This is yet another teaser for the upcoming Corded Ware/Yamnaya paper from the Reich lab. Sadly, it doesn't mention Y-chromosome haplogroups, so perhaps the authors are going to tackle this issue later. However, check out what they say about the German and Spanish farmers being of the same stock, and the resurgence of hunter-gatherer ancestry in Western Europe after the early Neolithic. Fascinating stuff.
Ancient DNA points to the Eurasian steppe as a proximate source for Indo-European migrations into Europe
David Reich and Nick Patterson
Abstract: We generated genome-wide data from 65 Europeans who lived between 8,000-3,000 years ago by enriching ancient DNA libraries for a target set of about 390,000 single nucleotide polymorphisms. This strategy decreases the sequencing required to obtain genome-wide data from ancient DNA samples by around 1000-fold, allowing us to study an order of magnitude more individuals than previous studies and to obtain new insights about the past. We show that in western Europe, the farmers of both Germany and Spain >7,000 years ago were descended from a common ancestral stock. These farmers did not replace the earlier hunter-gatherers, but continued to mix with them, leading to a resurgence of hunter-gatherer ancestry in both Germany and Spain ~1,000-2,000 years later. In eastern Europe, the hunter-gatherers of Russia >7,000 years ago were distinct from those of the west, having an increased affinity to a ~24,000 year old individual from Siberia, but this affinity was reduced by ~5,000 years ago in the Yamnaya steppe pastoralists because of admixture with a population of Near Eastern ancestry. Western and Eastern Europe collided ~4,500 years ago with the appearance of the Corded Ware people in Central Europe, who derived at least two thirds of their ancestry from an eastern population closely related to the Yamnaya. The evidence for mass migration into Europe thousands of years after the arrival of agriculture, in combination with linguistic and archaeological data, makes a compelling case for the steppe as a proximate source for the spread of Indo-European languages into Europe.
Source: INA Kolloquium Ws 2014/15
Update 11/02/2015: Massive migration from the steppe is a source for Indo-European languages in Europe (Haak et al. 2015 preprint) .
Lets see the claims......
ReplyDelete''In eastern Europe, the hunter-gatherers of Russia >7,000 years ago were distinct from those of the west, having an increased affinity to a ~24,000 year old individual from Siberia, but this affinity was reduced by ~5,000 years ago in the Yamnaya steppe pastoralists because of admixture with a population of Near Eastern ancestry. Western and Eastern Europe collided ~4,500 years ago with the appearance of the Corded Ware people in Central Europe, who derived at least two thirds of their ancestry from an eastern population closely related to the Yamnaya. ''
ReplyDeleteWhats new on that!!! Where are the Y-DNA DATA for gods sake!!....
Weren't you claiming recently that someone from Iran brought ANE to Eastern Europe during the Neolithic?
ReplyDeleteThat didn't work out, did it?
@David
ReplyDeleteNo i said of the ineffectiveness of ANE as it is very old and inconclusive for Asia at least! And to the Fact that we can't differentiate the non-IE ANE ancestry in Europe(E Europe Mostly) BTW the Near Eastern type population was from Iran....
I wish we would get a more clear picture of how this "mass migration into Europe thousands of years after the arrival of agriculture" affected the other areas of Europe besides the North and East.
ReplyDelete@ZeGrammarNazi and Others
ReplyDeleteIt will all depend on the Periodic presentation of Y-DNA of E European Area ONE thing I am 100% sure that Samara etc aDNA will provide YHG-I from the Earliest Levels.....
Before this gets bogged down by Nirjhar's posts
ReplyDelete"a compelling case for the steppe as a proximate source for the spread of Indo-European languages into Europe."
Very exact / cautious phrasing there.
"proximate" (perhaps pointedly not "ultimate") and "into Europe" (not generally). They may only be hedging their bets, or may have found something that has caused them to adopt this phrasing.
Interesting to me they found a resurgence in HG ancestry in Spain as well and no mention of any divergence between the Neolithic Farmers in Spain and Germany, despite mtdna divergence between the two being reportedly substantial (and indicating significant drift) according to one Skoglund's newest.
If a resurgence of HG ancestry in Spain is notable (how it compares to say the TRB) I wonder how much freedom Yamnaya would have to contribute to Spain population without shifting Spain "north" or "west" of where Spain actually is, without later Iron Age population flows from an opposite direction.
Finally, the number of hunters in Western Europe may have decreased with the arrival of IE,and not the contrary
ReplyDelete@Matt
ReplyDelete''Before this gets bogged down by Nirjhar's posts''
It will not happen Sorry:)
I'm not sure on the Y-DNA haplogroups of the earliest Samara peoples, but if Y-DNA's R and Q are linked to ANE ancestry and ANE was actually reduced over time in the Samara region due to contact with farmers from the Near East, R1a may have been present in Samara from very early on.
ReplyDeleteAlso, Y-DNA I2, which has been found commonly in Mesolithic European remains, seems to have no clear affinity to ANE ancestry and makes linking Y-DNA I lineages with early Samara inhabitants questionable, at best.
''"proximate" (perhaps pointedly not "ultimate") and "into Europe" (not generally). They may only be hedging their bets, or may have found something that has caused them to adopt this phrasing.''
ReplyDeleteIts obvious that they have presumed the Steppe Theory as the Basal anyway their main documentation so far we can see is the spread of ANE in Europe from Post-Neolithic period which is logical and reasonable...
