern-day and ancient DNA, so they strongly imply gene flow, and indeed considerable maternal gene flow, from Eastern Europe deep into Asia.


Worthy of note are the lineages belonging to such relatively young (likely post-Neolithic) haplogroups as U5a1a1, U5a1d2b, U5a2a1, and U5b2a1, all of which have already been found in ancient remains from the Pontic-Caspian steppe.

I'm no longer wondering whether there were massive population movements from Eastern Europe to South Central Asia during the metal ages. It's a given that they happened, and I'm now looking forward to learning about the details from ancient DNA. For instance, what was the ratio of men to women amongst these migrants? And how fair were they exactly?

See also...

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

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

R1a-rich ancient Siberians may have been as blond as today's Northern Europeans

228 comments:

1 – 200 of 228   Newer›   Newest»
EastPole said...

David, Tajik language is a Persian language and not Sarmatian/Scythian.

https://upload.wikimedia.org/wikipedia/commons/e/e0/Iranian_Family_Tree_v2.0.png

Sarmatian/Scythian languages are not known and it is assumed that they are similar to Ossetian.

Razib Khan said...

what you say may be true of pamiris re: gene flow. but obviously "european-like appearance" for highland south asian like appearance is not ipso facto evidence of elevated e. european gene flow. eg if you google "burusho people" you can observe that most of these people would fit in in the balkans, and some of them are blonde. but they also have a uniform 10% east asian admixture and don't seem as 'indo-aryan' as the kalash (which makes sense since they are a linguistic isolate).

also there is unplished stuff on the brokpa people which indicates descent mostly from west asians (mtDNA) and they look are very fair and promote descent from alexander's soldiers for tourism.

Davidski said...

@EastPole

Pretty sure that Pamiri is an East Iranian language group, while Persian West Iranian.

@Razib

Yep, agree that the situation is complex in SC Asia, and correlations between Euro ancestry and looks not always strong.

supernord said...

Z-value = -1 is small in absolute value, the values at the arrows to D8 look implausible.

Ryan said...

Am I correct in saying that D4 and D5 have no drift specific to them, and that they differ from each other only in their proportions of D2 and B? IE - that this model isn't very able to distinguish between CHG and Iran Chalcolithic?

Samuel Andrews said...

"Worthy of note are the lineages belonging to such relatively young (likely post-Neolithic) haplogroups as U5a1a1, U5a1d2b, U5a2a1, and U5b2a1, all of which have already been found in ancient remains from the Pontic-Caspian steppe."

Exactly. There's definitely more coming from Pamir's Steppe ancestry than U5 lineages.

I'd be done with a pretty serious study of haplotype-sharing, mitogenomes, and age estimates if not for school work. But once I am done, I think I'll be able to detect them. Some of it obviously, like U5a1d2b, won't be found in most of Europe because all Bronze age Steppe groups weren't identical to each other.

EastPole said...

@Davidski
“Pretty sure that Pamiri is an East Iranian language group, while Persian West Iranian”.

Yes, you are right, my mistake.
But there is some confusion about their exact classification some put it with Pashto and Saka as Southeastern and not with Northestern Ossetians and Sarmatians. In this article they link Pamir with Saka Scythians derived from Andronovo and not Sarmatians:

“Around 4000 years before present, the migrations of Andronovo Culture reached the Pamirs. And then, the Scythians (also known as Saka and Sai), the speakers of Indo-Iranian languages, began to dominate this plateau [7]. It indicates a wave of eastward expansion of the Indo-European language family [7]. Although the Scythians are evident in archeological and historical records [6], their population history remains unclear. Nowadays, the Pamirs are the homelands to the highland Tajiks [4]. Both linguistic and physical anthropological evidence suggest the highland Tajiks likely being the descendants of ancient Scythians [7, 8].”

If they came from Andronovo then there is a link with Corded Ware culture because Andronovo came from Sinashta culture which was influenced by eastern CWC. It is very interesting because H5a1 with mutation 15833 is present in Pamiri Tajiks, Lowland Kyrgyz, and can also be found in India and in Pakistana:

https://s7.postimg.org/x4km7eiy3/screenshot_310.png

On the other hand H5a1 originated somewhere around Poland at the time when CWC was formed there and started to expand east. It was found in Corded Ware Estonian sample:

http://eurogenes.blogspot.com/2017/03/baltic-corded-ware-rich-in-r1a-z645.html?showComment=1488537435683#c2361904565646502138

I am looking for link between Slavs and Indo-Iranians because they show many similarities in language and religion. If that link is 5000-4000 years old as the case of H5a1 seems to suggest (which is in line with time depth of R1a-Z645 split into R1a-Z93 and R1a-Z283) then it will be very important for the history of our languages and religions.

Shahanshah of Persia said...

@Davidski

Yaghnobis and the Tajiks of China are more Steppe derived than the Pamiris, aren't they? Why haven't you mentioned them here? People know that many of these peoples exist.

On a related note, how Steppe derived are the Nuristanis?

Razib Khan said...

the whole region used to be dominated by sogdian. an east iranian language. the intrusion of dari persian is in the last 1,000 years....

Sofia Aurora said...

@EastPole

Sarmatian and Scythian languages belong to the Northern Iranian peoples and to the Eastern Iranian language group.
They are known from the Tarim basin and especially from the cities of Karga, Khotan, etc.
The Khotanic or simply Khotan language is a very well known language of the Saka-Sarmatian group and it is well documented from texts of the Tarim area where together with the Tokharian were the only Indoeuropean languages that are known to have reached so deep in East Asia. Thus it is not only the Ossetian that descends from Sarmatian.
Additionaly the Alans, The Scythians of Bessarabia and the Saka of Hyrkania (Varkana) have left words in the various substrata of languages such as the ancient Persian, the Slavic and various Turkic languages.

Sofia Aurora said...

@Davidski

The Pamir mountains were known for the presence of the Pamir-Ferghana anthropological type.

Although that type was by no means restricted to the area (quite the opposite actually! it can be found almost everywhere in western Asia, Central Asia, Eastern Asia etc. with the exception of the insular Southern Asia. Also in Europe it was present in eastern countries and southeastern ones) nowadays it survives "en masse" in the Pamir mountains foetalized.
It has lost its robustness, diminished in Height and is less curvoccipital than the past.

Thus because of these alterations even if we observe the type nowadays i don't know how much can we say about its past phenotype.
I mean, yes, probably the soft parts of the skull were similar but how can we be sure about its color complexions?

If i remember well, Khokhlov in 2004, had made a huge research in the Eastern European area covering a space from Ukraine to Trans-Urals and found that in Eneolithic and Bronze age times there was a third proto-Europoid type living from Lower Volga to Trans-Urals and was almost identical with the Pamir-Ferghana type. The other two types were Vasilievka I and III. He had found also another type which was Europoid and considered it intrusive, the Voloskoye type, which almost disappeared with the domination of the Vasilievka III type at the end of Mesolithic and survived in small numbers only to Crimea and totally gone extinct during the Neolithic era.

The thing is that Khokhlov considered the Vasilievka and the Lower Volga types indigenous. But if the Lower Volga type was almost identical to the Pamir-Ferghana type then can we dare say that it appeared in the eastern Asian places from the Pontic- Caspian steppe?

And if it is so did it have the same color complexions such as them?
Or is it that before the disolving of the ancient Glacial Ob river the ancestral form of the Pamir-Ferghana type splintered in various locations and acquired different non-metric morphs while the basic skeletal form remained the same?
If we could answer that it could have helped a great deal!

Chad Rohlfsen said...

What?!? H5a1 from Poland at the time of Corded Ware? Wha??? Based on what? You need to go check on that formation date and the spread of this group. It's much older.

andrew said...

I put these results in context, consider some historical and linguistic context, and do some analysis is a post at my blog: https://dispatchesfromturtleisland.blogspot.com/2017/12/southern-central-asians-have-about-50.html

aniasi said...

@Davidski

You've always been on point, but do you really need to entertain the sophomoric 'phenotype' commentators here? It dilutes arguments of real genetic ancestry, and honestly plays into some really dated notions of race. Boas's study already showed how quickly physical changes can happen due to a change in environment, and sexual selection. I've seen Berbers with light hair and eyes, but they have very little steppe ancestry, and we still don't know where the West Eurasian light skin gene arose. Is it necessary to go back to the 1930s with "blonde haired blue eyed people are sign of ancient Aryan European presence" line?

Davidski said...

@supernord

Can't see anything implausible there. You're just taking the model too literally.

@Shahanshah

Can't say. I only have one Yaghnobi in this dataset and no Nuristanis.

Davidski said...

@aniasi

Well, if some people have outdated notions about pigmentation and ancestry, then let's update and correct them.

I'm pretty sure that the scientists who sequence and analyze the ancient samples for the upcoming paper(s) on South and Central Asia will cover the topics of pigmentation and height in some detail, like they did for ancient Europeans. Are you going to criticize them for doing so?

Arkaim said...

@aniasi
Just passing to say that the likes of Boas and Cavalli-Sforza already died in their fields along with their writings, and people are passing through them as they now know that they were wrong. Similar to how only know people are rejecting Freud and using Jung in the psychologies.
Outdated models and fraudulent research has been exposed in modern times, don't idolize liars and ideologues.

Davidski said...

Calm down. Cavalli-Sforza was the precursor to all of this.

Samuel Andrews said...

@aniasi,
"we still don't know where the West Eurasian light skin gene arose."

Let's be honest, when people talk about 'West Eurasian' light skin they're really talking about European light skin.

Seinundzeit said...

Shahanshah,

For what it's worth, the Nuristani seem to be identical to the Kalasha (in the context of ADMIXTURE; I've seen some results), so we can assume somewhere around 40%-50% Steppe_EMBA/Srubnaya_outlier/Potopovka_Outlier-related ancestry.

David,

"And how fair were they?"

This is a very interesting question, because I don't think that we have a complete handle on pigmentation variation amongst Steppe_EMBA-related populations.

Looking at the genotype data, and based on what I see IRL, nearly all South Central Asian populations (talking about the average) are much, much fairer than what one would expect for Steppe_EMBA + Iran_N/Iran_Chal mixtures.

For example, looking at genotype data, the Kalash are moderately fairer than currently sampled Yamnaya, and vastly more fair in comparison to Iran_N (and can confirm from personal experience; the Kalasha near Chitral are mostly what one could call "brunet white" and light beige, with a very strong/noticable minority of blonde people who could pass in northern Europe).

So, unless it's a matter of selection, there must have been Yamnaya-like populations that were much fairer than the samples that we currently have (would be interesting to know more about how Srubnaya_Outlier stacks up when it comes to pigmentation genetics).

But if all Steppe_EMBA were as dark as Yamnaya, it will turn out to be a matter of selection.

Can't wait to see those Swat valley genomes.

Mike the Jedi said...

@Dave

Regarding your qpGraph for the Tajik Pamiris, is D8 supposed to represent EBA steppe? If so, I am surprised to see its D4 (CHG-like) ancestry at 72% while the EHG-related D3 is only 28%. Isn't CHG ancestry in Yamnaya around 40% and the rest EHG? What am I missing?

@Razib
"the brokpa people... look are very fair and promote descent from alexander's soldiers for tourism."

Yeah, that ol' chestnut has always burned my ass. Pinning higher rates of light pigmentation on the Greeks is one of the stupidest things I've ever heard. What else is new, though? The only time you get to see Mediterraneans depicted realistically in pop culture is in mafia films or My Big Fat Greek Wedding. The rest of the time it's Brad Pitt in a Corinthian helmet.

And I think you're right about tourism being to blame for the myth's continued promotion. Everybody's heard of the ancient Greeks and Alexander, but only the educated are familiar with Scythians, Sarmatians, and the like. If you mention the Greeks, images of skirted hoplites and marble temples are immediately conjured. The steppe? Nothing but grass. One popular science writer (that Dave clowned recently) was so unfamiliar with the steppe that he even called it a hilly part of Asia. I mean-- what the fuck? And that guy is a journalist, so the common people don't have a prayer. Marketing is always going to use what the laypeople can be counted on to know. Maybe these Hindu Kush and Pamir mountain people will honor their actual ancestors one day and leave us Greeks the hell out of it.

@Aniasi
"Boas's study already showed how quickly physical changes can happen due to a change in environment, and sexual selection"

Sure, but some things haven't changed much. I'm pretty sure the Han were also brachycephalic in the Bronze Age. I'd wager the Andeans had hawk noses and high altitude breathing adaptations long before the Moche.

There's nothing wrong with anthropometry; it's just been relied on too zealously in the past (when nobody knew anything about DNA). Unfortunately, metric work isn't being done with the same enthusiasm as it was in the old days. I think anthropologists are afraid of getting the calipers out again lest they be called Nazis. Too bad, maybe the genomics revolution will inspire more widespread interest in physical variation.

People living in and near the Pamirs are apparently very brachycephalic on average. I can't help but wonder if this was also true of their ancient ancestors or if its the result of more recent selection (like in Central Europe). Their East Asian ancestry is minor so I doubt that has much to do with it, but maybe?

@Seinundzeit
"Looking at the genotype data, and based on what I see IRL, nearly all South Central Asian populations (talking about the average) are much, much fairer than what one would expect for Steppe_EMBA + Iran_N/Iran_Chal mixtures."

I wonder if inbreeding is a factor.

I've seen Samaritans cluster with both Levantines and Arabians, but they have a much higher rate of light hair/eye (maybe even skin) pigmentation than one would expect given such a robustly Near Eastern people. I suspect it might be because they are so inbred. In the 1920s they numbered less than 200 people.

Mike the Jedi said...

^ Oh, in case I get some smart-assed comment, I know the Han didn't exist in the Bronze Age; I'm talking about their ancestors obviously.

Ebizur said...

Is there any reason why we should not call the East Iranic-speaking populations of the Pamir Mountains and their immediate vicinity "Pamiris" as opposed to "Tajiks" or "Pamiri Tajiks," reserving the term "Tajiks" to refer to native speakers of Persian originating from areas east and northeast of Iran? Would that be offensive to ethnic Kyrgyz who also inhabit some parts of the Pamirs?

(Of course, in reality, a great part or even the majority of the ancestors of many Persian-speaking "Tajiks" born in the vicinity of the Pamirs may have been "Pamiris" until quite recently in history, and, conversely, many "Pamiris" may have some ancestors who have arrived to the region recently as Persian-speaking immigrants, but it seems like a convenient way of making an ethnolinguistic distinction for the sake of avoiding ambiguity.)

Davidski said...

@Mike

What am I missing?

Sarmatians are Yamnaya plus extra CHG, some ENA, and probably a bit of Iran_N and EEF. So modeling them as only EHG-related and CHG-related, plus Han, means the algorithm has to compensate. This might help...

http://eurogenes.blogspot.com.au/2017/06/iron-age-nomads-vs-bronze-age-herders.html

In other words, there's some hidden stuff in that 72% CHG-related edge, but overall the model works fine, despite being a major simplification.

Seinundzeit said...

Mike,

"I wonder if inbreeding is a factor?"

Although this could apply to the Kalasha, it wouldn't really work in the case of the Burusho, the Pashayi, the Tirahi, the Chitrali, the Nuristani peoples, the Kunar and Bajaur Pashtuns, the Panjsheri Tajiks, the Pamiri peoples, etc.

Also, for what it's worth, many Kalasha exhibit considerable unease about the Greek/Albanian/Macedonian stuff; such notions are mostly propagated by Western journalists (and are often backed, with considerable energy, by ethnic Greeks/Albanians/Macedonians).

Anyway, talk of "Aryans" is considered unseemly in our time; were that not the case, I'd bet that the "Aryan" angle would be played far more often (at least it would be factually correct).

Gaspar said...

I recall russian studies of 2010 and 2013 stating that proto-bulgarians originate from the Pamiris people

Chetan Vit said...

@Davidsky Any news about when the South Asian paper (or any important papers) will be out?

Davidski said...

@Chetan

No idea what's happening with the South Asian paper. I guess it's coming soonish. That could mean next year.

Salden said...

Davidski have you heard or read anything about Ancient DNA studies for either Predynastic to First Intermediate Egypt or Prehistoric to Bronze Age Mespotamia?

Davidski said...

Davidski have you heard or read anything about Ancient DNA studies for either Predynastic to First Intermediate Egypt or Prehistoric to Bronze Age Mespotamia?