''R1a may have been present in Samara from very early on''
ReplyDeleteThat's the Cornerstone Question but there are many other factors one is Asia which's aDNA is Vastly lacking from Important Archaeological sites....
''Also, Y-DNA I2, which has been found commonly in Mesolithic European remains, seems to have no clear affinity to ANE ancestry and makes linking Y-DNA I lineages with early Samara inhabitants questionable, at best.''
They Say-
'''In eastern Europe, the hunter-gatherers of Russia >7,000 years ago were distinct from those of the west, having an increased affinity to a ~24,000 year old individual from Siberia,''
The Word Is Increased Affinity NOT total Affinity which means With ANE-R1a the I-WHG type ancestry should also be expected....
All but one Scandinavian was I, so it is probably there as well. WHG goes pretty far to the East. It's probably in the minority.
ReplyDeleteAs Matt said, they may be leaving various routes open.
Yep but the I will be there Keep that in Mind but Chad Along with R1a,I which other Y-DNA's do you expect R1b,E,G2?
ReplyDeleteWe may find some form of Y-DNA C in Samara remains. La Brana and Kostenki both had C haplogroups, as well as some of the Neolithic remains from Hungary and I believe one of the later Andronovo guys was C(xC3), too.
ReplyDeleteYes that is possible ZeG good suggestion:).
ReplyDeleteNow we have prove in ancient genomes native/HG ancestry gradually rose in Neolithic-Copper age Spain, Germany, and Hungary. All present-day Euros fit as mostly being a mixture of those genomes. That may be a suprise to people several years ago who either supported Palaeolithic continuation or Neolithic-Iron age replacement.
ReplyDeleteI'll be very surprised if Reich doesn't suggest that Ajv58-like people contributed ancestry to north Europeans, because they obviously did.
I agree krefter but what is your opinion with the high WHG type ancestry that Indians show? They also have high ANE.
ReplyDelete"I agree krefter but what is your opinion with the high WHG type ancestry that Indians show? They also have high ANE."
ReplyDeleteI haven't looked into that. My impression from skimming through West Eurasia K8 spreadsheet is that WHG doesn't show any patterns in Asia, and that it's pretty much non-existent there.
Maybe it pops up in some Indians and is non-existent in others. It might be real ancestry or noise.
I tend to think Yamna-type people with 30-40% WHG brought Indo Iranian languages to Asia, and so I do expect to see WHG in Indo Iranian speakers.
I think the Near eastern component is absorbing WHG ancestry Asians got from Yamna-type people, and that Siberian ancestry in central Asia is erasing signs of Yamna ancestry, by making them much more ANE than WHG.
There very well may be significant Yamna-type ancestry in Asia.
But the Corded ware were the most successful Indo Europeans in erasing native gene pools.
''There very well may be significant Yamna-type ancestry in Asia.''
ReplyDeleteWithout aDNA from Asia that is meaning less Krefter.
''I tend to think Yamna-type people with 30-40% WHG brought Indo Iranian languages to Asia, and so I do expect to see WHG in Indo Iranian speakers. ''
I disagree no such mass movement can be detected from there to SC Asia and Iran on Evidence basis.
''But the Corded ware were the most successful Indo Europeans in erasing native gene pools.''
That is proven again on the basis ofaDNA and there are also Archaeological correlation isn't there Krefter?
"I disagree no such mass movement can be detected from there to SC Asia and Iran on Evidence basis."
ReplyDeleteThat's why I said I tend to think. It could be a totally different story in SC Asia. The IEs who arrived in SC I'm sure were very different from the Yamna, but probably still had Yamna-type ancestry.
It's very confusing why Yamna genomes from Samara which is pretty far east had so much WHG(maybe our estimations were wrong and they had under 30% WHG), when Asians have so little WHG.
It's a pitty they went the chip route and didn't do whole genome sequencing. IMO, if they had tested Y-DNA, the wording on their abstract wouldn't be so wishy-washy.
ReplyDelete@Krefter
ReplyDelete'' It could be a totally different story in SC Asia. The IEs who arrived in SC I'm sure were very different from the Yamna, but probably still had Yamna-type ancestry.''
I tentatively agree though i am quite sure that N Iran was the Seed Of PIE from where it dispersed...
''It's very confusing why Yamna genomes from Samara which is pretty far east had so much WHG(maybe our estimations were wrong and they had under 30% WHG), when Asians have so little WHG.''
PIE didn't had much WHG type ancestry that is the best proposal.
@RR
''It's a pitty they went the chip route and didn't do whole genome sequencing. IMO, if they had tested Y-DNA, the wording on their abstract wouldn't be so wishy-washy.''
Yup!:).
Now keep the discussion on i meet you guys in 9 Hours.....
"It's a pitty they went the chip route and didn't do whole genome sequencing. IMO, if they had tested Y-DNA, the wording on their abstract wouldn't be so wishy-washy."
ReplyDeleteWhat? Autosomal tells more than parental. Anyways, we know what Neolithic and WHG Y DNA was like, we know Yamna will probably be mostly R1a, and that EHG will probably be P(R and Q), C, and I.
"These farmers did not replace the earlier hunter-gatherers, but continued to mix with them, leading to a resurgence of hunter-gatherer ancestry in both Germany and Spain ~1,000-2,000 years later".
ReplyDeleteExactly as I expected. The resurgence must be attributed to Megalithic (and BB) expansion.
"In eastern Europe, the hunter-gatherers of Russia >7,000 years ago were distinct from those of the west, having an increased affinity to a ~24,000 year old individual from Siberia"...
Again exactly as I expected.
"... but this affinity was reduced by ~5,000 years ago in the Yamnaya steppe pastoralists because of admixture with a population of Near Eastern ancestry".
This is interestingly novel. Is it maybe when the R1a people arrived from Iran/Turkey (as per Underhill)?
Archaeologically speaking, we know little of the exact origins of the PIE or early Kurgan people. Samara has only been dug up to the Neolithic layer, before it is a mystery.
In any case it seems to imply that IE-speaking Kurgan peoples, carriers of ANE into Europe, were also carrying an unknown fraction of both WHG-like and EEF-like genomes (although surely subtly different from those Western references), rendering the analysis even more complicated.
"Corded Ware people in Central Europe (...) derived at least two thirds of their ancestry from an eastern population closely related to the Yamnaya".
That's a figure. And a big one indeed! Wonder how Corded compares with their Eastern BB successors, my impression is that they were not fully related, as BB probably gathered rather the pre-IE substrate, even if under IE language and still coalescing identity clues.
"It's very confusing why Yamna genomes from Samara which is pretty far east had so much WHG(maybe our estimations were wrong and they had under 30% WHG), when Asians have so little WHG".
ReplyDeleteThat's because both Western and Eastern pre-Neolithic Europeans descended from the same original founder populations, associated to Aurignacian and Gravettian cultures. Soon they diverged, with the Eastern group admixing with paleo-Siberians (different from East Asians and rather akin by origin to Europeans, West Asians and maybe Indians, as well as partly to Native Americans), but they still retained a "European aboriginal" affinity, detected here as WHG.
It's important to understand that:
1. Earliest UP or "Aurignacoid" expanded from probably West Asia (Iran?, Uzbekistan?) to Europe, Altai, West Asia and probably much of NE Africa, as well as maybe India as well (not too clear how earliest UP procceded in the southern arch yet).
2. Already in Europe, Aurignacian expanded from Central Europe in West and East directions c. 41 Ka BP.
3. Later, also in Europe, a second wave (maybe of West Asian origin) known as Gravettian and associated with the famous "Crô-Magnon" man, spread in the same fashion from Central Europe, reaching as far as Central Siberia (Mal'ta for example belongs to Gravettian culture). This happened after c. 32 Ka BP.
4. After that, Western and Eastern Europe evolved separately. So the affinity between Eastern and Western paleo-Europeans must be attributed to steps 1, 2 and 3, and shouldn't be any more recent.
Maju, according to hand written notes from ASHG 2014 (from Razib?), Late Bronze Age Europeans (presumably Unetice and Bell Beaker) plot very close to Corded Ware peoples.
ReplyDeleteSo, Eastern BB were likely similar to Corded Ware groups, but with a bit more ancestry from Old European groups.
Sorry, that was supposed to be Early Bronze Age Europeans; not Late.
ReplyDeleteThe more I hear about this paper the more disappointing I think it will be.
ReplyDeleteIt is all vague, inaccurate, sloppy...
In the last sentence of the abstract quoted in the post, at least they could have written North Eastern Europe, instead of Europe (2 times). Just to gain some credibility.
Anyway no one doubts that the steppe people expanded to NE Europe. We all know that for decades (Balto-Slavic languages, R1a, Kurgans,...). Nothing to see there.
At least they do write something coherent that will inform some people:
"These farmers [EEF] did not replace the earlier hunter-gatherers, but continued to mix with them, leading to a resurgence of hunter-gatherer ancestry in both Germany and Spain ~1,000-2,000 years later."
Though of course "resurgence" is a very wrong and sloppy word to use in that context. The HG did not come back from the dead. They were there all along. And they (obviously) had more HG ancestry before mixing than after.
I do hope the final paper is much better written. But better to keep the expectations low...
"Y-DNA's R and Q are linked to ANE ancestry"...
ReplyDeleteNot necessarily in an exclusive form. Actually we only know that R(xR1,R2) was present in Mal'ta. Full stop.
We can infer that Q1 was already present there (or had been earlier) because of its association with Native Americans and its necessary origins near Iran (where it's less common but much more basally diverse).
Similarly we can infer that R did not originate in Siberia but further South and we have absolutely no reason to associate R1 (nor R2, nor R1a, nor R1b) with the paleo-Siberians of Mal'ta and Afontova Gora (ANE). That's a conclusion that has no evidence whatsoever.
In fact basal diversity strongly suggests that R and R1 arose in the NW of the Indian subcontinent, while R1a and R1b did in the Fertile Crescent. We also know with some certainty (elaborating on Underhill's data) that the spread of R1b is much older than that of R1a, which could be even as late as Neolithic.
"Late Bronze Age Europeans (presumably Unetice and Bell Beaker)"
ReplyDeleteLate Bronze Age are Urnfields, i.e. proto-Celts and possibly also proto-Italics and proto-Germanics. At most they could be talking of their predecessors of the Tumuli culture (actually Middle Bronze Age but whatever) and never of Bell Beaker (late Chalcolithic, not yet Bronze).
Interesting bit anyhow.
"Interesting to me they found a resurgence in HG ancestry in Spain"...
ReplyDeleteThat should be obvious. Modern Spaniards are more WHG-like than the individual of El Mirador (likely local EEF proxy), who was close to modern Italians (particularly Tuscans) per a previous Swedish thesis.