Nope.

kony1_1 said...

@Mike

The steppe newcommers to Ireland were swarthy, but todays Irish are the fairiest people in the world.
This theme repeats everywhere.
The steppe ancestry brings a lowish frequency of light skin genes to an area, which gets driven to up to 90s of percents afterwards.

Rob said...

The Bulgars formed in the Don region when Kubrat rebelled against the Avars and carved out a following of post-Hun and other EE peopes; not in the mythical Pamirs

EastPole said...

@Chad Rohlfsen
“What?!? H5a1 from Poland at the time of Corded Ware? Wha??? Based on what? You need to go check on that formation date and the spread of this group. It's much older.”

“Recent studies on mtDNA hg H5 have revealed that phylogenetically older subbranches, H5a3, H5a4 and H5e, are observed primarily in modern populations from southern Europe, while the younger ones, including H5a1 that was found among RoIA individuals in our study, date to around 4.000 years ago (kya) and are found predominantly among Slavic populations of Central and East Europe, including contemporary Poles”.

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0110839

H5a1 with mutation 15833 was found in Estonian CWC Sope sample 4,575–4,350 BP.

It looks like H5a1 ancestors came with Neolithic Farmers from southern Europe but H5a1 with mutation 15833 expanded with CWC from Poland. Those samples H5a1 with mutation 15833 found in Pamiri Tajiks or in India and Pakistan probably came with Andronovo.
If you have a better theory, I would be glad to hear it!

Chad Rohlfsen said...

You're not understanding. H5a1 is too old to have expanded with Corded Ware. It just happens to be another European haplogroup that ended up in South Asia.

EastPole said...

@Chad
You have not convinced me because you didn’t present any evidence. In which culture older than CWC by several thousand years was H5a1 found? It is not common in southern Europe where ancestral clades are found. It was found in Estonian CWC Sope sample 4,575–4,350 BP, in Roman Iron Age Poland and presently is found predominantly among Slavic populations of Central and East Europe. Fits CWC and R1a expansion from Poland.
Indo-Iranian languages including Pamiri, where H5a1 was found and where R1a-Z93 is common, are closely related to Balto-Slavic languages. Indo-Iranian R1a-Z93 like Balto-Slavic R1a-Z283 originated from R1a-Z645 which was found in CWC and expanded with it.
Theory that H5a1 arrived to Pamir from Andronvo and CWC seems reasonable to me.
Look, we are not looking for absolute knowledge but present theories and choose the best one. I don’t buy theory that Tocharians were Celts or Germanics or similar crap based on pseudoscientific speculations.

Chetan Vit said...

@Everyone Do we have any ancient genomes from the eastern CWC (like Fatyanovo-Balanovo or Abashevo)? What are the odds that these cultures would have been Uralic speaking?

Davidski said...

@Chetan

Do we have any ancient genomes from the eastern CWC (like Fatyanovo-Balanovo or Abashevo)? What are the odds that these cultures would have been Uralic speaking?

Abashevo was part of the early chariot complex and closely related to Sintashta. You think early Uralics resembled Sintashta and drove chariots?

Rob said...

Sintashta could have precipitated the expansion of FU around it's periphery

Davidski said...

Abashevo and Sintashta had nothing to do whatsoever with Uralics.

Chetan Vit said...

@Davidsky Parpola seems to think Abashevo and Fatyanovo were early Uralic cultures. Since the chariots only appear from Sintashta onwards, maybe the Abashevo were assimilated by the Indo-Iranians from Poltavka. That would explain the shared vocabulary/ loanwords between IIr and Uralic. Also I remember they found two Andronovo individuals with Asiatic features.

Davidski said...

You mean the two Andronovo individuals with Asiatic features in South Siberia?

Yes, surprising, considering the location.

Chetan Vit said...

*two haplogroup c individuals in Andronovo?

Davidski said...

Parpola also thinks that Cucuteni-Tripolye was Late PIE and migrated to India. Guess he didn't bother to wait for aDNA from Ukraine before writing his book.

Chetan Vit said...

@Davidski Parpola did make some mistakes but a lot of his theories are turning out to be true.

Davidski said...

But let's be reasonable, a couple of his main theories, that Cucuteni-Tripolye people migrated to India and that early Uralics were Abashevo people who drove chariots, clearly aren't working out.

Otherwise we'd be seeing a lot of Y-hg I2 and EEF ancestry in India, and a lot of Y-hg R1a-Z93 among Uralics. And that's a negative on both counts.

Rob said...

"Abashevo and Sintashta had nothing to do whatsoever with Uralics."

^ I laughs in PJW

Davidski said...

@Rob

Sintashta could have precipitated the expansion of FU around it's periphery

Awesome theory. You were there are saw this happening?

Matt said...

seinundzeit: "Looking at the genotype data, and based on what I see IRL, nearly all South Central Asian populations (talking about the average) are much, much fairer than what one would expect for Steppe_EMBA + Iran_N/Iran_Chal mixtures."

kony: The steppe newcommers to Ireland were swarthy, but todays Irish are the fairiest people in the world. This theme repeats everywhere. The steppe ancestry brings a lowish frequency of light skin genes to an area, which gets driven to up to 90s of percents afterwards.

Interesting question of whether populations with Indo-European ancestry in West Asia and Middle East generally tend to have higher frequency of derived pigment variants that peak in Europeans.

I've had a try at looking at that by simply downloading frequencies from ALFRED of 3 most distinct SNPs at SLC24A5, HERC2, SLC45A2: https://imgur.com/a/k33K0

(Occasionally in a few pop some of these frequencies stand out as shaky, but I think the general patterns look OK, and enough to estimate general patterns).

First two just plot frequency of SLC24A5 v HERC2, then SLC24A5 v SLC45A2, then I used PCA and Neighbour Joining dendrogram to try and visualise them together.

General pattern is that East Asian and African populations low in all derived variants (basically at 0), then two other poles are formed with a) Europeans having high derived variants of all 3, and b) the Middle East having derived variant at SLC24A5 only. Beyond that level of detail, South Europeans have specifically relatively lower frequencies of HERC2 derived variants and higher SLC45A2 than expected.

In general it seems like there is very much overlap between 1) NW Africans (Aqua in my plots), 2) populations from Pakistan with high levels of Yamnaya ancestry in Lazaridis (Darkred) and 3) also the Near Eastern Levant to Arabia.

Of these three, the set with that edges highest levels of the specific European SNPs seems to be marginally be NW Africans, despite any influence of West African ancestry, and who in theory would have the least Yamnaya ancestry of the lot (and rather be a mix of mainly Levant N expansion, some MN EEF, plus more marginal ancient NW African and/or West African ancestry)...

This is just a simple test though - someone could use a much denser set of pigmentation related SNPs and populations to predict a best guess map of population distance in pigmentation phenotype.

Rob said...

@ Dave

I said "There was a migration of GAC to steppe" from Kuyavia, didn't imply any mass settling. Nevertheless as I outlined it certainly did reach the steppe, the steppe parts of Transylvania, and even some GAC type cist graves in Kemi Oba.
Now that I have afresssd your distraction

"Awesome theory. You were there are saw this happening?"

Of course it's awesome. Thanks .
I haven't mastered time travel yet, but do know enough to realise that I-I to F-U loans weren't realised via an internet chatline

Rob said...

Just to clarify, I am not suggesting that Sintashta was FU, rather that the forts and activities surrounding them would have knock on effects in the foreststeppe & southern forest zone, incl. language spread. An idea neither wholly novel, nor controversial, so not sure what is the issue

supernord said...

Chetan Vit said...
"Parpola seems to think Abashevo and Fatyanovo were early Uralic cultures."

Why Parpola right? It is a common thing, at any cost to cling to the famous cultures.


"Since the chariots only appear from Sintashta onwards, maybe the Abashevo were assimilated by the Indo-Iranians from Poltavka."

Poltavka was not Indo-Iranians. This Abashevo (and others) began to destroy Poltavka, not Vice versa. ~50% of all Poltavka were killed by the breach of the skull. Abashevo member of the charioteers culture, there was found the oldest elements of horse harness.


"That would explain the shared vocabulary/ loanwords between IIr and Uralic."

Sintashta & Andronovo explain.

supernord said...

@Rob

"Sintashta could have precipitated the expansion of FU around it's periphery."

Rather Seima-Turbino people.

Shaikorth said...

@Chetan
Parpola correlated Fatyanovo with Northwest Indo-European and Abashevo with early Proto-Indo-Aryan.

Source: Parpola, Asko, 2012 (2013). "Formation of the Indo-European and Uralic (Finno-Ugric) language families in the light of archaeology". Pp. 119-184 in: R. Grünthal & P. Kallio (eds), A Linguistic Map of Prehistoric Northern Europe. Helsinki. (page 146)

supernord said...

"Parpola correlated Abashevo with early Proto-Indo-Aryan."

Parpola are wrong, Abashevo can't the Proto-Indo-Aryans, culturally not suitable, but genetically may be relatives.


First from his text "The Sejma-Turbino transcultural phenomenon (c. 2100/1900–1600 BCE), which seems to have its origin in the Abashevo culture"

It seems Parpola does not know anything, the whole text in continuous errors. It cannot be taken seriously.
Parpola is not authoritative.

Shaikorth said...

His idea of Volga/Southern Urals origin of Seima-Turbino instead of an Altaian origin was backed with following:

"The majority of the Sejma-Turbino objects are of the better quality tin-
bronze, and while tin is absent in the Urals, the Altai and Sayan mountains are
an important source of both copper and tin. Tin is also available in southern Cen-
tral Asia. Chernykh & Kuz’minykh have accordingly suggested an eastern ori-
gin for the Sejma-Turbino network, backing this hypothesis also by the depiction
on the Sejma-Turbino knives of mountain sheep and horses characteristic of that
area. However, Christian Carpelan has emphasized that the local Afanas’evo
and Okunevo metallurgy of the Sayan-Altai area was initially rather primitive,
and could not possibly have achieved the advanced and difficult technology of
casting socketed spearheads as one piece around a blank. Carpelan points out that the first spearheads of this type appear in the Middle Bronze Age Caucasia c. 2000 BCE, diffusing early on to the Mid-Volga-Kama-southern Urals area, where “it was the experienced Abashevo craftsmen who were able to take up the new techniques and develop and distribute new types of spearheads” (Carpelan
& Parpola 2001: 106, cf. 99–106, 110). The animal argument is countered by reference to a dagger from Sejma on the Oka river depicting an elk’s head, with
earlier north European prototypes (Carpelan & Parpola 2001: 106–109).
Also the metal analysis speaks for the Abashevo origin of the Sejma-
Turbino network. Out of 353 artefacts analyzed, 47% were of tin-bronze, 36% of
arsenical bronze, and 8.5% of pure copper. Both the arsenical bronze and pure
copper are very clearly associated with the Abashevo metallurgy. While the tin-
bronze dominates in Siberia, it covers less that 30% of the European artefacts;
arsenical bronze covers 45% and pure copper 10% of the European material, but
their distribution in Siberia is very small (cf. Chernykh 1992: 222–224). How-
ever, objects made of pure copper have been found in the Altai (cf. Chernykh &
Kuz’minykh 1987: 94, map 20)..." etc.

Ancient DNA will solve the origin of the people themselves eventually.

Chetan Vit said...

@Shaikroth Actually it's in that same paper Parpola refers to Fatyanovo-Balanova as representing proto-Uralic p.20

" The Abashevo culture (c. 2200–1850 BCE)(Figure 5), which extended along the border of the forest steppe and the forest zone from the Upper Don to the Upper Tobol river, had its origin in these eastern Late Yamnaya cultures. Its eastward expansion was motivated by the sandstone deposits with pure copper, for which the Abashevo people on the Lower Kama and Belaya rivers fought with the Balanovo people. While the Abashevo burials with their kurgans are similar to those of the Poltavka culture, the early Abashevo ceramics resemble the Fat’yanovo-Balanovo Corded Ware, which had been in these parts earlier along with the Volosovo and Garino-Bor cultures,candidates to archaeological correlates for Late Proto-Uralic. "
@supernord Parpola is an authoritative source. Obviously he didn't have access to the ancient genomes we have now, but most of his predictions about the Urheimat are being proven correct.

Davidski said...

I won't be around much for the next three-four weeks. Merry Xmas everyone...

Shaikorth said...

@Chetan
He obviously means Volosovo and Garino Bor only since page 146 mentions Fatyanovo as a NW-IE Corded Ware derivative. The same paper suggests that the Volga-Kama remnants of Abashevo were only uralicized after Sintashta had formed, so Sintashta has nothing to do with Uralic in that scheme.

Chetan Vit said...

He predicted the early Khvalynsk represented PIE, Khvalynsk migrations westwards to form Sredny Stog and the early Hittite migrations through the Balkans c.4000 BC. All of this is being proven correct now, isn't it? Correct me if I am wrong.

Shaikorth said...

I don't know about the Hittite in Balkans but Khvalynsk PIE looks reasonable.

Rob said...

" ChetanVit

''He predicted the early Khvalynsk represented PIE, Khvalynsk migrations westwards to form Sredny Stog and the early Hittite migrations through the Balkans c.4000 BC. All of this is being proven correct now, isn't it''

None of that has been proven at all. In fact, the data suggests otherwise.

supernord said...

"Parpola is an authoritative source."
No, he is a layman in these questions. He deals with other matters, and in these he climbed completely without understanding them.

"most of his predictions about the Urheimat are being proven correct."
No, what he says is a mistake and always has been. He always was contrary to all known archaeological and linguistic data.

It is not necessary to discuss he at all, he is not worthy of it.

Chetan Vit said...

@Rob What else does the data suggest?

supernord said...

Shaikorth said...
" His idea of Volga/Southern Urals origin of Seima-Turbino instead of an Altaian origin was backed with following:"

Not a good idea. The raw materials were taken from the nearest source, no one was carrying raw materials for thousands of kilometers through the Taiga. These people are not in one generation has spread to Estonia.

Matt said...

Hmmm... On the purely autosomal level, is low level GAC migration - essentially 25% WHG: 75% AN - as an element in forming Steppe_EMBA steppe so crazy?

When I use the Fst and PCoA based method with nMonte, and restricted the calc to only be pre-Yamnaya ancients without pure HG and only including Steppe_Eneo, I find that, although the largest contributor of non-Steppe Eneolithic ancestry to Steppe_EMBA is CHG, other populations that show at more minor contributions are the Europe_MNChl, Greece_LN, Iran_N (actual model was Steppe_Eneo - 60.5%, CHG - 15%, Europe_MNChl - 11%, Greek_LN 6%, Iran_N 6%).

This is not necessarily the best method, not using formal stats and using Fst measures that may have some fine error. However I do think that we have found in all the new papers that, even using formal stats when a more dense set of ancient outgroups are considered, there are offsets that mean that, though CHG is still the *best* single reference, it doesn't *totally* explain relatedness to "Western farmers" which is greater in Yamnaya than a pure CHG:EHG or CHG:Steppe_Eneo combo (and also Davidski has found some slight WHG affinity as well).

With Anatolia_N and EEF in Europe we are "lucky" in that at the time of early Neolithic groups setting out to Europe, there really was only one group of farmers in a position to move there, and substructure was quite differentiated in early farming regions, and then we have fairly better sampling of the trail from the Near East to Europe at the right time.

At the moment with Steppe_EMBA, though we can be sure that their non-EHG related ancestry was predominantly CHG like, checking the fine details is more difficult, and seems possible that with more Near Eastern population mixture by that time, and population movements from multiple directions into the steppe, there might be more complexity at the margins.

@Davidski, Merry xmas.

Olympus Mons said...



This is why I am pushing questions about H15a1

a. Shulaveri H15a1 is found by 6000bc in the arriving population of Shulaveri to Armenia. And we know that at least their cattle/sheep was coming from Anatolia not iran (just next to them).

b. We have H15 then in steppe Yamnaya (as we have I1 another Shulaveri mtdna found).

c. We have downclades H15a1a1 in Armenia/Iran but mostly in Pamiris people, the ones Davidaky so keenly “proves” are Yamnaya derived, steppe, steppe. Pamiris that show lots of R1a/r1b-M269.

d. We have downclade H15a1b1 also prevailing in Pamiris the “yamanya steppe” people.

e. Sister clade H15b is found in Armenia, not in east!, but also Europe (mostly Denmark).