From Razib:
ReplyDelete"But, I am nearly 100% sure that R1a1a coalesces to a period more recently than 10,000 years ago in the past. The reason is that I saw some posters using whole genome sequences from the Y chromosome at ASHG. These should be a more precise estimates because of the enormous marker set of more slowly mutating SNPs, and they too arrived at a relatively recent period for the last common ancestor of these common male lineages. In fact, if I recall correctly the divergence between R1b and R1a dates to ~10,000 years before the present in these studies, so R1a must have a much more recent coalescence."
It is a bit early to say where and when R1 lineages diverged. Perhaps the information Razib is referring to will become public before too long and we will have a clearer picture.
Maju, you will see that I meant to type Early Bronze Age in previous post. To be even more precise, the group was labeled "Late Neolithic/Early Bronze Age." (LNE/EBA)
ReplyDeleteYes, Grammar, I didn't read that. My bad.
ReplyDeleteA problem detectable here in mtDNA (see here) is that while true BB groups had massive mtDNA H, "fake" ones like Kromsdorf or post-BB ones like Unetice had exactly zero. Pooling them is surely erroneous, while only using some of the samples (say Unetice ones) is also misleading.
Neither Unetice nor (presumably CW-derived) Kromsdorf can explain modern levels of mtDNA H in Germans, which are ~40%. You need BB and whatever preceded it (Megalithism, Funnelbeaker) and almost certainly the origin of all that excess mtDNA H is to the West, not the East.
"4. After that, Western and Eastern Europe evolved separately. So the affinity between Eastern and Western paleo-Europeans must be attributed to steps 1, 2 and 3, and shouldn't be any more recent."
ReplyDeleteThanks for the archaeologically-minded info.
K-14(http://anthropology.net/2014/11/07/kostenki-14-a-36000-year-old-european/), suggests that WHG-type ancestry in East Europe and west Asia arrived later than 30,000 years ago. This is because he was neither, and Loschbour is more related to Stuttgart than to K-14.
Mesolithic(even bronze age and modern) east European's U5 was mostly or solely U5a, suggesting a very old introduction of WHG into east Europe.
More (weak)evidence that WHG-type ancestry in east Europe is old but younger than Aurignacian, is two FMS U5* aka pre-modern U5(xU5a'b) samples from 30k< Gravettian in central Europe.
ANE and WHG might not be clean cut ancestral components, but they are differnt, and we can confidentially say Mesolithic Russians were a mix of east and west Upper Palaeolithic north Eurasians, who had been living in that part of the world for over 30,000 years.
I agree, Maju. I believe the majority of mtDNA H found in Western and Central Europe today is derived from BB groups that expanded from the West of Europe, likely the Iberian peninsula.
ReplyDeleteI also do not think Kromsdorf was a "true" BB site, but likely influenced by BB. The mtDNA looks more like steppe populations than the typical BB site, which is usually >80 mtDNA H. Also, the grave goods were lacking. Two beakers and BB-style burial positions was about all that could be used to label it "Beaker". No v-shaped button, wrist guards, etc.
Very lacking.
I believe Reich told someone (Jean M?)that at least one mtDNA H was found in hunter-gatherer remains from Samara, but I doubt these Eastern groups are responsible for the presence of the majority of H lineages in Central and Eastern Europe.
Central and Western Europe*. Geez, I can't think or type correctly today, it seems.
ReplyDeleteMaju, I think you might like this source for ancient mtDNA.
ReplyDeletehttp://www.anthrogenica.com/showthread.php?3701-The-new-way-to-look-at-Ancient-mtDNA
Maps, haplotypes, frequency charts, etc.
I can't find the haplotypes of most ancient Iberian mtDNA from recent studies, because of pay walls and other obstacles. The results from old studies don't make sense. There's a load of R*(xR0, most basal U clades, JT) and L3*(xall west Eurasian N and R* clades).
Also, isn't it possible there was natural selection for mtDNA H? mtDNA frequencies of H, JT, U, etc. don't matter very much to me. Mesolithic and Neolithic Euros had very different mtDNA frequencies than modern ones but still are the main ancestors of modern ones.
Most Euros have 40-75% Mesolithic blood, but only 15-25% Mesolithic(confirmed)-derived mtDNA.
Okay, firstly, I find it very difficult to believe that the Near Eastern farmers who lowered the levels of ANE on the steppe by ~5,000 YBP (but probably appeared there much earlier during the Neolithic) also brought R1a to Eastern Europe.
ReplyDeleteWhat I think happened is what we already know happened in Central Europe, which is that is male hunter-gatherers often mated with female farmers. In other words, R1a looks to me like an eastern hunter-gatherer lineage that expanded during the Chalcolithic with a subset of the mixed hunter-gatherer/farmer population from the Russian steppe. It arrived in the Near East with ANE and the Indo-Europeans.
Also, I find it hard to believe that there's no WHG in Central Asia, considering that this is what the K8 results show, and also that Tajiks have higher affinity to Loschbour than Armenians do (so their WHG affinity can't be mediated entirely via farmer gene flow from the Near East).
https://drive.google.com/file/d/0B9o3EYTdM8lQMkNod1ZNZldvUHc/view?usp=sharing
"In eastern Europe, the hunter-gatherers of Russia ... The evidence for mass migration into Europe"
ReplyDeleteIf the Eastern HG were long in Russia, then why do the authors say that there is evidence for "mass migration into Europe"? Why could then EHGs not have just spread from eastern Europe into the West?
"These farmers did not replace the earlier hunter-gatherers, but continued to mix with them, leading to a resurgence of hunter-gatherer ancestry in both Germany and Spain ~1,000-2,000 years later."
Surely a resurgence of HGs would have depended on those who did not mix.
The claim that Neos did not replace HGs seems false. They took over most of the land.
The authors seem to have a preset
agenda of mass migration into Europe and mixing. Maybe they are just PC?
When I read the line "mass migration into Europe" in the context of the entire abstract I actually see "mass migration from the Russian steppe deeper into Europe".
ReplyDeleteAlso, even if the farmers took over most of the land (which doesn't sound right to me, because only some of the land in Europe at the time was suitable for the early Neolithic farming package), that doesn't mean pure hunter-gatherers didn't manage to hang on for over a thousand years on the peripheries of the Neolithic settlements. It's these people who would have mixed with the farmers very gradually, as both groups became more familiar with each other and less hostile.
As a result, hunter-gatherer ancestry would rise among the farmer population, and thus hunter-gatherer ancestry would increase as part of the total ancestry in Europe, because farmers had the demographic advantage.
@Krefter: Kostenki-14 is certainly pre-Gravettian (not sure if he was part of Aurignacian). What we need are samples from Gravettian or Epigravettian Eastern Europe.
ReplyDelete→ http://forwhattheywereweare.blogspot.com/2012/02/gravettian-culture-in-central-europe.html
Per Svoboda 2007, it would seem that Gravettian is an intrusive wave from West Asia (allegedly related to Amharian) that truly replaced the previous Aurignacian layer, at least partly and maybe in different degrees in different areas.
I know well that these kind of analyses are often controversial but the fact that such a (partial?, patchy?) replacement is possible and even likely underlines the need to sequence more recent samples than K14, which is even older than Mal'ta by 12,000 years! It's plausible and even likely that Gravettian substantially altered the genetic pool of paleo-Europeans.
We cannot discard later reshuffling of the European populations and their genetic pool (interesting question) but it seems most unlikely that they affected both Western and Eastern Europe for all I know.
"I can't find the haplotypes of most ancient Iberian mtDNA from recent studies, because of pay walls and other obstacles".
→ http://forwhattheywereweare.blogspot.com/2012/06/mtdna-h-found-in-epipaleolithic-basques.html (linked: freely accessible doctoral thesis in French)
→ http://forwhattheywereweare.blogspot.com/2012/04/ancient-mitochondrial-dna-from-basque.html (linked: open access paper, Hervella personally confirmed to me the methodology used)
Both studies evidence the presence of notable frequencies of mtDNA H in the pre-Neolithic Cantabrian strip. To that add the sequences of Chandler 2005 from Portugal (hard to find but last time I could find it at Zilhao's personal page, being co-author), which is a conference "paper" only but still the best we have re. the Neolithic transition in Portugal. All that, as well as other indirect evidence (Skoglund's Swedish Megalithic samples being near identical to modern Basques, Iberians, South French) point to a Western origin of the bulk of the Chalcolithic rearrangement we see in Germany particularly, be it Megalithic or Bell Beaker (both original from Iberia, although with secondary centers further North).
(continues)
(continued)
ReplyDeleteI can't but agree with the sad state of old aDNA studies, which often only sequenced the HVS-I region, which is totally inconclusive for mtDNA haplogroup H. But much of the R* data (not R(xR0) but rather R0(xR0a) very often) can perfectly be H and in a few cases looks like H1b (Portugal) or H17'27 (Sunghir).
We can't but continue the quest for knowledge of course.
"L3*(xall west Eurasian N and R* clades)"
Not really (with maybe rare exceptions). L3* per HVS-I can be M or N(xR). Neither M nor N are defined by any HVS-I mutation, as happens with H (but this one is downstream of R, so only R* sequences can be it).
I compiled (and revised carefully) much of that ancient DNA data HERE. Notice that Jean Manco and I fell out on her irrational rejection of the finding of H in North Iberian late Paleolithic, so I had to personally revise every single sequence that had even the slightest doubt when doing such maps according to the most up-to-date mtDNA phylogeny (PhyloTree). In many cases I had to review the original classification by the researchers, which had become outdated.
"isn't it possible there was natural selection for mtDNA H?"
It's always a possibility but so far unproven.
"Most Euros have 40-75% Mesolithic blood, but only 15-25% Mesolithic(confirmed)-derived mtDNA".
That depends on what you consider "Mesolithic mtDNA". It's not something clear at all, among other reasons because we do not have enough data for West Asia, for example. I'd say that X, W and probably much of N1 is Neolithic as is J and T (as such sublineages of JT). But H (which subclades?), HV0 (incl. V), U5 and U4 were surely present in Paleolithic Europe.
Re. H, I think that H1, H3, H4 and probably also H7 should be among the pre-Neolithic ones, because they are shared across the Strait of Gibraltar (with apparent "Iberian" origin) and that is best explained by Gravetto-Solutrean influence in the Oranian (Iberomaurusian) genesis (later largely replaced surely by Capsian arrivals from the Nile, Afroasiatic speakers surely).
It's very complicated I reckon but I have been following the matter, in great detail and with critical thought, for almost a decade now, so I believe I have an idea. Detailing all that and sourcing everything in comment format is not really possible however.
"If the Eastern HG were long in Russia, then why do the authors say that there is evidence for "mass migration into Europe"?"
ReplyDeleteTerminology. It's not like "Europe" is a well defined objective region like could be Australia. In this case clearly the ad-hoc definition of "Europe" is West of the Odessa-Kaliningrad isthmus, rather than the equally arbitrary Ural border decided by some tsar or the Volga border used by ancient Greeks. Arbitrary sure, wrong?, not necessarily so.