So, if earliest H15a1 is found in Shulaveri in south caucasus and it is seen later following “the great steppe” into east but not west doesn’t it imply or at least strongly supports that the Yamnaya seen thus far in samara was an offshoot of those Shulaveri, the south Caucasus master of domestication (plant and animal) moving to steppe, and explaining why the steppe R1b clade (Z…) died in steppe?

Yeah right…. Steppist are very smart to what suits their confirmation bias.

Shaikorth said...

The idea in the cultural spread scheme is not that they carried raw materials, but that the tin-usage in Altai was preceded and developed by arriving pure copper metallurgists from the west.

Olympus Mons said...

What I have stated above can also be infered by H2.

H2+152 seen in Shulaveri, south caucasus, by 6000bc is seen in Kum6 by 4700 BC as H2a1 in anatolia, in Khvalynsk man 4500bc as H2a1, Sredny Stog and also common with Pamir.

supernord said...

His idea is in "Also the metal analysis speaks for the Abashevo origin of the Sejma-
Turbino network. Out of 353 artefacts analyzed, 47% were of tin-bronze, 36% of
arsenical bronze, and 8.5% of pure copper. Both the arsenical bronze and pure
copper are very clearly associated with the Abashevo metallurgy. While the tin-
bronze dominates in Siberia, it covers less that 30% of the European artefacts;
arsenical bronze covers 45% and pure copper 10% of the European material, but
their distribution in Siberia is very small (cf. Chernykh 1992: 222–224). How-
ever, objects made of pure copper have been found in the Altai (cf. Chernykh &
Kuz’minykh 1987: 94, map 20)"

From this text it is clear that he contradicts himself. And how the Bronze age came to Altai it is quite another matter, but Abashevo to him is irrelevant, as he writes.
But in general, all this text focuses on the most marginal ideas.

Rob said...

@ Chetan

I'd need 7 pages ;)
But if you briefly want to appraise the Khvalynsk model specifically and some potential problems with it - but not against a Caucaso-Ponto-Caspian scheme in general, then here are some factors i have seen so far. Most of it seems to hinge on the ad hoc attribution of Khvalynsk as something remarkable in the scheme of things, and that's just tradition amongst the Indo-Europeanists that most are familiar with, but further details provided by less mainstream works had already shown pitfalls, as does the aDNA.

- the earliest strata of lineages in Ukraine (where we have more data from) are R1b XP297, R1a and I2a1. By the Sredny Stog & Mariupol horizon (which you mentioned), there is a swamping by I2a2 and R1b-V88, which could only have arrived from the west, as did the domesticates and ealry Copper they brought with them. These men were not Anatolian farmers, but the first line of European hunter-gatherers to adopt aspects of the ANF package, profited and in turn expanded mostly northeast, whilst the main bulk of mainstream ANF, haplogroup G2 heavy farmers continued northwest up the Danube to LBK land, etc.

- Khvalynsk is part of Sredni Stog II horizon, thus after the above event. The arguement goes that these individuals invaded the Balkans with their alleged horses, and then quickly moved onto Anatolia as well. However, there is no evidence for any steppe-derived settlements or monuments or mounds in the East Balkans at this time. In fact, after 4000 BC there is not much evidence of anything in the former core area of Varna -KVK VI, for several hundred years. In fact, the downfall of centres like Varna impacted the steppe, because apparently for a couple of hundred years there seems to an impoverishment of former grand burials like at Khvalynsk itself.

- lets also note that Khvalynsk so far has not produced any R1b-M269 from admittedly paultry 3 samples. And even if we accept that the CHG rise was simply due to Caucasian or Kazakh-Aral steppe women, it still needs some taking into account for who these people were.

- in place of Khvalynsk, the Repin culture emerges in the shadows of Majkop influence, and in turn these Repin groups moved west as proto-Yammnaya and then classical Yamnaya groups, somewhat displacing or absorbing the earlier groups of Sredny Stog, and finally appearing in the Balkans after 3000 BC, although some appeared in Hungary as early as 3300 BC.

- there is evidence of movement from Balkans to Anatolia, but this is from a muchmuch earlier period, like late Neolithic and Copper Age, and from E balkans centres, not Khvalynsk.

Rob said...

@ Matt

"there might be more complexity at the margins."

And in the centre of the steppe too! (We'll eventually see).

Shaikorth said...

@supernord
He elaborates later on in the paper that the pure S-T copper metallurgy of the Altai is of western origin because the Seima techniques are of advanced type which appeared first in the west, pure copper processing early Abashevo would be predecessor to both western S-T arsenical bronze and eastern S-T tin-bronze artefacts.

All written before ancient DNA from the area, but the genetic correlate to this would be the appearence of R1a-Z93 in Bronze Age Altai.

Chad Rohlfsen said...

Some issues with the models and nMonte based on Dstats are based in the alignment issues. For instance, Natufians and Iran are misaligned probably the worst of all. This may potentially lower the importance of these groups, or inflate depending on outgroups.

CHG is aligned much better than Iran N, which might be why it is favored over Iran, even with its high level of ROH. We already know ROH can cause issues with F_st.

Anatolia is also definitely aligned worse than European farmers, so models with it will likely underrepresent actual farmer ancestry. There's a whole host of issues, but I'm not sure how much of and impact each level of misalignment has on models.

One sanity check is with f3, which Nick told me is less affected by this. So, I've done about 2000 f3s just regarding farmers and found many interesting things. I'm going to be putting something together to share, once I catch up on schoolwork.

Rob said...

@ superNord

"Rather Seima-Turbino people."

That's what I always favoured, but I guess it depends which dating one accepts for FU expansion- an earlier MBA one or a LBA (S-T) one
It would slot nicely if ST is Uralic, coming from Altai and all, but it's a strange phenomena, so many epitaphs etc. The lack of detailed study on the Chinese end of the culture sites doesn't help either

Rob said...

^ cenotaphs, not epitaphs

supernord said...

@Rob
"he arguement goes that these individuals invaded the Balkans with their alleged horses, and then quickly moved onto Anatolia as well."

No, no everybody who say that "quickly moved". The Hittites appeared in Anatolia not earlier than ~2000 BC.

"However, there is no evidence for any steppe-derived settlements or monuments or mounds in the Balkans at this time."

They are exist in biggest count, more the genetics facts.

"In fact, after 4000 BC there is not much evidence of anything in the former core area of Varna -KVK VI, for several hundred years."

Do not matter.

"- lets also note that Khvalynsk so far has not produced any R1b-M269 from admittedly paultry 3 samples."

R1b i not matter. Expansion PIE is not associated with R1b.

"the Repin culture emerges in the shadows of Majkop influence"

This is an very big exaggeration.


Shaikorth said...

"pure copper processing early Abashevo would be predecessor"

Pure copper processing earliest was Everywhere! But that's not important. I already wrote that its concentration in Abashevo in the issue of the appearance of the Bronze age in the Altai does not hold water. Papola not an archaeologist, he is in these questions do nothing about. Пenerally, all his text is meaningless and irrelevant.

Rob said...

@ Supernord

''
"However, there is no evidence for any steppe-derived settlements or monuments or mounds in the Balkans at this time."

They are exist in biggest count, more the genetics facts.'

Where ?

Rob said...

@ S/ N

''No, no everybody who say that "quickly moved". The Hittites appeared in Anatolia not earlier than ~2000 BC.'

There are other Anatolian languages also, dont forget. And they all are of a broadly similar category, but by attestation, significantly diverged, moreso than Slavic at 1000 AD or Romance in late Empire days. This sets a sanity check which stipulates that the Anatolian languages must have been drifting apart for at least 1500 years.

The other reason why Hitties entering in or after 2000 BC is nonsense is becuase at thi time the moving direction as from Anatolia to Europe - with south Anatolians moving to Crete, and NW Anatolia colonisgts in Thrace.

Shaikorth said...

@Supernord
It's not just the material, pure copper, but the argument goes that the technique which in Seima-Turbino is advanced like in the west, it's less likely that the more primitive Altaians developed the casting in situ. This is not actually Parpola's research, but Carpelan's who is an archaeologist.

If R1a-Z93 appears in Abashevo and keeps popping up in the Bronze Age Altai during the S-T timeframe the correlation between that theory and genetics will become clearer.

Rob said...

"R1b i not matter. Expansion PIE is not associated with R1b"

Exactly the R1b in khvamysnk doesn't matter
Neither does the Q and R1x M17
All irrelevant

supernord said...

@ Shaikorth
"This is not actually Parpola's research, but Carpelan's who is an archaeologist."

Carpelan does not write about Abashevo descendant. About Abashevo it is Parpola claim. It doesn't change anything.

The largest number of finds of the Sejm-Turbino was on the territory of the Abashevo culture, since the Turbino was easier to attack them than the well-fortified city of Sintashta.


@Rob
"There are other Anatolian languages also, dont forget."

Here in the Hittites refers to all Hittite-Luwic, to simplify the text. History of Hittite-Luwic to 1900 BC in Anatolia does not fix. They all appear in the history later of the Hittites.

R1a is Relevant by fact.

Rob said...

Don't really understand what you're saying
But the Anatolian languages split from the maintrunk before 3500 BC, and internally amongst themselves soon after them
Therefore your scenario , similarly proposed by others like Folker, requires a special set of pleading where multiple groups all moved into Anatolia at same time who all happens to speak the same language, at the very moment that the archaeological amd genetic evidence shows people moving out of it
Not very convincing or even comment worthy

Rob said...

Supernord
And where is the "plenty of evidence " you suggested ?
Back up your claims, for once

Shaikorth said...

@supernord
Carpelan notes that the advanced casting techniques involved in the Seima Turbino spearheads appear first near the Caucasus, then they diffuse to Volga. Abashevo (or Abashevo derivatives, Abashevo horizon proper did not extend to Altai) as the source of the diffusion from there to Altai is Parpola's idea but it is based on Carpelan's work, with citations. If there will be Z93-findings from Abashevo they should be compared to BA Altai Z93 to test the theory. If Z93 is missing in Abashevo their connection to BA Altai and Sintashta will both become questionable.

Rami said...

@David you truly are weird at times
Razib is right here , Steppe ancestry is not fundamental for lighter phenotypes, those easily can be attributed to selection. You can find Near Eastern populations with light skin, eyes and hair.

Also there is no way Pamiri Tajiks are 50% Yamnaya, the Sintashta Iranic groups they descend from themselves have substantial Farmer ancestry via Western/Mainland Europeans, a 1/3 to 2/5 of their ancestry. Actual Yamnaya ancestry is 17-30%.

supernord said...

@Rob

You do not understand that between the Steppe and Anatolia is a huge territory. There were different intermediate culture like Cernavoda, Ezero, etc. This is your invention about the rapid penetration into Anatolia from the Steppes, not one we've ever wrote, so it's irrelevant and not serious.

Shaikorth, thank you, I do not need to explain what I understand. But I can't assume that reasoning Parpola is correct, they are just simply not true at all. The opinions of Carpelan should not be conclusions Parpola.

Shaikorth said...

@supernord
Noteworthy that Kuz'mina (The Origin of the Indo-Iranians p.253) came to a similar hypothesis, deriving S-T metallurgy from Abashevo derivatives and early Fedorovo traditions interacting in the east instead of the earlier Altaian and South Siberian traditions.

Rob said...

But hey, when R1a and Q start appearing in 4000 BC balkans, I'll buy you some Slivavitze

supernord said...

@Rob

But then the Khvalynsk in your statement about the "quickly moved" to Anatolia from Khvalynsk? You just arrogant invented a kind of "quickly moved" in Khvalynskiy time, and when I explained to you that no one has invented any quickly movement in Anatolia, you just smack some nonsense. About the history of Anatolia you know nothing. Your building is empty.


@Shaikorth

The transfer from West to East metallurgical traditions, no one denies, but on the way through Central Asia, and the fact that in Siberia the metal came from Andronovo, mentioned Kuz'mina, nor about something else I didn't read she. But she certainly couldn't write from that Sejma-Turbinо from Abashevo. So the text doesn't apply to such hypothesis, they are not even similar.

MomOfZoha said...

Christmas? Since you like appropriating Iranian culture, enjoy Yalda Night, David. In your hemisphere there is actually something to enjoy: The shortest night, the longest day of the year. In this hemisphere we'll contend with the fact that days can only get longer henceforth. No wonder we need another celebration to cheer things up...

Nothing against the baby Jesus, whose symbolic birth I will celebrate in the first week of January. After all, Isa is the father of my mother-in-law Maryam. Only along the Caspian does one find Jesus fathering Mary.

Shaikorth said...

She does propose Abashevo/Fedorovo hybrid origin, contrasting Chernykh's theory of a local origin. Direct quote from Kuz'mina's The Origin of Indo-Iranians:
"
What is the origin of Turbino-Seyma bronzes? E. Chernykh and S. Kuz'mi-
nykh (1989: 259-261; 270) dated them to the 16th century BC and presumed that
the complex formed in Siberia as a result of the interaction between the culture
of the hunter-fishers of Baikal and that of the Altai's horse-breeders and metal-
lurgists. A clan of armed nomad-metallurgists would carry out distant raids on
horseback and spread their products in the west. V. Bochkarev (1986) estab-
lished the wide European relations of the bronzes and determined the chronolo-
gical sequence of the complexes: Turbino-Seyma-Rostovka and Samus' IV,
Turbino being synchronized with the Abashevo culture. O. Kuz'mina (2000: 65-
134) confirmed the relationship between the metallurgy of the European Aba-
shevo culture and Turbino and demonstrated that many types of early Andronovo
metal from Sintashta advanced traditions of Abashevo metalworking (adzes,
double-edged and single-edged knives, spears, shafted arrows, hooks, bracelets).
This bears out the role of the European traditions for the metalworking of
Turbino-Abashevo-Sintashta. However, there is a group of bronzes of eastern
origin at Seyma (Chernykh 1970: 155-173). In Sintashta two articles of tin
bronze alloyed with lead (Pb) and antimony (Sb) were found, which points to
their being exported from eastern Kazakhstan (Zaykova 2000). That is also the
likely source for the single -edged knife with a representation of an argali known
from Turbino and one from Seyma's one with a pair of horses (Bader 1964: 115-
123, fig. 113; 1970: fig. 52).

This allows one to pose a working hypothesis concerning the formation of
Turbino-Seyma bronzes as a result of the interaction between the population of
Eastern Europe (above all, Abashevo and, partially, Catacomb tribes) and early
Andronovo tribes of the Fedorovo type in eastern Kazakhstan. There bronzes
have already been recovered from the Kanay cemetery which preserve the
Eneolithic traditions, and the early Marinino stage of the Fedorovo type has been
established (Tkacheva 1997: 12). The acceptance of the calibrated C 14 dates of
the Abashevo (Mikhaylova and Kuz'mina 1999: 119) and early Andronovo sites
of the Novy Kumak stage compels one to assign Turbino to an earlier time and
date it to the turn of the 3rd-2nd millennia BC. It may be presumed that it is the
early Andronovo tribes of Siberia and eastern Kazakhstan that were instrumen-
tal in the appearance in China, in the northern contact zone, of tin and bronze
articles, stone and clay molds, the technique of casting celts and spears with a
concealed socket, as well as types of adze, the single-edged knife and the ring-
headed dagger in Erlitou in Henan (Linduff 1994: fig. 3,18) and the dagger from
Zhukaigou, phase 5 (Linduff 1997: fig. 6 bottom)."

Seinundzeit said...

Matt,

I feel like there is something odd going on with the ALFRED data; Orcadians, Andalusians, and Adygei have slightly less derived alleles (on average) at SLC24A5 than most Pakistanis? Sindhis, IRL the darkest Pakistanis around who happen to be predominately West Eurasian, are overall somewhat more derived at light pigmentation-implicated SNPs compared to Pashtuns? And the Kalash are only around 70% for derived SLC24A5 (they're at 100% with the HGDP data, which makes far more sense with what one sees IRL)?

Very strange stuff, so I would take with a pinch of salt.

A far more accurate, rigorous catalog of human variation at pigmentation-implicated SNPs (with HGDP populations) can be found in S. Walsh's "DNA Phenotyping: The prediction of human pigmentation traits from genetic data".

For what it's worth, in that text (and if my memory serves me right), the Kalash cluster with Sardinians and Druze, when using all pigmentation-related SNPs.

Rami,

David didn't claim that Pamiris are 50% Steppe_EMBA.

His model above involves Sarmatians, who were East Iranian speakers, just like modern Pamiris and Pashtuns.

Samuel Andrews said...

@Seinundzeit,

The frequency of SNPs rs16891982 and rs1426654 alone can't measure a population's skin color. Many other SNPs, which have not been identified as making an influence on skin color, do make an influence on skin color. One population can have a lower frequency of derived alleles in those SNPs than another but be lighter skinned. That shouldn't be the case but don't be suprised if it is.

Btw, my own bunch, not based on any indepth research or anything, is the Iran Neo/ASI population of South Asia had very dark skin similar to Indians. And the R1a Z93 Steppe populations who moved in were fair, maybe not as much as Andronovo, but definitely not out of the ordinary for modern Europeans. But a minority or even 50% Steppe ancestry still results in for the most part brown skin.

Anthro Survey said...

@Davidski

I have to emphasize something.

Unfortunately, various 20th century nationalisms based on false premises have created certain paradigms in popular consciousness, which continue to be reinforced by the media and various Iranicists, Turanicists, etc. I'm sure I don't need to tell you(or Rob) about Greek confirmation bias, not to mention false notions about being cultural descendants of the old Hellenes. :-) Well, it is the same when it comes to Greater Iran.

No, Tajiks do not speak East Iranic if we use the term 'Tajik' properly in its traditional sense.
Historically, the term Tajik referred to urban or settled agriculturalist Farsi(New Persian) speakers residing in what are termed "Persianate" or "Turko-Persian" states. Farsi is derived from Middle Persian, which is a Western Iranic language. As the centuries passed, it was usually relegated to those hailing from east of the Great Salt Desert.

In a Persianate society, the Tajiks traditionally occupied professions like artisans, scribes, imams/muftis, grand viziers, etc. In other words, they were 'men of the Pen', whilst Turks, often semi-nomadic and illiterate, were usually 'men of the Sword' and the ruling families of any given sultanate.

A quote from a poem by Sa'adi reflecting this dichotomy:
"It's appropriate to tell the King,
Your Turk shed the blood of Tajik"

The Great Seljuk and Ghaznavid Sultanates were classic examples of and, in many ways, prototypes for future Persianate states.

Modern-day Western Iran was NOT the cultural epicenter for where that all began, either. The foundation for New Persian, or Farsi, was laid in Greater Khurasan sometime during Samanid times. The new standard was immortalized in Ferdowsi's epic Shahnama, tr. Book of Kings. Now, Middle Persian, from which Farsi derives, took hold in Khurasan's landed nobility during Sassanid times, at the expense of Sogdian and Bactrian. Khurasan is a region encompassing modern-day NE Iran, Mughrab region, north Afghanistan(esp Balkh), Oxus basin, and the western valleys of Tajikistan. The Persian "big shots" of the Islamic Golden Age I've enumerated hailed from there, not Zagros-proximal Iran. I'm sure you've heard of Avicenna(Ibni Sina), Ferdowsi, Rudaki, Mowlana Rumi, Nadir Tusi etc.

All this being said, it's more proper to call them Pamiris, not Pamiri Tajiks. They do not speak Farsi/Dari, don't practice most Persian customs and live a rural, mountain life in difference to the urban Tajiks of modern Afghanistan(esp in Herat and Mazar-i-Sharif) as well as those inhabiting Zerafshan and Oxus valleys of the former soviet union. They are called Tajiks in large part because they share borders with proper Tajiks who are dominant population in modern-day Tajikistan. I should stress that many "proper" Tajiks living there are, in fact, refugees from Afghanistan and Uzbekistan's cities thanks to ethnic cleansing taking place after Soviet Union fell apart.

Anthro Survey said...

In fact, proper Tajiks from cities like Samarkand, Mazar-i-Sharif, Herat, etc. range between 20 and 35% steppe and nowhere near 50% if you subtract out their minor Turkic ancestry.

Unlike with Pamiris annd Yaghnobis, Eastern Euro looking individuals tend not to be commonplace among them. Typical Tajiks who don't sport "Turanid" looks often resemble singers Valy and Sadriddin, as well as this gentleman: https://en.wikipedia.org/wiki/Taqiyah_(cap)#/media/File:Young_Tajikistani_man_with_guitar.jpg

In Afghanistan, Pashtuns definitely have more East/North-Euro looking individuals, but I'm not sure if that's due to their steppe ancestry being considerably higher or more so because of bottlenecking dynamics.

Matt said...

Sein: I feel like there is something odd going on with the ALFRED data; Orcadians, Andalusians, and Adygei have slightly less derived alleles (on average) at SLC24A5 than most Pakistanis?

Yeah, like I say, there's some noise there in some of the panels, so wouldn't fixate on any particular individual population or think of it as invalidating the strong trends in these three variants between Europe, Middle East, rest of world. (Re: Sindhi and the Pakistani Pashtun/Pathan, Lazairidis estimated them at like, what, 5% difference in Steppe_EMBA ancestry - https://1.bp.blogspot.com/-9k5UilpTSEA/V2TqxrBdxgI/AAAAAAAAFCE/awWnaVwRo6cPa2xjxEKuNKRLfmLIM5f3gCPcB/s1600/MODEL.png?).

The main takeaway I get from it is that it seems unlikely that, considering all these populations and looking at the median of regions, the Steppe EMBA influenced populations in South and Central Asia don't seem to have higher frequency derived, lighter pigmentation variants at these alleles (SLC24A5, SLC45A2, HERC2), than seen generally in the Levant or North Africa, and so noticeably different effects, at least in terms of skin and eye. There might be something going on with KITLG blonde hair variant though. I wouldn't take these literally for individual populations!

Davidski said...

@Chad

Don't lose yourself in Admixtools too much.

Formal stats aren't the only way to analyze data. And obviously there are different types of data, like haplotypes and uniparental markers. So the best sanity check is not f3, but how results from different types of ancient data correlate with results from modern data.

Ask yourself, are modern samples behaving as if they were the descendants and mixtures of the ancient samples, as per the formal stats, haplotype data and uniparental markers? If they are, and it does appear that way, then the problems you're talking about must be having a fairly minor impact, in this context anyway.

If you think Iran_N is so much more important, then why doesn't the haplotype data and uniparental markers show this? Coincidence?

@Rami

All South Central Asians have significant steppe ancestry, even the Burusho, and even the groups that are fairly recent arrivals from West Asia, since West Asians also have steppe ancestry. So I'm not sure why you think that European-like phenotypes in the region are not a reflection of this in varying degrees?

Of course, the major proviso is that, thousands of years after the steppe migrations, the level of steppe ancestry in each individual, and even in each population, is unlikely to be closely reflected in their looks. But no one is arguing that this is the case, certainly not me, so calm your tits.

And if you want to lecture me about how hopeless and wrong my mixture models are, then first demonstrate that you're in a position to do so, because right now you're not in such a position.

Challenge my qpGraph models with your own qpGraph models, and let's see whether you have anything more to offer than empty opinions based on your own preferences. Fair enough?

@MomOfZoha

Control your complexes. Try not to be just another Rami. One is enough.

Anthro Survey said...

@Davidski
Oh, did you already cover this paper here? Just stumbled upon it and seems like there are more new EEF genomes. PCA has horrible projection bias, though.
http://rspb.royalsocietypublishing.org/content/284/1867/20171540

Davidski said...

@AS

I've commented on the same samples at my other blog, because they were also part of Mathieson et al. 2017.

http://polishgenes.blogspot.com.au/2017/05/globular-amphora-people-starkly.html

Nothing more to add, although feel free to comment on the new paper in this thread.

Shahanshah of Persia said...

@Davidski

"Can't say. I only have one Yaghnobi in this dataset and no Nuristanis."

Thanks, appreciate it.

@Seinundzeit

"For what it's worth, the Nuristani seem to be identical to the Kalasha (in the context of ADMIXTURE; I've seen some results), so we can assume somewhere around 40%-50% Steppe_EMBA/Srubnaya_outlier/Potopovka_Outlier-related ancestry."

Alright, makes sense. Thanks for sharing your input, mate.

Shahanshah of Persia said...

@Anthro Survey

"Unfortunately, various 20th century nationalisms based on false premises have created certain paradigms in popular consciousness, which continue to be reinforced by the media and various Iranicists, Turanicists, etc. I'm sure I don't need to tell you(or Rob) about Greek confirmation bias, not to mention false notions about being cultural descendants of the old Hellenes. :-) Well, it is the same when it comes to Greater Iran."

Well, to an extent, Greeks are the cultural descendants of the "old" Hellenes, whatever that means. What you have to remember is that the Greeks of today are the only ones who speak the Greek language as their native tongue, and thus are the true inheritors of ancient Greek civilization, whether you like it or not, mate. Although, I do agree, that a lot of modern nationalism do have false premises.

"No, Tajiks do not speak East Iranic if we use the term 'Tajik' properly in its traditional sense."

He never claimed that they do. The Pamiris do, however, speak an east Iranic language. The Pamiri language is not mutually intelligible with Tajik, and it is not a Western Iranic language.

"Historically, the term Tajik referred to urban or settled agriculturalist Farsi(New Persian) speakers residing in what are termed "Persianate" or "Turko-Persian" states. Farsi is derived from Middle Persian, which is a Western Iranic language. As the centuries passed, it was usually relegated to those hailing from east of the Great Salt Desert."

This is true.

"In a Persianate society, the Tajiks traditionally occupied professions like artisans, scribes, imams/muftis, grand viziers, etc. In other words, they were 'men of the Pen', whilst Turks, often semi-nomadic and illiterate, were usually 'men of the Sword' and the ruling families of any given sultanate."

Correct again. In fact, many Iranians from the plateau were also referred to as Tajik by their Turkic overlords.

"A quote from a poem by Sa'adi reflecting this dichotomy:
"It's appropriate to tell the King,
Your Turk shed the blood of Tajik"

Thanks for sharing this quote, but it is a bit ironic considering that the Turks were the soldiers and martial elites. If anything, the Turks were the ones shedding the most blood in most conflicts, with the obvious exception being the Mongol invasions.

Shahanshah of Persia said...

@Anthro Survey

"The Great Seljuk and Ghaznavid Sultanates were classic examples of and, in many ways, prototypes for future Persianate states."

Yes, this was actually proven true with the Khwarezmians.

"Modern-day Western Iran was NOT the cultural epicenter for where that all began, either. The foundation for New Persian, or Farsi, was laid in Greater Khurasan sometime during Samanid times. The new standard was immortalized in Ferdowsi's epic Shahnama, tr. Book of Kings. Now, Middle Persian, from which Farsi derives, took hold in Khurasan's landed nobility during Sassanid times, at the expense of Sogdian and Bactrian. Khurasan is a region encompassing modern-day NE Iran, Mughrab region, north Afghanistan(esp Balkh), Oxus basin, and the western valleys of Tajikistan. The Persian "big shots" of the Islamic Golden Age I've enumerated hailed from there, not Zagros-proximal Iran. I'm sure you've heard of Avicenna(Ibni Sina), Ferdowsi, Rudaki, Mowlana Rumi, Nadir Tusi etc."

You are correct again, for the most part. However, your assertion that Middle Persian took hold among Sassanian noblemen in Khurasan is not well substantiated. I think you are forgetting the the Sassanid heartland was very much in western Iran, not in Khurasan. The Sassanids viewed their northeastern provinces as nothing more than buffer against Turkic and Iranian tribes of the Central Asian steppes.

"All this being said, it's more proper to call them Pamiris, not Pamiri Tajiks. They do not speak Farsi/Dari, don't practice most Persian customs and live a rural, mountain life in difference to the urban Tajiks of modern Afghanistan(esp in Herat and Mazar-i-Sharif) as well as those inhabiting Zerafshan and Oxus valleys of the former soviet union. They are called Tajiks in large part because they share borders with proper Tajiks who are dominant population in modern-day Tajikistan. I should stress that many "proper" Tajiks living there are, in fact, refugees from Afghanistan and Uzbekistan's cities thanks to ethnic cleansing taking place after Soviet Union fell apart."

Well said, agreed!

Shahanshah of Persia said...

@Anthro Survey

"In fact, proper Tajiks from cities like Samarkand, Mazar-i-Sharif, Herat, etc. range between 20 and 35% steppe and nowhere near 50% if you subtract out their minor Turkic ancestry."

No, they range between 35 and 45% Steppe, closer to the low 40s.

"Unlike with Pamiris annd Yaghnobis, Eastern Euro looking individuals tend not to be commonplace among them. Typical Tajiks who don't sport "Turanid" looks often resemble singers Valy and Sadriddin, as well as this gentleman: https://en.wikipedia.org/wiki/Taqiyah_(cap)#/media/File:Young_Tajikistani_man_with_guitar.jpg"

True.

"In Afghanistan, Pashtuns definitely have more East/North-Euro looking individuals, but I'm not sure if that's due to their steppe ancestry being considerably higher or more so because of bottlenecking dynamics."

How so?

Roy King said...

@Davidski,
I think that part of Luca's legacy was an obsession with superficial phenotypes which he inherited from 19th and early 20th century human biologists. Population geneticists like Luca Cavalli-Sforza and his students to this day continue to reflect this preoccupation. There is so much more genetic data that would be interesting to butress your case: geneotype assignments to the DRD2 markers associated with impulsivity and the oxytocin receptor markers associated with empathy or its lack. As a psychiatrist, I see behavioral genetics as a potentially important influence on cultural history. It's just that many are stuck on the external: skin color, hair and eye color as if those phenotypes carried some behavioral meaning.

Seinundzeit said...

Matt,

"Re: Sindhi and the Pakistani Pashtun/Pathan, Lazairidis estimated them at like, what, 5% difference in Steppe_EMBA ancestry - https://1.bp.blogspot.com/-9k5UilpTSEA/V2TqxrBdxgI/AAAAAAAAFCE/awWnaVwRo6cPa2xjxEKuNKRLfmLIM5f3gCPcB/s1600/MODEL.png?"

Although this is a distinct topic, I would note that Sindhis vs the Pashtuns originally sampled by the HGDP do differ by around 10%-15%, in terms of Steppe_EMBA.

In addition, when looking at total ancestry, the Sindhi have a far weaker Iran_Chl affinity, a much stronger Iran_N affinity, and 50% more ASI (when I have more time, I'll post some models).

All of that divergence in streams of ancestry, taken together, naturally leads one to expect some consistent phenotypic divergence, a notion which pans out IRL (Sindhis are a much, much darker population in comparison to Pashtuns. The difference in unexposed skin pigmentation is probably equivalent to the differentiation seen between Estonians and Sicilians. Differences in exposed skin color are even starker, and no one would ever suggest that Sindhis have higher percentages of light eye and hair color).

"Yeah, like I say, there's some noise there in some of the panels, so wouldn't fixate on any particular individual population..."

That being said, I would fixate on one particular individual population, the Kalasha, as their frequencies are quite off from what we've seen with the HGDP data.

If my memory serves me right, the Kalash basically cluster near Sardinians and Druze in terms of pigmentation genetics.

The HGDP Pashtuns are fairly close. In fact, they'd be in the same position, if one excludes samples like HGDP00239, HGDP00237, HGDP00232, and HGDP00258.

These samples belong to Qasibghar, and are thus Pashto-speaking ethnic Punjabis (for a European analog, it's like including Polish Jews in a genetic analysis of ethnic Polish people).

Anyway, my main thought was that South Central Asians are far more depigmented than what one would expect for mixtures between Steppe_EMBA and Iran_N.

Again, using the Kalash as an example, they have slightly/moderately higher frequencies of derived alleles for depigmentation-implicated SNPs than currently sampled Yamnaya/Afanesevo, and massively higher frequencies of derived alleles for depigmentation-implicated SNPs than currently sampled Iran_N.

So, either it's the case that some Steppe_EMBA populations were much more depigmented in comparison to Yamnaya/Afansevo, or selection is at work.

We'll find out soon enough, with those Swat Valley samples.

Rami,

Although David's modelling involves the East Iranian Sarmatians (which makes perfect sense), I should note (because I kinda enjoy reading your pissed/annoyed reactions) that I can usually get Pamiris to be Steppe_EMBA/Srubnaya_outlier + Eastern Scythian. For example:

50% Steppe_EMBA + 5.40% CHG + 3.10% Scythian_East
28.80% Iran_Chal

6.95% ASI

5.75% Xibo

distance=3.8000

So, basically around 60% Steppe-related (mostly EMBA).