What is clear is that "Europe" is a relative and somewhat elastic concept. And in any case a geographic accident of Asia.
Well, we have a common use of the word Europe and migration from the eastern half of Europe into the western is correctly stated as "migration ACROSS Europe" and not "into Europe".
ReplyDeleteTo say that HGs survived on the periphery does not mean that farmers did not replace them in much of Europe. The statement that farmers "did not replace" HGs misses all of the replacement that did take place.
And the suggestion that "mixing led to" a resurgence of HGs is crass.
A better statement would have been that "farmers gradually replaced HGs in much of Europe, although some mixing did take place; HGs survived on the peripheries, which led to a resurgence, and some HGs spread from east Europe to the west."
But no, they are clearly pushing the idea of "mass immigration into Europe and mixing."
It is plain PC.
@Davidski and others:
ReplyDeleteWhat most likely happened is that there was greater remix with paleo-Europeans (WHG) in the Atlantic zone. The Atlantic Neolithic is complex and often neglected (both in Prehistorical and Genetic approaches) but shows a vast array of unique cultures or groups since early on, from Portugal to Belgium, etc. Some areas like Britain or Denmark remained in HG status until the Megalithic period.
The Oceanic climate must have been challenging to the neo-European farmers (EEF) with their crops used to drier conditions. Even in Central Europe it is very probable that, after the initial bubble, coincident with the Neolithic climate optimum (warmest temperatures ever until recently) there was a poof as climate worsened. There are several archaeo-statistical studies that seem to confirm this.
While in the SW (Portugal-to-Belgium) we see these unique cultures early on (probably more mixed populations similar to Gokhem), in the NW (Denmark-to-Ireland) the Neolithic only arrived a millennium later and with roots in those probably mixed areas of Atlantic mainland Europe. It also arrived with Dolmenic Megalithism (in a complex manner, no doubt), which has its ultimate origins in the Portuguese early Neolithic.
The Megalithic backflow, surely carrying that remixed Atlantic genetic pool, is very apparent in Germany in much of the Funnelbeaker phenomenon, particularly Michelsberg culture, which displaced Danubian (epi-Rössen) from North to South West of the Elbe and even in North France. So there was a backflow from the Atlantic and North Sea, surely complex and hard to interpret at times, but that is necessary to explain what we see in the data.
apostateimpressions,
ReplyDeleteWell, I think it's pretty obvious they're talking about intra-European population movements when they say "Western and Eastern Europe collided ~4,500 years ago".
But anyway, feel free to point out your concerns to the authors when the paper goes up for public comment at bioRxiv, which will probably happen soon.
Maju,
ReplyDeleteI have all the samples from those studies you listed(except the one from Greece, but I have one from Italy) in my "Ancient mtDNA package's" spreadsheet. All except the Neolithic samples are in my "Likely false positives" section.
https://docs.google.com/spreadsheets/d/1HcAhe7QvggT792VruuoZX6IsTg4LhWXV-Z_ZfTk2KGA/edit#gid=1562774546
I think it is possible they could have some true results, but I'm very skeptical. I have several reasons for being skeptical.
1.Upper Palaeolithic-Mesolithic Euro mtDNA from recent studies comes out 100% U, except C1, pre-H2a2b, and Z1a from Russia and R3 from Hungary. La Brana-1 was basically identical to Loschbour and K01(Hungary), so why would Mesolithic Iberians have such a different mtDNA gene pool?
2.Chandler 2005 found very similar haplotypes from Mesolithic and Neolithic Portugal, not found in any other ancient mtDNA study. Which makes me think the connection is a common contamination source.
3.The L3*, M*, or N*(xR, western Ns). You're right that R0 people oftnly lack 73A, but it R-73G is still fairlly common in Neolithic Iberia.
4.The samples from that French-language study have contradicting calls.
I also understand the evidence that they're are legit results, so I'm seriously considering they're.
If the R*s and R0s are mostly H(some are HV6-17 and HV0), then H was very popular in Neolithic SW Europe.
Maju,
ReplyDeleteSomething important to remember, incase you don't already know, is that ANE not WHG(like previously believed by many) is the main thing that makes north Europeans different from Neolithic Europeans and modern SW Europeans.
SW Euros don't have much less WHG than north Euros, and Basque have as much as NW Euros. The only modern Euros with more WHG than Gok2 are Balts.
C01 from copper age Hungary had a modern European-like level of WHG, so WHG was also rising in Hungary.
The Late Neolithic WHG-rise could have come from anywhere in central-west Europe.
How much ANE did Raghavan 2013 estimate for Karitiana?
ReplyDeleteNice!
ReplyDeleteNothing we didn't know already however. The main body of the paper is still the main thing.
Addressing the discussion above, I think that many older archaeological movements would not have great informative power for the autosomes of populations alive today, because it seems, while the earliest divergences of the roots of the tree we have now are present from a timeframe circa 50kya ago, the recent situation for the distribution of the three branches of WHG, ANE and the sub-branches of ENA is much more recent than the UP.
Ust-Ishim is almost certainly not the progenitor of Mal'ta, K14 is almost certainly not the progenitor of Loschbour, and Tianyuan is definitely an outgroup to all East Asians+Papuans, and so also not a progenitor to ENA.