This is based on 13 PCoA dimensions, using an Fst sheet David provided me, a while back.

I did include the Sarmatians, but this is what happens. Same story with Global_10 data.

Based on some things I've heard about the Swat valley samples, the Steppe_EMBA vs Steppe_MLBA conundrum for Central and South Asia is going to be figured out pretty soon.

Anthro Survey said...

@Shahanshah

"Greeks of today are the only ones who speak the Greek language as their native tongue, and thus are the true inheritors of ancient Greek civilization, whether you like it or not, mate"

When it comes down to this, they are more ancestral inheritors than cultural(they derive about 20% of their ancestry from post-AD400 migrations, from what I gather with uniparentals and nMonte).

In short, there were many intervening events between the Hellenic golden age and Modern Greece standing in the way cultural continuity there. Yes, the language was retained, but using that logic, we can say Romanians are the inheritors of "Roman civilization". In reality, Romanians are a cultural and ancestral amalgam of Slavs and Paleo-Balkanics, much like Macedonians, Albanians, Serbs, and Greeks for that matter. Their language contains many Slavic loanwords and, if in the absence of such words, expressions often follow Slavic structure. Traditional customs, dress, etc. are entirely Balkanic and no relation whatsoever to "proper" Latins like French or North Italians. Thankfully, Romanians are starting to recognize this today after a spell of Romanicist nationalism induced by French intellectuals.

From the medieval times onward, Latinate Westerners(deemed "Franks" by outsiders) came to be the true inheritors of Hellenism, for it is they who continued to hold Archimedes, etc. in high esteem and expounded on their work. The narrative of "1452 Greeks came and took Europe back out of a dark age" is bollocks. In 1444, Cardinal Bessarion described sending his son to Northern Italy to learn of Western technological practices. The Byzantine Archimedes palimpsest, a prayer written over an erased mathematical treatise by the polymath, speaks for itself, too.

Like aforementioned Romanians, continental Greeks are a Balkanesque people in most regards. There wouldn't be a huge cultural shock going from, say, Ioannina to Sarajevo. Modern Greeks began to identify strongly with Hellenic civilization after Lord Byron and other adventurers arrived to help with their war of independence in the early 1800s.

"The Pamiri language is not mutually intelligible with Tajik, and it is not a Western Iranic language."

Absolutely agree. No quarrel there---Pamiri along with other East Iranic languages spoken in those highlands is in the same family as Pashtun. I was referring to one of his now-corrected sentences, btw.

"In fact, many Iranians from the plateau were also referred to as Tajik by their Turkic overlords."

Oh, indeed. I jokingly asked a Western Iranian once if he was "Turk" or "Tajik" to which he unironically replied "Tajik". So, yes, it's still in the vocabulary but the term fell out of everyday usage in the West as time passed, remaining common in some Persian areas abutting Turkic communities on the plateau.

"If anything, the Turks were the ones shedding the most blood in most conflicts, with the obvious exception being the Mongol invasions."

Well, maybe one of those rank-n-file Turkish "chads" lost his grip and killed an urbanite in a fit of rage. (No question there were consequences for this handed down to him by his superiors! :D ) After all, Muslim Turkish communities were quite primitive and uncouth at the beginning, often consisting of pastoral yurt-dwellers living outside city walls. Later on, they became more literate and integrated into urban communities. That's the culprit behind ENA in Tajiks, as well as added Iran_Chl-like and reduced ENA ancestry in urban Uzbeks(corresponding to their less Turanid look).Turks integrated in a similar fashion in Western Iran.

Btw, In the furthest reaches of Central Asia like Taraz and Tarim regions, there was only a Sart "caste" and the Turk-Tajik dichotomy was not so relevant, IIRC.

Vara said...

@Shahanshah

"You are correct again, for the most part. However, your assertion that Middle Persian took hold among Sassanian noblemen in Khurasan is not well substantiated. I think you are forgetting the the Sassanid heartland was very much in western Iran, not in Khurasan. The Sassanids viewed their northeastern provinces as nothing more than buffer against Turkic and Iranian tribes of the Central Asian steppes. "

While Anthro Survey knows his stuff, he has what some would call bias towards Western Iran. However, in this case he is correct if you consider Transoxiana and Chorasmia as part of Khorasan even though they were not part of the Satrapy as they were ruled by their own kings and princes. It was the Sassanid governers that first spoke Middle Persian there even though the natives spoke Sogdian.

On the other hand, Eastern Iran was not a buffer state, as you claim. Most of the Younger Avestan texts that were written or edited in the Sassanian period is focused on that area and reclaiming it from the Xionites (Hyaonas in the Yasnas). Bactria in particular was the important region on the Silk Road.

Anthro Survey said...

@Shahanshah

"However, your assertion that Middle Persian took hold among Sassanian noblemen in Khurasan is not well substantiated. I think you are forgetting the the Sassanid heartland was very much in western Iran, not in Khurasan. The Sassanids viewed their northeastern provinces as nothing more than buffer against Turkic and Iranian tribes of the Central Asian steppes."

Oh, I don't disagree here and my rather succinct paragraph failed to convey this, if contradicting it somewhat. Let me digress by saying that Khorasan-westernIran relatinship is analogous to that of Central Europe and Lazio/Rome. Much like Khorasan was a staging ground for Iranization of the western plateau, so too was Central Europe a staging ground for the Italicizing of Italy. Up until, say, 700BC, there was a stream of peoples and metallurgical customs from the north. Enter a centralized Rome and the situation reverses(though imperial Roman legionnaires were outfitted in a Celtic manner and discarded early Roman hoplite/"Mediterranean" fashion). Like Fars/Elam during Achaemenid & Sassanid eras, Lazio became a focal point in Western Europe. Khorasan was analogous to the Rhine frontier and Gaul, I guess.

Yes, Middle Persian originated in Fars--then the Sassanid heartland---and spread to Khurasan's landed aristocrats during this era much like Latin spread to aristocrats in Rome's Gaulish provinces. People previously spoke east Iranic languages there.

I should note here that 'Khurasan' started acquiring importance by late Sassanian times, but really "broke out" during early Islamic times and maintained its dominant position until the disastrous Mongol incursions. (They really destroyed much infrastructure, demographics and caused a brain-drain.) It is there that New Persian developed and subsequently radiated outwards, alongside other new and exciting cultural trends. They became leaders in the Eastern Caiphate: be it the Barmakid family or Abu Muslim Khurasani or Ismail Samani himself. In Western Europe, former Gaulish provinces started to set the trends after Rome fell(and actually somewhat prior to that). It was no longer Rome or Capua but Milan, Cologne, Bologna, Oxford, Nuremberg, etc. just like it was no longer about Firuzabad or Taysfun but Balkh, Bukhara, Tus, Nishapur, etc. in Greater Iran.

Samuel Andrews said...

@Anthro Survey,
"In short, there were many intervening events between the Hellenic golden age and Modern Greece standing in the way cultural continuity there. Yes, the language was retained, but using that logic, we can say Romanians are the inheritors of "Roman civilization". "

Romania and Greece aren't comparable. Romans came to Romania, subjected the people and essentially forced them to speak Latin. Greeks didn't subject Greeks. Modern Greeks are Greeks. They are a continuation of the same ethnic group the classical Greeks belonged to. Yes they have some forigen admixture but that doesn't change the fact that they are Greeks.

Samuel Andrews said...

@Anthro Survey,

Maybe the relation between modern Greece and ancient Greeks is sort of comparable to the relation between Germania and modern Germany.

Things change, admixture with other people's occur, but the identity, the language, and for the most part genes in modern Greece goes back to classical Greeks. The same is even true for Coptic Christains and ancient Egyptians and more remotely fro Ashkenazi Jews and ancient jews.

Anthro Survey said...

@Shahanshah

"No, they range between 35 and 45% Steppe, closer to the low 40s."

Is that after we subtract out the steppe associated with Turkic admixture and normalize the remainder, though?

What I mean is I ballpark Turkic ancestry as ~50-50 ENA/Siberian-West Eurasian. The WestEurasian portion I take to be steppe-like for a couple of good reasons. First, they are believed to have "stormed out" of the eastern steppe, close to the Altai range, then well-visited and settled by our favorite kurgan folks. In fact, we have a late eastern Scythian genome(Pazryk) near Altai and it carries ~50-60% Siberian admixture. R1a-z93 is fairly common in Kyrgyz, too. Perhaps early Siberian Turkic-speaking groups mixed with Scythian populations and acquired their warrior horseman package, only to play a cruel joke on them.

Anyway, if a Tajik scores, say, 10% ENA/Siberian(above noise level, that is, since North Euros and steppe-rich Iranic groups get a bit due to overlap resulting from ANE in Siberians), I subtract out 10% from his other steppe and divide the remaining steppe by 80. Is it still close to 40% after that? Maybe I miscalculated things. :-)

"How so?"
In regards to what? The question of relative percentage of NorthEuro-looking folks in Pashtuns? Or selective and bottle-necking mechanisms operating on a few loci responsible for spiking such traits in some populations, like Nuristanis, relative to others with comparable steppe(sintashta-like?) ancestry?

Anthro Survey said...

@SamuelAndrews

That's fair---perhaps using Romanians as a counterexample was a bit extreme.

Again, I agree that most of their ancestry is probably local(pre AD400 or so). My 20% estimate is based on uniparentals, archaeological/historical data and nMonte modeling. They get about 10% Slav_Czech569/West_Ukranian-like ancestry when a similar scheme is applied to other Balkaners, but it's likely that it arrived bundled with intermediaries carrying some paleo-Balkanic ancestry from the north. Btw, recall also that large portions of Greek peninsula were speaking Slavic at one time, to be later re-Hellenized by Byzantine authorities. Northern Greece is full of such toponyms.

What we can say is that Modern Greeks were living in a state of historical/cultural amnesia by the time Lord Byron arrived or even when Norman knights were laying siege to Dyrrhachium. The general population, as well as religious and intellectual elites, ceased to identify themselves with their genetic ancestors who built the parthenon, etc. It was not a matter of mere Christianization, mind you, although its progressive entrenchment in Greece was correlated with losing those "Hellenic" aspects.

I think the example of Copts and Assyrians works, but would there would be a bigger cultural disparity between them and their ancestors 2000 years ago than between modern and ancient Greeks?

At any rate, the paradigm of the northern Greek border being some "rubicon" separating "Greek" and "Slavic" worlds doesn't hold water. Greeks share much more in the way of material and symbolic culture with other Balkaners(including those Albanians and Macedonians they berate so much) than with 'Italians' or anyone else, for that matter. Now, perhaps Ionian islanders and Cretans DO pull closer to "Italy" in this sense because Venetians dominated those islands. "The Balkan sprachbund" is a term that can be extended to more than just matters of language. Greece is certainly well within it.

Chad Rohlfsen said...

I think it is more than minimal. I'm not overly concerned with uniparentals. A neighboring and unsampled group may be the smoking gun. Just as one EMBA steppe group might be dominated by R1b, a neighbor has a lot of I2. They don't look any different. We have a lot of gaps in West Asia. I wouldn't doubt the chance of a group having a lot of uniparental matches with Barcin and Boncuklu, but being more like Iran autosomally.

What I'm talking about isn't to do with the steppes. I haven't even gotten there. I'm working on the farmers and I have found a lot of stuff. Some is nothing to do with Iran and was simply missed.

I'll show you when I'm done.

Chetan Vit said...

@Rob However, there is no evidence for any steppe-derived settlements or monuments or mounds in the East Balkans at this time.

"During the last quarter of the fifth millennium, Skelya pastoralists (of Sredny Stog II) invaded the Balkans. Widely distributed Skelya type burials with characteristic Skelya grave goods are intrusive in this area." (p.5, Parpola 2012)

He can't possibly be making this up can he?

And they needn't have immediately moved to Anatolia although that kind of movement is also not to be discounted. Later Yamna settlements did move directly into the Hungarian plain while skipping the cultures in between. we have candidates for intermediate culture between the Balkans and Anatolia which show steppe influence could have spoke early PIE related languages. Cernavoda And Ezero cultures.

@supernord R1b i not matter. Expansion PIE is not associated with R1b.

That is incorrect. Some subclades of R1b are indeed associated with the expansion of IE.

Aram said...

Imho there will be 2 migrations from Steppe to Anatolia. One via Caucasus and other via Balkanes little bit later. Both were IE. The Balkanian one affected mostly NW Anatolia and didn't go deep.

Rob said...

@ Chetan V

""During the last quarter of the fifth millennium, Skelya pastoralists (of Sredny Stog II) invaded the Balkans. Widely distributed Skelya type burials with characteristic Skelya grave goods are intrusive in this area." (p.5, Parpola 2012)
- - He can't possibly be making this up can he? "

Of course not. But did you pay attention ? :)

This is last q. of M5. parpola is actually just repeating Anthony's essay. Anthony argues these are intrusive invaders. That's his speculation. More like traders and prospectors, who seem to have co-existed at the time of Varna, literally there are 7 or so Skelya time monuments in lower Danube. And the distribution of these finds actually procede west to East, reaching Khvalynsk last, also Meshoko in the north Caucasus. So as Govedarica and Jeuneusse point out, this Gimbutas' "Wave I' flows in the opposite direction.

Then c. 4000 BC the Skelya phenomenon ends as does the Varna / KVK VI culture. It disappears with the Balkan collapse. There is no sudden appearance of more and more Skelya or skelya type finds anywhere, let alone in the east Balkans. In fact, south of the Danube river, there is not much to find at all for several hundred years. So there can be no invasion to speak of.

Even in the steppe, in place of the rich Skelya finds, there is a period of lower quality culture.

@ Aram

"Imho there will be 2 migrations from Steppe to Anatolia. One via Caucasus and other via Balkanes little bit later. Both were IE. The Balkanian one affected mostly NW Anatolia and didn't go deep."


Well, the Balkan one wasn't from the steppe. That's the point. It was from the Balkans and it occurred c. 4500 BC. when metal starts being used all of a sudden in Anatolia.

Aram said...

Rob

Do You think Kum4 represents the migration You mean?

Olympos I replied to You.

Rob said...

@ Aram

Kum IV is late on for that, but could be part of an ongoing process which began as early as 5000 BC. The collapse of Varna c. 4200 BC could be what broke the continuum for a while, with separation of proto- Anatolian.

Anthro Survey said...

I think some people erroneously assume that proto-Anatolian speakers must have reached Anatolia around 2000BC because that is roughly when they start to enter the spotlight with Hittite "badasses".

Perhaps proto-Anatolics entered (as some Sredny Stog derivatives) from the Balkans ~4000BC and their linguistic heirs "hibernated", co-existing with non-IE speakers there for the next couple of thousand years.

In other words, we really shouldn't expect to stumble upon any steppic Hittite genomes, but possibly elevated frequencies of R1b-pf7562(one of the more basal clades of m269).

Chetan Vit said...

Anthony argues these are intrusive invaders. That's his speculation. More like traders and prospectors, who seem to have co-existed at the time of Varna

The appearance of Skelya graves occurs together with the burning and abandonment of settlements in the Balkans which makes the possibility of an invasion more likely. Anthony makes the same point again in his 2013 paper. p.8

"The oldest archaeological evidence for a post-Neolithic migration from the Pontic-Caspian steppes into neighboring regions is a movement into southeastern Europe about 4400–4200 bc, linked chronologically and geographically with the sudden abandonment and burning of hundreds of tell settlements in the lower Danube valley and eastern Bulgaria about 4400–4200 bc, with associated rapid changes in pottery, metallurgy,mortuary customs, ritual figurines, and other behaviors (Fig. 2). During the same period Balkan copper bracelets, beads, and rings were obtained by small-scale steppe elites in the lower-Dnieper and middle Volga steppes."

And you are right. The distribution exists west to east because once the Skelya invaders seized the settlements and goods of the Balkan people, they traded them with their steppe neighbors and relatives to the east. So Balkan goods appear in the east later. And the destruction of these cultures meant that the production of valuable metal goods was cut off thus the impoverishment of the graves after the initial Skelya period for the next few centuries.

Anyway it's almost undeniable that this represented a steppe incursion rather than a peaceful integration.

Anthro Survey said...

@Rob

Of course, and that clade's position in the R1b-m269 tree is exactly why I think it may have been common in some of the earliest IE speakers in the Balkans and Ukraine---perhaps of proto-Anatolic languages. Something more downstream as L51, on the other hand, I see as being widespread in proto-CeltoItaloLigurian speakers, seen by some as derivatives of Western Yamna.

Why supernord can't understand R1b's relevance to IE at this point is beyond me.

Chetan Vit said...

@Rob I don't think I am wrong about there being a clear steppe incursion into the Balkans in the late fifth millennium BC. Guess we have to wait for more evidence.

"The steppe group were mutually antagonistic with each other too, and competitive, which is why there was constant turnover of lineages in the steppe and bludgenoning each other to death as well. "

That is such a blanket statement. I guess the steppe groups didn't hate each other always as much as they hated the non IE outsiders in the Balkans ;)

Rob said...

@ Chetan

As I said,the major cycle of violence in the lower Danube occurs in 4500 BC, before any Skelya, perhaps relating to the rise of coercive heriarchy in the Black Sea region..
And these events of abandonement and crisis all began even further south, in Anatolia. So the opposite of what is generally assumed.
Then, you're not piecing the other evidence from nearby. What came after 4000 BC ? Where did new centres rise ? If the Skelya people invaded the balkans - what did they achive - they neither settled it, nor did they come the new power.
Does the story add up ?

Rob said...

@ Chetan
"I don't think I am wrong about there being a clear steppe incursion into the Balkans in the late fifth millennium BC''

Have you ever seen a map of distribution types ?

Rob said...

@ Chetan

''I guess the steppe groups didn't hate each other always as much as they hated the non IE outsiders in the Balkans''

The nationalising of language occurred in the 19/20th century AD Chetan.
I'd educate yourself as to how group identity was manifest in the Copper Age.

Chetan Vit said...

@Rob They did settle. The later Cernavoda and Ezero cultures are fusions of native and indigenous elements.

"Have you ever seen a map of distribution types ?"

Haplogroup distribution maps? Yep

Rob said...

@ Chetan

You're not getting it .
The Ezero culture formed 9-700 years after the said incursions. 4200 vs 3300 BC.

Not haplogroups, the map of Skelya finds in the Balkans. Have you seen where they are exactly and how many there are

Chetan Vit said...

Now Rob, your statement about the steppe groups hating each other is pretty general and vague. I'm not even sure how I can offer a counter argument to that. Maybe some steppe groups hated some other steppe groups at some times, but such feelings of antagonism are difficult to discern in any archaeological sense. And mutual antagonism doesn't preclude trade relations. Two mutually antagonistic nations may still engage in trade relations in the modern world. I don't see any other explanation of Balkan artifacts in the Dniepr and Volga region.

Chetan Vit said...

@Rob "the map of Skelya finds in the Balkans. Have you seen where they are exactly and how many there are" NO can you link me to one?

Chetan Vit said...

"The later Cernavoda and Ezero cultures are fusions of native and indigenous elements."

*sorry steppe and indigenous elements.

Cernavoda culture (c. 4000–3200 BCE). The time seems right

Olympus Mons said...

...guys, thats the definition of religious confirmation bias.
Arriving populations that did nothing historically noteworthy in their homeland of the steppes and Balkans are the source of the most remarkable civilizations of bronze age?? Wtf do you all even notice how mad that sounds?
Didnt the Minoans and myceneeans genome tels ypu nothing? And Gac?

Rob said...

@ Chetan

" NO can you link me to one?"

So then your assertions are baseless

""The later Cernavoda and Ezero cultures are fusions of steppe and indigenous elements."

Did you just read that on Wikipedia ? Don't waste my time


Final remark

Khvalynsk: Q, R1a , R1x M269
Ezero : I2a1b, Yamnaya Bulgaria I2a1b

Chetan Vit said...

@Rob Why do you get so angry when something contradicts your views? I'm prepared to change my views if we obtain evidence in the future that contradicts steppe invasions into copper age Balkans.

But as of now, it definitely appears that Sredny Stog invaded the Balkans around 4000 BC, "steppified" the native cultures which produced the Cernavoda and Ezero cultures. That's what both Parpola (2012) and Anthony (2006 and 2013) concludes. And would you say Anthony is not an authority on the topic? Anthony was in on the 2013 genetics paper that pretty much established the kurgan steppe theory of IE origins. He knows what he is talking about.

Chetan Vit said...

@Rob The archaeological record is clear. As for the genetic, we need more than just 3 genomes to make any meaningful conclusions on the matter

Rob said...

@ Chetan
You're not contradicting me, you just don;t understand the details, despite my attemts to explain them to you. So, yes, Im out of patience.

ak2014b said...

Off-topic.

There's this this pre-print on Cucuteni-Tripolye from earlier in November. Has it already been discussed? Or maybe because it's just mtDNA, it didn't need its own discussion page? Still, if these are new genomes, it could result in full genome studies of these CT samples in future.

Analysis of ancient human mitochondrial DNA from Verteba Cave, Ukraine: insights into the origins and expansions of the Late Neolithic-Chalcolithic Cututeni-Tripolye Culture
Ken Wakabayashi, Ryan Schmidt, Takashi Gakuhari, Kae Koganebuchi, Motoyuki Ogawa, Jordan Karsten, Mykhailo Sokhatsky, Hiroki Oota
doi: https://doi.org/10.1101/217109

November 10, 2017.

Abstract
Background: The Eneolithic (~5,500 yrBP) site of Verteba Cave in Western Ukraine contains the largest collection of human skeletal remains associated with the archaeological Cucuteni-Tripolye Culture. Their subsistence economy is based largely on agro-pastoralism and had some of the largest and most dense settlement sites during the Middle Neolithic in all of Europe. To help understand the evolutionary history of the Tripolye people, we performed mtDNA analyses on ancient human remains excavated from several chambers within the cave. Results: Burials at Verteba Cave are largely commingled and secondary in nature. A total of 68 individual bone specimens were analyzed. Most of these specimens were found in association with well-defined Tripolye artifacts. We determined 28 mtDNA D-Loop (368 bp) sequences and defined 8 sequence types, belonging to haplogroups H, HV, W, K, and T. These results do not suggest continuity with local pre-Eneolithic peoples, but rather complete population replacement. We constructed maximum parsimonious networks from the data and generated population genetic statistics. Nucleotide diversity (π) is low among all sequence types and our network analysis indicates highly similar mtDNA sequence types for samples in chamber G3. Using different sample sizes due to the uncertainly in number of individuals (11, 28, or 15), we found Tajima's D statistic to vary. When all sequence types are included (11 or 28), we do not find a trend for demographic expansion (negative but not significantly different from zero); however, when only samples from Site 7 (peak occupation) are included, we find a significantly negative value, indicative of demographic expansion. Conclusions: Our results suggest individuals buried at Verteba Cave had overall low mtDNA diversity, most likely due to increased conflict among sedentary farmers and nomadic pastoralists to the East and North. Early Farmers tend to show demographic expansion. We find different signatures of demographic expansion for the Tripolye people that may be caused by existing population structure or the spatiotemporal nature of ancient data. Regardless, peoples of the Tripolye Culture are more closely related to early European farmers and lack genetic continuity with Mesolithic hunter-gatherers or pre-Eneolithic groups in Ukraine.

Rob said...

@ Chetan

"The archaeological record is clear. "

That's a some what bold and comedic statement from someone who's basis of "knowledge" on the Balkans and Anatolia is the writings of a Uralicist

" we need more than just 3 genomes to make any meaningful conclusions on the matter"

We've got more than 3. Certainly, plenty of I2a1b around before and after Yamnaya & Ezero

Davidski said...

@Olympus Mons

Obviously, GAC pastoralists got rolled over by Corded Ware pastoralists from the steppes. The ancient DNA doesn't lie.

Globular Amphora people starkly different from Yamnaya people

So what was the point you were trying to make about GAC and Minoans? Or haven't you learned anything from the ancient DNA yet?

Davidski said...

@Chad

Uniparental markers were only a small part of my argument.

Matt said...

Sein: Anyway, my main thought was that South Central Asians are far more depigmented than what one would expect for mixtures between Steppe_EMBA and Iran_N.

So, either it's the case that some Steppe_EMBA populations were much more depigmented in comparison to Yamnaya/Afansevo, or selection is at work.

I guess my main interest here is in the idea of whether Steppe_EMBA ancestry tends to be more associated today with an effect where (and this is where I'm paraphrasing the poster kony ITT) steppe ancestry brings in a low frequency of derived variants like SLC45A2 that are generally less common in pre-existing populations (unlike SLC24A5 variants). Then a selective shift happens in specifically these populations (and not others that didn't get the steppe ancestry).

So this is for context why I was really interested in the comparison of other populations at a similar latitude band (NW Africa, Levant) that don't have steppe ancestry or have much more minimal amounts, and in the general picture (and not so specifically in frequency in any particular pop).

I actually don't disagree that selection on SLC45A2 and HERC2 seems likely in the recent past on populations in West Eurasia and North Africa (results on Europe, Levant and North Africa all seen to agree as well) and that selection on SLC24A5 seems ubiquitous in all populations where the derived variant has been introduced (e.g. East Africa, though there are also signals for selection toward darker skin phenotypes in that region as well). That's a very sensible idea. Though obvs needs a full scientific review of allele frequencies and models.

The question that grabbed me ITT (which I probably did not communicate that well earlier) was really just whether any of this effect is linked to or enhanced by Steppe_EMBA ancestry specifically, or if these are more of a latitude based effect. Both within Europe for HERC2 at smaller levels of divergence in Steppe_EMBA ancestry and more generally for HERC2 and SLC45A2 within the broader West Eurasian context, it looks to me much more like both are effected by latitude* based selection than any balance of any particular BA ancestry.

*latitude based selection at least for HERC2 / OCA2 variants linked to eye colour fits with Mathieson's results for OCA2 for Euro HGs in "Genomic History of Southeastern Europe". SLC45A2 for HGs doesn't show such a clear gradient - if anything it's increasing West->East, but the Ukraine_HG shows low frequencies that same as Iron Gates and Latvia HG despite having more EHG ancestry than SHG...

Davidski said...

@Roy King

In spite of the panic from some of the crowd here, traits like pigmentation and height aren't controversial any more, at least not in western countries, and their genetic architecture is relatively simple. So they're a good way to learn how to infer traits from ancient DNA, and this is why they're such a big part of ancient DNA papers.

The stuff you're talking about is very difficult and complicated, not only genetically, but also politically. I mean, imagine comparing the genetic origins and predisposition to impulsivity and empathy in Europeans and Africans.

Rob said...

@ Dave

"Obviously, GAC pastoralists got rolled over by Corded Ware pastoralists from the steppes. The ancient DNA doesn't lie"

The GAC continues as late as 2400BC
It only really ends with the arrival of BB, as they seem to have been "shut off" from the developments which led to BA cultures like Trziniec, deemed to develop out of the Unetice - BB- late CWC interactions

Davidski said...

@Rob

GAC got rolled over by Corded Ware. That's why modern day Northern and Eastern Europeans cluster with Corded Ware, not GAC.

Shaikorth said...

@Rob
Anthony and Parpola are specialists in Indo-European and South Asian studies respectively, both AFAIK support AIT and at least in Parpola's case Dravidian IVC/Harappa.

Roy King said...

@Davidski
"The stuff you're talking about is very difficult and complicated, not only genetically, but also politically. I mean, imagine comparing the genetic origins and predisposition to impulsivity and empathy in Europeans and Africans."

I was not interested in modern comparisons, but ancient comparison. For example what are the calls of the DRD2 and Oxytocin receptor variants of the Yamnaya, Corded Ware, and other Steppe populations compared to the European and Anatolian Neolithic populations. It could be important for understanding the cultural/genetic interactions in the Bronze Age transformation of Western Europe.
It terms of political concerns, I can definitely say skin color remains highly laden in Western societies--trust me.

ak2014b said...

Samuel Andrews said...
@aniasi,
"we still don't know where the West Eurasian light skin gene arose."

Let's be honest, when people talk about 'West Eurasian' light skin they're really talking about European light skin.



It's not clear which of SLC24A5, SLC45A2 or TYR count as the gene that aniasi alludes to. But SLC24A5 was not exclusively European, as the derived allele at rs1426654 already occurred among most of the Iran Neolithic samples of Broushaki et al 2016: only AH2 was homozygous for the ancestral allele, whereas both WC1 and AH1 were heterozygous at SLC24A5, and AH4 was homozygous derived for the same. The Iran Neolithic sample, GD13a, of the Jones et al 2016 paper was also homozygous derived. And an Iran Chalcolithic individual of the Lazaridis paper, sample I1661 from the Seh Gabi site of the 6th millennium BC, was homozygous derived too for SLC24A5.

The same Iran Chalcolithic sample I1661 was furthermore heterozygous at rs16891982, so for SLC45A2. Two of the Chalcolithic Armenian samples (Areni cave, 4330-4000 calBCE) were heterozygous at rs16891982, so they carried a derived allele for SLC45A2.

So the SLC45A2 mutation was floating around for quite some time in Asia. And SLC45A2 does not appear to be exclusive to Europe either, or at least did not exclusively spread with IE.

And brought this up before: Iran Neolithic sample WC1 from 7,455-7,082 calBCE of Broushaki et al 2016's dataset was heterozygous at rs12913832 (HERC2), so it had a derived allele for blue eyes. And it was heterozygous for most or possibly all of the remaining locations for the blue eye haplotype that's common in Europeans.

So rather than any of those mutations (HERC2, SLC24A5 and SLC45A2), which would have entered S/SC Asia early on with the Iran Neolithic and Iran Chalcolithic admixture, perhaps the mutation at rs12821256 for blonde hair will be the way to track the IE newcomers to S/SC Asia in the ancient DNA? However, I thought the Yamna samples so far were dark haired? What about those from later, like Poltavka and Sintashta or Srubna?

Olympus Mons said...

They cluster with CWC because BB screw and snatch their women. Those CWC womem were the only thing that got rolled over. Big time!

Olympus Mons said...

@roy King
That would be an amazing work.
Are you activingly pursuing that? Where we read about it?

Davidski said...

@Olympus Mons

They cluster with CWC because BB screw and snatch their women. Those CWC womem were the only thing that got rolled over. Big time!

No obviously they cluster with CWC because CWC rolled over GAC, and CWC spread their DNA, including the male marker R1a-M417, from the steppes across Northern and Eastern Europe.

It's hilarious that you still can't grasp the simple fact that the data in the GAC paper support the steppe hypothesis.

Rob said...

@ Dave

Nonsense
Most modern R1a is Slavic
CWC got removed from Central Europe

Davidski said...

@Rob

Nonsense
Most modern R1a is Slavic
CWC got removed from Central Europe


Most modern R1a in Europe is Corded Ware, via derived cultures like Trzciniec in Eastern Europe, and other derived cultures in Scandinavia.

Most modern R1a in the world is either Corded Ware or something very similar that ended up in South Asia.

Deal with it: Corded Ware rolled over GAC, and that's why GAC is usually dated to 2,800 BC latest.

ak2014b said...

If blonde hair originated in ANE, since Afontova Gora-3 is to have had the mutation at rs12821256 and it hasn't (yet) been found in WHG and EEF, then is it generally expected that there would be no ANE anywhere in South Central or South Asia before the Bronze Age invasions from the steppes? For Europe it seems to be the case that ANE only dates to EHG and thereafter the Bronge Age steppe invasions into the rest of Europe. However, is it necessary that there would be no ANE in Asia anywhere until the IE speakers were to have brought it? Or can ANE be expected to have had an early spread in various parts of Asia, such as among the ancestors of Mongols, but perhaps also other parts? For instance, wasn't the Mesolithic Iran sample thought to have had a bit of ANE or something EHG-like, and EHG contains ANE? (Unless that's not really ANE, like it's not really EHG.)

At anthrogenica, Kurd was recently discussing his findings on how several ANE and other ancient samples were skewed more towards Europe than they may have been, due to modern contamination. Someone asked him whether the derived allele for blonde hair in AG3 might have been on account of that contamination too. It seemed like Kurd's response implies that the mutation for blonde hair in ANE sample AG3 could be part of the contamination. What is then the oldest sample after AG3 that had the blonde hair mutation?