UP movements before the LGM would probably inform older movements of ghost pops that we haven't defined yet, strongly suggested by the failure of the tree to address K14 and the weird positions of Tianyuan and Ust-Ishim. Mal'ta, Loschbour, and Stuttgart, are all at or post LGM, and inform modern genomes far more greatly. The fact that all E+SE. Asian form a clade to the exclusion of Jomon, and that the phylogeographic isolation this corresponds to, aka the cutting off of Japan, only happened after the LGM, also implies that the formation of E+SEAsian autosomes dates from periods after the LGM when the Jomon were cut off.
For this reason I think that most population structure in northern Eurasia took form post-LGM, and most events with relevance to modern autosomes do not have such great time depth. The northern tier of Eurasia was probably repopulated from the south over and over, and we are looking at the most recent repopulation here.
Have you bothered asking Hervella? Nope. Exactly the same as Joan Manco (JoanL in your forum, I believe). I did when doubts arose and Dr. Hervella gently confirmed to me that the methodology excluded any "false positives", that her H is H because it was tested with RFPLs, a method so far uncontested, even if a bit old.
ReplyDeleteSo if you want to join the heavily biased camp of Jean (who has too many personal interests invested in a book she published just before those results and whose ultimate "archaeological" reasoning is also extremely problematic and not accepted by anyone else I know of), feel free. But you will be wrong.
In the long term, facts will pile to the ones already existing in the direction I say. Why am I so sure: because there's already a huge pile of evidence in that direction, pile of evidence that some are in denial about for their own extremely biased reasons.
Who should I trust more: a historian of art who has never been able to properly argue her jingoist denial of hardcore aDNA evidence or real geneticists like Hervella and Lacan, two independent sources pointing in the same direction? And never mind mtDNA H in Epipaleolithic Karelia, different source, and the old Portugal conference paper of the very prestigious team made up Chandler, Sykes and Zilhao.
This is not even considering the many old studies that do suggest but can't confirm mtDNA H in Paleolithic Europe, East and West. Those are indeed doubtful but not "unlikely", just unclear.
"Upper Palaeolithic-Mesolithic Euro mtDNA from recent studies comes out 100% U".
False! Two different, independent, recent and good quality studies in the Late Paleolithic/Epipaleolithic Cantabrian Strip have provided three unmistakable H sequences. Another one found also H in Epipaleolithic Karelia. Why would you deny those FACTS?! Beats me.
What seems clear is that that was the case in Central Europe (Germany in essence) but Iberia presents a more mixed picture with both U5 and H, as well as some other oddball lineages. France sadly enough remains unresearched.
"Chandler 2005 found very similar haplotypes from Mesolithic and Neolithic Portugal".
Yes but they are not identical. The Epipaleolithic sample has some L3* (N*?) and lacks V (or HV0). Both have lots of apparent H but the Neolithic sample has even more (and at least one sequence is unmistakable H1b in my analysis). The individual sequences seldom are the same ones, what in mtDNA wouldn't be so rare. So it's still possible that there was some degree of replacement. Still I concur with you that (partial) continuity in Meso-Neolithic Portugal is possible, so we both disagree with the conclusions of that paper, because their own data does not seem to support unmistakable massive replacement at that point (the genetic pools are too similar before and after the arrival of Neolithic).
"I also understand the evidence that they're are legit results, so I'm seriously considering they're."
Cool. That's my stand: they look very real and there are several independent studies showing H in Late UP Europe (mostly in Iberia but also in Karelia). One may be wrong but which are the odds that all them are? Virtually zero.
"Something important to remember, incase you don't already know, is that ANE not WHG(like previously believed by many) is the main thing that makes north Europeans different from Neolithic Europeans and modern SW Europeans".
ReplyDeleteThere is more than just that. Their WHG is more distant from the actual samples by IBD, meaning that Northern Europeans (at least not Danes and Scots) are not direct descendants from the Lochsbour meta-population but from some relative. Instead the French look more directly descended from Lochsbour, even if they have less WHG apportion. I've been told (who?) that the same happens with La Braña and Basques/Iberians but haven't read anything specific yet.
→ http://forwhattheywereweare.blogspot.com/2014/04/lochsbours-ibd-in-modern-europeans-is.html
A mystery to me is how much of that Northwestern excessive ANE arrived with Indoeuropeans and how much corresponds to Nordic paleo-Europeans, who (as Motala data shows) had already a significant amount of ANE affinity. This has not been sufficiently clarified yet and I see no particular reason to think that IE demic flows (Bronze Age onwards) had greater impact in Britain/Ireland than in France or Italy.
Furthermore, another recent study showed that diet in Britain changed radically from fish to milk with the arrival of Neolithic and that this pattern continued until the Viking invasions introduced again the like for fish. This suggests that the population was stable in Britain between the Neolithic and (at least) the Danelaw.
→ http://forwhattheywereweare.blogspot.com/2014/02/neolithic-peoples-from-britain-and.html
It also suggests that those Atlantic farmers were lactose tolerant, what is a topic on its own right. Interestingly enough we see how two such populations (one lactose tolerant and another intolerant) were just beginning to mix in the southern Basque Country (Upper Ebro) in the Chalcolithic:
→ http://forwhattheywereweare.blogspot.com/2012/01/caught-in-act-lactose-intolerant-and.html
EEFs were in general lactose intolerant (at least judging by this gene, there may be others). So I presume that the lactose intolerant population we see in the Chalcolithic Ebro were EEFs (probably related to El Mirador sequence), while the lactose tolerant one were probably "Atlantic farmers" (Gokhem-like), which for whatever reason carried the gene (fluke or selection?)