There have been pictures of Burusho, Kalash and some other South Central Asians having blondish hair. The blonde hair may have various sources, depending on whether or not ANE had some presence in those parts of Asia before IE spread there. Does anyone have similar examples of blonde hair among South Asians? (Outside of persons who colour their hair or have recent European or other steppe-derived admixture dating from the European colonial or Turkic periods)

Samuel Andrews said...

ak2014b,
"There's this this pre-print on Cucuteni-Tripolye from earlier in November. Has it already been discussed? Or maybe because it's just mtDNA, it didn't need its own discussion page?"

Becasuse the mtDNA results in that study are very few and low coverage. The authors wrongly predicited the mHG for several samples. For example quite a few are obviously Ukraine HG derived U4 (16356c mutation) but labeled HV12b1. Plus, a whole paper shouldn't be dedicated to a handful of low coverage mtDNA results.

ak2014b said...

Kurd's posted new findings today, about skin colour phenotype of certain steppe samples he analysed. He's found that his diploid versions of 3 published genomes, one each of EHG, Yamna and Srubna samples, all confirm the published genotype and phenotype for them, and concludes that the individuals would have been white in colour. Kurd's not posted the results for TYR for his diploid genomes, but his conclusions are based on several other SNPs.

His 3 steppe genomes are all both AA at rs1426654 for SLC24A5, and heterozygous as CG at rs16891982 for SLC45A2, just like the above mentioned Seh Gabi Iran Chalcolithic sample I1661, which was similarly AA for SLC24A5 and CG for SLC45A2. (I think Europeans tend to be mostly homozygous derived for SLC45A2, slightly more so in northern Europe.)

This part of Kurd's conclusions looks unconvincing:

"The lightest pigmented appears to be EHG and the darkest Srubna I0232
For Yamna their graph shows 100% derived for rs1426654 which basically means AA, which is like all Europeans and the vast majority of W Asians, namely light. Again this SNP contributes alot to pigmentation. S Asians and Africans have the ancestral; either AC or CC, which is a big reason they are dark."


I think he's saying that AA for SLC24A5 results in greater depigmentation than having derived alleles for SLC45A2, which is to contribute less to depigmentation.

However, Razib Khan is a South Asian and owns to being homozygous derived at rs1426654 (SLC24A5) as per his caption to his own photo. Maybe some subtlety about the region is lost to me, but Razib looks as dark as the South Asians that Kurd speaks of contrastively with West Asians and Europeans, despite Razib being homozygous derived. Razib's nowhere near the "European light skin" as Samuel Andrews phrases it, nor what Kurd seems to have in mind when he speaks of West Asians being lighter on account of being SLC24A5 homozygous derived than those with one or both ancestral alleles.

So is SLC24A5 really more relevant for significant depigmentation than SLC45A2, as Kurd thinks? That seems rather unlikely to me, because if SLC24A5 really is the most effective of the two, then wouldn't the 3 steppe samples be only very slightly lighter than Razib? But that doesn't match with Kurd's actual results for the phenotypes of the steppe samples, which turned out actually white with very high to high probability going from EHG to Yamna to Srubna. Is SLC45A2 or perhaps TYR more effective in depigmentation, or combinations of mutations?

Rob said...

@ Dave

"Most modern R1a in Europe is Corded Ware, via derived cultures like Trzciniec in Eastern Europe,"

Exactly . After CWC got replaced in Central Europe, it had to re-expand later
But this is nothing new- it's the same of continual flux which had been gowing on ainbce before CWC and continued long after
Only difference is that CWC was so autosomally divergent, thus so notable to statisticians who otherwise disregard other phenomena.

"Deal with it: Corded Ware rolled over GAC, and that's why "

CWC is coterminous with GAC. And what replaced both was BB. (See pic below)


"GAC is usually dated to 2,800 BC latest."

You're off by 800 years

http://imgur.com/5cvVSrI

Is it too hard to look at things soberly and in an informed, balanced manner ?

ak2014b said...

@Samuel Andrews
"Becasuse the mtDNA results in that study are very few and low coverage. The authors wrongly predicited the mHG for several samples. For example quite a few are obviously Ukraine HG derived U4 (16356c mutation) but labeled HV12b1."

So that would explain why I couldn't find a discussion page for it. I must have missed out on criticism against the paper in the comment sections, but I'm still catching up.

Did you correct the authors about the mistakes you found, since it's just a pre-print?

Hopefully these CT samples will get a full genome analysis in a future paper. Maybe Planck, Allentoft or Reich labs will acquire them.

Davidski said...

@Rob

Is it too hard to look at things soberly and in an informed, balanced manner?

You need to finally accept that GAC and other "Old Europeans" got rolled over by Corded Ware and other steppe derived groups, and that's how the modern-day European gene pool formed.

Samuel Andrews said...

@ak2014b,

The reason geneticists looked for and discovered these West Eurasian light skin mutations in the first place was to discover reasons why Europeans have light skin. The main reason ancient DNA studies test them is to measure selection levels in Europe and the origin of uniform light skin in Europe.

This is why I said "Let's be honest, when people talk about 'West Eurasian' light skin they're really talking about European light skin."

I have nothing against you or anyone else here. Other than Rob, all you guys seem cool. However, I've got to say I find some of the "West Eurasian light skin" discussion politically correct. Sometimes it seems like a way to non-racealize Europeans, to make "whiteness" a Middle Eastern as well as a European thing. One way or another something unique to Europe relating to light skin is going on.

I'm open to whatever the facts around this topic are. Maybe the same depigmentation mutations have existed in both Europe and the Middle East for millenia and it just so happens theyw ere selected for strongly in Europe which is why there's such a big differnce in pigmentation between the two regions today.

Rob said...

^ and where did the I2, I1 and G2 in Wielbark come from ?

Kurd Dgk said...

I'm not sure when and where the hype about Yamna being dark pigmented started, but the evidence does not support it. Yamna is 100% derived at the important pigment locus rs1426654, which is highly correlated with dark pigmentation. In other words, Yamna is AA at that locus, just like all Europeans, and 95%+ W Asians. How do I know that? I happen to have DIPLOID genotyped Yamna I0231, EHG I0061, and some Steppe MLBA at over 2 million positions. So not only do I have its diploid genotype at rs1426654, but also at about 30 other less important pigmentation SNPs, such as rs16891982.

Based on 9 pigmentation SNPs (not just the 2 used by many papers) Yamna is about the same skin color as modern Caucasians and Ukrainians ( I suppose a shade or 2 or 3 darker white than most N Europeans)

By contrast Iran-N-WC1 has an ancestral allele at rs1426654, it is AG. So based on 5 diploid pigment SNPs I have for WC1 his prediction is brown, darker than the vast majority of Kurds and Iranians ( some of who are actually pretty dark). About the color of dark Punjabis.

Therefore, we can rule out Pamiris or fair SC Asians getting their pigmentation from Iran-N related. However, their pigmentation is consistent with Yamna and MLBA.

Details at http://www.anthrogenica.com/showthread.php?3433-Waves-of-migration-into-South-Asia&p=319142#post319142

Olympus Mons said...

@Rob,
Can you explain a bit the " Yes there was a large steppe arrival, added with extra WHG..."?

Olympus Mons said...

@Davidski,
The only places you have R1A above 50% is like Ukraine and poland. I mean, its not like R1b. most places were "abundant" its not over 30%. Actually at most, R1a stayed "home" in Ukraine and poland and its incursion west was just to lend some women to BB males moving up the Rhine and Elbe- A friend call it the Eastern brothel of the Bell beakers males.
Even chalcolithic R1a areas in germany that were full R1a and CWC, later do show full of I2, didnt it?



Rob said...

@ Shaikorth

Parpola and Anthony hardly qualify as specialists for sEE and Anatolia

Olympus Mons said...

@Rob,
yes, it seems WHG got an upheaval. But was it bronze, Iron age, later? Its also fascinating.

Ric Hern said...

My "uninformed" guess would be that Unetice I2a migrated from the Alps....

Shaikorth said...

@Rob
They're not regional specialists but what they propose is around the usual steppe theory. I've not seen conclusive evidence for Hittite Balkans route specifically but something did come from steppe to SEE.

Onto another matter: I1 is almost certainly from a founder effect around the North Germany-South Scandinavia range, the recent MRCA being some CW or BB-assimilated local. All I1 we know of descends from this. Before this it was uncommon, enough so that I'll be very surprised if high frequencies of I1 will be found in any Neolithic or Mesolithic culture anywhere. I2 another matter ofc.

Matt said...

@Kurd, few questions:

a) Did you put the genotypes through a model like Beleza's to try and predict outcomes? As variants at some SNPs have more weight than others.

b) Do you think that pigmentation phenotypes of current SCA populations, NW Africans, Levant make sense as likely to be linear products of genotypes around at time of Yamnaya, or selection afterwards?

(Obviously there's a difficulty here in that we don't actually have any genotypes of populations for SCA at the time of Yamnaya, closest proxy being Iran_N from thousands of years earlier and another place and time.)

Certainly, there does still seem like selection within Europe to explain frequencies of at SLC45A2 (rs16891982) and at HERC/OCA2, even if other pig. SNPs are more similar between Yamnaya / Europe_LNBA and present day Europeans. Considering both the Bell Beaker paper and Mathieson's last selection paper and the results from the paper on the Bronze Age Levant for example (and even earlier papers like this - http://dienekes.blogspot.co.uk/2014/03/dark-pigmentation-of-eneolithic-and.html). (Also, for the British Isles, the strong selective signal of pigmentation associated derived variant at IRF4 that essentially doesn't exist at high frequencies even in other NC Europeans).

c) Have you done any of this generally across other ancient samples (e.g. Iberia Chal, etc.) as well as Iran_N and Steppe samples, as comparison?

Kurd Dgk said...

@Matt,

Excellent questions.

a) Yes, so the basis of the model is you basically take a cohort of say 1000 individuals --> put individuals into a few groups classified by skin color --> count the number of individuals having certain genotypes at selected SNPs (pigment related) --> create a probability matrix based on those %

b) Absolutely, because those SNPs are part of coding genes, and you may know the mutation prevention mechanisms are more robust than with intergenic neutral alleles.

c) Plan to do that with higher coverage samples (>3 X avg read depth). Should have something for Kotias, Satsurbila, WHGs. Stuttgart by tomorrow


I am also working on something that uses 36 SNPs for prediction.

Shaikorth said...

@Matt

Kurd's results also suggest that pseudo-haploidity has a noticeable effect on genome-wide relatedness to various populations. In the example case of western Scythian it looks like Tatars and West/South-Central Asians show relatively lower IBS in pseudo-diploid results, while eg. Swedes and Estonians show higher. Ukrainians and Belarusians look stable.

http://www.eurasiandna.com/wp-content/uploads/2017/11/IBS-Atlas-Scythian-Means-Normalized.jpg

http://www.eurasiandna.com/wp-content/uploads/2017/11/IBS-Haploid-Scythian-Means-Normalized.jpg

Kurd Dgk said...

@ Shaikorth

Just to clarify:

1- The genomes I diploid genotyped are more accurate than the published pseudo-haploid ones for IBS work for obvious reasons. Humans are diploid at a substantial portion of their genome. Thus, We are looking at 1 wrong allele at potentially 30% of the positions in Human Origins for pseudo-haploids

2- on the same page you referenced, I have graphs for modern human contamination. The Scythian is higher than the rest at 8%. One of these daysI I will update the page with analysis of the Scythian genome using sites that show ancient DNA damage pattern only (thus backing out any positions that may have modern human contamination) as per the PMDS technique published by Skoglund.
I have already done that and the results that it was significantly Tatar shifted which is consistent with the IBS graph corresponding to my diploid genotyped Scythian

Mike the Jedi said...

@Kurd

You do hear many say Yamnaya had slightly darker skin on average than modern Europeans, but I hear it more often claimed that they were mostly dark-haired and brown-eyed, at least on the western steppe. Would you also dispute that characterization?

@Anthro Survey

Of course modern Greeks are not culturally identical to the classical ones-- the golden age was 2500 years ago. So what? They still predominantly descend from the ancients and speak the Greek language. If some medieval or pre-Independence Greeks were ignorant of this connection or didn't care, well, such is life.

Philhellenism led to the Greeks calling themselves Hellenes again (as opposed to Rhomaioi), but that's just one more in a long line of name changes for this age-old people. Similarity to other Balkaners is also no surprise; centuries of Byzantine and Ottoman rule will do that. I maintain the biggest break in continuity with the classical era happened in late antiquity with the rapid Christianization of the Roman Empire. Thankfully the Enlightenment also led to better attitudes towards women, though male chauvinism is still a problem in Greece.

Matt said...

@Kurd, thanks for answering. Sorry, is b) "Absolutely" (selection afterwards) or "Absolutely" (no selection needed)?

@Shaikorth, that's interesting. There wasn't such an obvious effect in the IBS scores Davidski ran off for various other CA+BA Europeans+Steppe before, which I would assume are pseudo-haploid. What noise there was I attributed to being a "real phenomena" rather than loss of information. But that doesn't mean there wasn't such an effect there and information hasn't been lost from pseudo-haploidity rather than representing the true diploid situation. It would be good to see a set of full lists of populations and see what happens (if that kind of approach is at do-able).

(I guess it's hard to for me gauge what the overall effect in the absolute IBS is compared to the rank effect though, since this is normalized, though of course for many purposes only rank effect matters).

Kurd Dgk said...

@ Matt

No selection needed

@ Mike the Jedi

No dispute on hair color

Kurd Dgk said...

@ Mike

Or eyes

@ Shaikorth

1- 1 wrong allele at every hetrozygous position ~30% of the positions in HO

Davidski said...

@Olympus Mons

yes, it seems WHG got an upheaval. But was it bronze, Iron age, later? Its also fascinating.

Pfft...

Yeah, during the Great Depression of the 1930s.

Davidski said...

@Olympus Mons

Actually at most, R1a stayed "home" in Ukraine and Poland.

During the Late Neolithic/Bronze Age R1a-M417 spread west across most of Northern Europe and east across most of Central Asia, as far as western China and India.

That's the opposite of staying home you fruitcake.

Do yourself and everyone else a huge favor and find another hobby. You're not cut out for this.

Bob Floy said...


@Olympus Mons

^^^He's telling the truth, you know.

Davidski said...

@All

Olympus Mons is now banned.

I gave this idiot too many chances.

Chetan Vit said...

@Rob (if you are still reading this)

I2a seems to have been present among the steppe groups too.

There was a post about this here http://eurogenes.blogspot.in/2017/05/steppe-invaders-in-bronze-age-balkans.html

So where from comes the conclusion that the I2a in Balkan finds didn't have a steppe origin? When the burial type and grave goods clearly indicate steppe origin.

Rob said...

@ Chetan

Yes of course I'm aware of it, but if probe deeper into the matter you'd have realised that I2a2a1b is present in Neolithic Bulgaria and middle Neolithic East Hungary, it similarly appears in Mariupol Ukraine, and then appears also in Yamnaya Bulgaria, Ezero Bulgaria as well as Ulan IV (late yamnaya - early Catacomb). I2a2 as a whole appears to originate in east-central European Palaeolithic.
So as I suggested in a preceding paragraph (which I'll repeat for you, this represents a western component moving east, present since the Mesolithic, and unlike the Khvalynskain dodo's, remained present in into the Bronze Age.
Why ? Because pastoralism began in the northern Balkans and then moved to the steppe.


Ebizur said...

Di Cristofaro et al. (2013)

Hazara (HGDP Pakistani Hazara)
1/25 I2a2a-M223

Esfahan province, Iran
1/42 I2a2a-M223

Khorasan province, Iran
1/20 I2a2a-M223

South Iran
1/9 I2a2-M436(xI2a2a-M223)
1/9 I2a2a-M223

Hazara (Bamyan province, Afghanistan)
1/69 I-M258(xI2-M438) [Is this result erroneous, or has a member of Y-DNA haplogroup I1 somehow ended up among the Hazara in Bamyan?]
1/69 I2a2a-M223

Tajik (Balkh province, Afghanistan)
2/54 I2-M438(xI2a1b-M423, I2a2-M436)

Tajik (Takhar province, Afghanistan)
1/35 I2a1b-M423

Mongol (NW Mongolia)
1/97 I2a2-M436(xI2a2a-M223)

Y-DNA haplogroup I actually is not an exclusively European haplogroup. It appears that I2a2-M436 and its subclade, I2a2a-M223, may be relatively common among members of Y-DNA haplogroup I in hypothetically steppe-influenced populations in Asia. Those clades have been found with greatest frequency among British and other NW European peoples if I remember correctly.

Rob said...

There is also I2c in Caucasus groups,

Ebizur said...

Also cf. id:YF02073 on YFull with a reported origin in Iran. This individual's Y-DNA has been classified as belonging to I-Y3670(xY7240, L1229), a subclade of I2a2a-M223.

I-M223 formed 17,400 [95% CI 15,700 <-> 19,200] ybp, TMRCA 12,100 [95% CI 10,900 <-> 13,300] ybp

I-Y3670 formed 10,400 [95% CI 9,100 <-> 11,900] ybp, TMRCA 9,200 [95% CI 8,100 <-> 10,300] ybp

This Iranian individual does not belong to a basal outgroup to European members of I-M223, but he does not yet have any close relatives on the YFull tree, either.

Davidski said...

@Chetan Vit

I2a seems to have been present among the steppe groups too.

It is, and I2a2a1b1b appears to be a marker of population movements from the steppe to the Balkans (rather than the other way around).

I2a2a1b1b is found in a forager sample on the North Pontic steppe but not in any foragers from the Balkans or Hungary, so it looks native to the steppe.

Outside of the steppe, it's found in samples with steppe ancestry on the autosomes and from steppe-like burials.

Obviously, the fact that I2, I2a2a and even I2a2a1b look like they have their origins outside of the steppe has no relevance here, since the focus is on I2a2a1b1b, the ancestral lineage of which probably entered the steppe during the Upper Paleolithic, with foragers not pastoralists.

I2a2a1b1b could well turn out to be one of the three main Y-haplogroup lineages associated with Eneolithic/Bronze Age steppe expansions, alongside R1a-M417 and R1b-M269.

Matt said...

@Sein, by the way, if you're still following this comment thread, off topic, but had a few comments about the PCoA+Fst method based on a quick experiment around with it. (All this will be of interest to only me and you ;) ).

Before I told you, if you recall (may not!), that following this procedure:

1A) setting the value Exponent=2 on PCoA (effectively ^2) allowed the preservation of Fst distances, such that when you
1B) take the PCoA output and check Euclidean distance matrix,

Returns the same matrix you put in. That is, it's reversible. And if you follow the same procedure setting Exponent=1, it's not.

Now, this actually seems true for quite simple matrices, but doesn't happen with the big matrices of 180+ populations that represent a big sampling of world populations. The distances end up extremely wrong and inflated / deflated at fine scales.

However, I have found that if you:

2A) run PCoA in Past3 using Exponent=1, then
2B) check Euclidean distances on the output, and then
2C) √2 that output distance matrix

then that actually does *perfectly* reproduce the original matrix that was input (as in error less than 0.001 Fst).

(Does this make sense?)

This kind of leaves me in a bit of a quandry. On the one hand, using Exponent=1 does preserve the original input Fst matrix totally... but you have to use a odd procedure to do so, while Exponent=2 fails to preserve most of the distances, but does do OK for some distances between very distant populations without having to go through an extra procedure. So not sure which one is better for nMonte now. Wish I knew more math to know if there was a simple transformation that I could apply to get the best qualities of both.

Still, one interesting thing this does allow you to do though, is predict Fsts for populations. So if you had a dimensional for your PCoA Exponent=1 for an in-theory ASI zombie, then you could simply insert this between steps 2A) and 2B) above, and you'd have an idea of predicted consistent Fst scores for ASI. Or the same for UHG or Basal Eurasian or something.

Another thing I did with it also was create zombie rows for 80%English:20%WHG and 90%English:10%WHG . Basically, I wanted to check if this zombie had lowest Fsts with Lithuanians and Latvians, as WHG rich populations. Likewise 75%Spanish:25%Steppe_EMBA.

Funnily enough though, the output showed that these zombies would actually be predicted to still have their lowest Fst with English (for English+WHG zombies) and Spanish (for Spanish+Steppe_EMBA). Morever predicted Fst was quite low even for 90%English:10%WHG and English (0.001), 80%English:20%WHG and English (0.003), while 75%Spanish:25%Steppe_EMBA (0.002). And Fst from Yoruba did not increase from adding in WHG... So this actually adds a bit more skepticism for me as to whether differences in ancient proportions actually drive Fsts in Europe, present day, at the fine-scale.

Possibly more of interest to you is that I tried the same thing with "(Pashtun-25%Chamar)/75%". Effectively exaggerating away from Pashtun in an anti-Chamar (e.g. ASI) direction. The resulting Fst did still come out closest to the Pashtun populations at predicted Fst 0.002, and closer than the other closest real relatives Sindhi (0.003) and Brahmin_UP (0.005).

I don't know totally what to make of this, and if this is reflecting anything real or just an artefact of method, but assuming it is representing something real, but again I guess would not be surprised if Fst measures found various ancient populations of ANI and ASI origin, but particularly ANI, from Pakistan or nearby places tend to have surprisingly low levels of divergence to present day people from Pakistan.

(This will all be gibberish to anyone else reading!)

Rob said...

@ David

"It is, and I2a2a1b1b appears to be a marker of population movements from the steppe to the Balkans (rather than the other way around)."

I'd say it is "native" to the Romanian - west Ukraine region as a whole. It, or its predecessor, appears in the Mariupol culture for the first time. (It'll be clearer when more data comes out and further pulled out).
But yes, it obviously linked to Yamnaya, in the West, and places east of the steppe, eg modern Kurds. The same group is found in the Ezero culture in Balkans, and its phylogenetic equivalents or older and related clades are also found in the Balkans/ Hungary, because they'd intermixed down the rivers since the Neolithic

Chad Rohlfsen said...

@David,

I know parental markers are a small part. WRT the haplotype sharing, why not be concerned that is way off too if they're both Iran and Natufians are almost misaligned as much as ENA?

Chad Rohlfsen said...

-they're... not awake and alert here

Davidski said...

@Rob

It, or its predecessor, appears in the Mariupol culture for the first time. (It'll be clearer when more data comes out and further pulled out).

That I2a2a1b1b sample is from a set of Ukraine_Neolithic foragers native to the North Pontic steppe with no links to the Balkans or the Carpathian Basin.

I know that they have no links to the Balkans or Carpathian Basin because they lack Anatolian ancestry, which was widespread in the Balkans and Carpathian Basin at the time and much earlier.

So there's no reason at all to posit that I2a2a1b1b is native to the Balkans and a signal of farmers or pastoralists moving from the Balkans to the steppe. In fact, this contradicts the data that we have, and I can't see more data overturning the current picture.


Davidski said...

@Chad

Haplotype data deals with sets of linked markers. So if the alignments were seriously off for, say, Iran_N WC1, the results would look like garbage, with strangely low haplotype sharing between WC1 and modern samples.

But of course WC1 shows very high haplotype sharing with Indus Valley samples, so much so, that it is preferred by them ahead of the Anatolians. It's impossible for this to be the result of bad alignments, and it obviously makes good sense.

So if WC1 is behaving like it should relative to modern-day samples, then all of the other Iran_N individuals are too, more or less, because they're very similar to WC1.

That's not to say that bad alignments aren't causing strange outcomes to varying degrees in formal stats analyses, but the point I'm making is that overall, it appears that we do have a solid idea of what the Iran_N population is.

Matt said...

By the way, all, I was randomly browsing around earlier and spotted something which provoked some thoughts around "For instance, what was the ratio of men to women amongst these migrants (to South Central Asia)?"

That, is, back in 2009 (way, waaay back in the time before adna) Reich produced this Supplementary Table 5 and Figure S7 https://imgur.com/a/R0Hl6 from his paper "Reconstructing Indian Population history" (https://www.nature.com/articles/nature08365).

This seems to show that the simple ANI:ASI fraction, as estimated by the brutally simple f4 ratio method, shows no pattern of enrichment on the X over the autosomes. (There's also either no correlation between autosome fraction ANI and autosome X difference. Or a slight positive one, where e.g. higher autosome fraction ANI correlates with a slight excess of estimated ANI on the X; that is what would be slightly higher female bias if anything. But Reich says this is likely due to standard error.).

Though we've learned a lot since then (e.g. ANI not a single population, essentially in very simple terms), and we could get some fine detail perhaps out of using a better simulated proxy for "ANI", it doesn't seem like this would change that there still ought to be a general acceleration of ANI estimated ancestry on autosome relative to X, which doesn't seem to be the case? (And in general terms we shouldn't need ancient dna to detect this if it's strong and obvious and we can estimate the cline using modern populations...)

I wouldn't mind knowing why I'm wrong though, because this result doesn't seem to show up in later papers on potential sex biased demography in South Central Asia (e.g. https://bmcevolbiol.biomedcentral.com/articles/10.1186/s12862-017-0936-9).

Samuel Andrews said...

@Rob,

Apprently Mesolithic Steppe and Balkans shared lots of Y DNA. Both had loads of R1b1a and I2a2a1b. Considering the Yamnaya founder effect R1b1a-Z2103 likely decends from the Mesolithic Steppe, shouldn't their I2a2a1b also be descended from the Mesolithic Steppe?

I don't have a strong position on this issue. If not for the loads of I2a2a from Ukraine HGs I'd probably say Yamnaya's I2a2a1b is from SE Europe. That hasn't changed your view because you're bias.

For a long time, you have been pushing a narrative which down plays the role of Steppe migration in exchange for Balkan migration. I remember last year you flirted with the idea R1b M269 orignated in Balkan farmers or something like that. Nobody is seeing evidence of people movement from the Balkans into the Steppe. I2a2a1b isn't strong evidence considering it was probably already in deep in the Steppe in the Mesolithic.

Chad Rohlfsen said...

@David,

Sure, we have a good idea how they look, but if 20% of the sample is incorrect, it will certainly affect the sharing with moderns. CHG is much, much more aligned than Iran, so I think that deserves serious consideration in the haplotype sharing. CHG didn't exist as a population during the Neolithic, bit was more of a trace to minor admixture deal like WHG. It is much more likely the Iran/CHG stuff, the great majority of it, is from SE Anatolia and Northern Mesopotamia.

Rob said...

Sam
Stop insinuating ulterior motives if you don't understand what I'm saying.
The steppe was liable to dynamic events, ecology and local population shifts. The demographic hubs for steppe lineages were adjacent areas around the steppe like the south forest zone (R1a) and upland regions to the west (I2a2), perhaps Crimea and the Caucasus etc

Rob said...

Anyhow, the earliest I2a2 is in south Germany and Iron Gates. It appears jn Dnieper- Ukraine c. 5000 BC.
What's to convince

Davidski said...

@Rob

The demographic hub for R1a was the western steppe, probably around the Sea of Azov, not the south forest zone.

This is not my opinion, we actually now have direct evidence from ancient DNA showing it, with R1a continuity on the western steppe from the Mesolithic and throughout the metal ages.

R1a never left the steppe, and indeed expanded from the steppe multiple times, mostly as R1a-M417.

Early indications are that the steppe was also the demographic hub for I2a2a1b1b, and this is why Yamnaya_Bulgaria looks the way it does, and is actually very, very similar to the R1a-M417 sample from the steppe in Alexandria.

You obviously would prefer it if this wasn't the case, but with the ancient DNA evidence available, you have to leave your personal preferences aside and seriously consider the possibility that it is the case.

Seinundzeit said...

Matt,

"However, I have found that if you:

2A) run PCoA in Past3 using Exponent=1, then
2B) check Euclidean distances on the output, and then
2C) √2 that output distance matrix

then that actually does *perfectly* reproduce the original matrix that was input (as in error less than 0.001 Fst).

... So if you had a dimensional for your PCoA Exponent=1 for an in-theory ASI zombie, then you could simply insert this between steps 2A) and 2B) above, and you'd have an idea of predicted consistent Fst scores for ASI. Or the same for UHG or Basal Eurasian or something."

Very interesting; for the fun of it, I'm going to give this a spin with my ASI zombie (thanks for sharing this idea!).

"All this will be of interest to only me and you ;)"

Haha; quite frankly, your thoughts on this should interest everyone, since this is a methodology that does bear some good fruit.

Davidski said...

@Chad

It's very difficult for me to accept that the Iran_N samples are so way off, considering how sensibly they behave in a wide range of analyses based on unliked genome-wide data, linked genome-wide data (haplotypes), and uniparental markers.

Arza said...

@ Matt
Another thing I did with it also was create zombie rows for 80%English:20%WHG and 90%English:10%WHG . Basically, I wanted to check if this zombie had lowest Fsts with Lithuanians and Latvians, as WHG rich populations. Likewise 75%Spanish:25%Steppe_EMBA.

Funnily enough though, the output showed that these zombies would actually be predicted to still have their lowest Fst with English (for English+WHG zombies)


Here are distances to 15 samples nearest to the point between English_Kent and Loschbour, that in turn is closest to Latvians:

in Global 10 (22.93% English_Kent)

Vatya:RISE479 0.00942943817076
Corded_Ware_Germany:I1540 0.0123677207047
Estonian 0.0144324494531
Latvian 0.0152484999307
Lithuanian 0.0153070270257
Hungary_BA:I1502 0.0155160376875
Russian_Smolensk 0.0156221421274
Polish 0.0165590318067
Belarusian 0.0176547433443
England_Anglo-Saxon:HS3-I0161 0.0183895736264
Icelandic 0.0185434347044
Nordic_LN:RISE98 0.0186573178525
Unetice:I0804 0.0189307447971
England_Roman:3DT16 0.019177768543
Swedish 0.0192816466902

in averaged DoHA (22.55% English_Kent)

Welzin_BA 0.0156557106506
Norwegian 0.0167121134513
Orcadian 0.018856491283
Swedish 0.020265904709
Irish 0.0205053979915
Ireland_EBA 0.0210310122225
Scottish 0.022250978483
English_Cornwall 0.0223334363433
Dutch 0.0228058999104
English_Kent 0.023
Bell_Beaker_Germany 0.0230937426223
Unetice 0.0231703672357
Bell_Beaker_Czech 0.0239610348702
Alberstedt_LN 0.0241294624436
French 0.0241345668436

IMHO Narva will be a little bit different kind of WHG (which, in the end, won't be a cluster but a cline between two closely related HG populations) and this difference is already visible on PC6 in DoHA, in the mentioned Fst sheet and in few other places (but not in Global 10).

That's why when you're moving towards "standard" WHG, at the same time you're moving away from the Balts.

If you want to see how it works, make a DoHA PC1/PC6 plot and see where English, Baltic and WHG samples are.

If you want to see where probably will be the missing (W)HG population placed, make a PC2/PC6 plot and draw two lines - one from Ranchot88 via Loschbour and further away, and the second from AG3 via Karelia_HG:I0061. And take a look where Hyper-Latvians, RISE568 and HG samples from the ANE-WHG "cline" are.

https://s6.postimg.org/8rr88wl5t/WHGcline.png

Kurd said...

@ Chad

Out of curiosity what exactly are you referring to with "alignment issues"? do you have a reference to it?

Reason I ask is that WC1 is one of the highest coverage ancients out there with an average read depth of 10. Broushaki used the BWA aligner which is used by probably over 95% of the folks out there me included. A read that is very off for whatever reason is simply not mapped. My guess is that only about 50% of the reads make it through mapping.

Like most others Broushaki then used GATK's BQSR (Base Quality Score Recalibration) to downgrade quality scores for bases for which there is no known variation in dbSNP. So if for ex a base at a certain position was a T, and Hg19 showed an A and dbSNP did not show that the particular position was polymorphic, then the assumption is that the T is a sequencing artifact, and the quality score for that base is downgraded.

So, IMO mapping mismatches after BQSR with the 10X WC1 should be negligible, that is why I was curious if you had read something regarding WC1.

IMO, incorrect genotypes would be a greater issue, but here again Broushaki tested the robustness of his pipeline (I can't be sure, but I believe he used something similar to ATLAS which is what I use) for the diploid call in WC1 by downsampling Ust which is 42X down to 10X and running it through his pipeline to see whether the diploid genotypes of the downsampled Ust agreed with the published 42X Ust, and they did in 99.1% of the calls. The only difference was that the downsampled 10X Ust had lower hetrozygosity than the 42X Ust, but that is to be expected with 10X vs 42X. I've run across that also. What ends up happening is some of the hetro sites are called homo, but 1 of the alleles is still correct.

«Oldest ‹Older   1 – 200 of 228   Newer› Newest»