Well, food for thought. I do think that there is something really interesting to tap in the Atlantic Neolithic, that was important in the shaping of modern European populations. What else?
Like I said I'm considering the mtDNA from old Iberian studies, I'm not die-hard for no-H and alot of H.
ReplyDeleteThere's Mesolithic H from France, you gave me a link to the study, it's written in French.
Not counting Hervella, Karellia, and Hungary my statement about 100% U is true. Also, U2-K14, U5*, U8c, and U-MA1 from Upper Palaeolithic north Eurasia, is convincing evidence there is something special about U and the WHG-ANE-K14 branch.
The only Mesolithic Iberian-French H is from the 2011 French-language study. The other Hs you're mentioning were tested for R0 at best, besides that all we know is that they had R.
There was an abstract a few weeks ago about genomes from Mesolithic-bronze age Spain, and so in the next few years there might be undisputed Mesolithic-H from Iberia.
@David,
ReplyDelete''Okay, firstly, I find it very difficult to believe that the Near Eastern farmers who lowered the levels of ANE on the steppe by ~5,000 YBP (but probably appeared there much earlier during the Neolithic) also brought R1a to Eastern Europe.''
Who said of ANE David? ANE was obviously there from the times of Dinosaurs but its more about the Y-DNA and Mtdna structure......
''When I read the line "mass migration into Europe" in the context of the entire abstract I actually see "mass migration from the Russian steppe deeper into Europe".''
Isn't that obvious also?? Russian steppe was a refuge of the Indo-European folks when they arrived there....
@ZeG
ReplyDelete''"But, I am nearly 100% sure that R1a1a coalesces to a period more recently than 10,000 years ago in the past. The reason is that I saw some posters using whole genome sequences from the Y chromosome at ASHG. These should be a more precise estimates because of the enormous marker set of more slowly mutating SNPs, and they too arrived at a relatively recent period for the last common ancestor of these common male lineages. In fact, if I recall correctly the divergence between R1b and R1a dates to ~10,000 years before the present in these studies, so R1a must have a much more recent coalescence."
Then I think it can be around 6000 bc
Which correlates the Zarzian Migration towards the Urals Which brought R1a!.
"A mystery to me is how much of that Northwestern excessive ANE arrived with Indoeuropeans and how much corresponds to Nordic paleo-Europeans, who (as Motala data shows) had already a significant amount of ANE affinity. This has not been sufficiently clarified yet and I see no particular reason to think that IE demic flows (Bronze Age onwards) had greater impact in Britain/Ireland than in France or Italy."
ReplyDeleteI agree, because based on leaks about the upcoming study with 65 ancient genomes, north Euros can't be explained as Gok2+Yamna, there has to be significant Loschbour and Motala12-like ancestry.
Lithuanians actually look like a mix of SHG and Yamna.
http://www.anthrogenica.com/showthread.php?3694-ANE-K8-oracles-for-Euros-using-ancient-Euros-and-modern-Middle-easterns
But it has been confirmed Yamna-types made a big genetic impact, and more so in north Europe than in south Europe.
Corded ware, Bell Beaker, and Unetice from Germany were practically modern northern Europeans, and could fit as being mostly Yamna, with EEF and I'm sure also WHG ancestry.
Now we know the steps that took place to create north and east Europeans, and Yamna may be their main ancestor.
Southwest Europeans, even non-IE Basque, show the same signal of Yamna ancestry but much less. By much less I mean possible as much as 20%, which is alot considering how far away Russia is from Iberia.
@ Krefter @ David
ReplyDeleteOne weird issue is how to reconcile the expected high ANE figures from Corded Ware, with the very low/no ANE from the BR1 and BR2 genomes, who are post corded-ware.
There seems to be a theme of autosomal diversity and inhomogeneity. We have very EEF-like genomes from bulgaria well into the Iron Age, even when IR2 and BR1-like genomes existed before. It seems like Europe was a jumble of diverse genotypes for a long period.
This makes me think that BR1 and BR2 were IE-ized culturally but not similar to Corded Ware genomically, and that the process of spread of ANE across Europe is not 100% correlated with the spread of IE culture across history, even when in the broadest sense movements from the steppe introduced ANE into Europe. Issues of class, ethnocultural status and adoption of culture with little introgression, probably kept high WHG, high EEF and high ANE genomes highly distinct for some period.
@ Krefter
ReplyDeleteAbout the Hap U connection, agree strongly.
David pointed out the reach of U into mesolithic Siberia years ago as well.
Addendum to previous comment: The whole picture painted by aDNA makes me think a consolidated, modern-looking picture for regional European autosomes took place very late, maybe well into historic times. The figures you quote seem to suggest slavic expansion driving up affinity to autochthonous WHG on top of Corded Ware+Farmers. This might be one such phenomenon.
This was certainly the case for e.g. Japan, SEAsia and South China. So might be similar for Europe.
I've never seen it. Offer it up. I might know more about Archaeology and the IE tree than you think. Several others too, I'm sure.
ReplyDelete@ryu
ReplyDelete"One weird issue is how to reconcile the expected high ANE figures from Corded Ware, with the very low/no ANE from the BR1 and BR2 genomes, who are post corded-ware."
Maybe different routes i.e.one group drifting west in the forest zone beyond the northern limit of LBK and another heading down the Danube with the two strands only mixing later?
//
@maju
"it would seem that Gravettian is an intrusive wave from West Asia (allegedly related to Amharian) that truly replaced the previous Aurignacian layer, at least partly and maybe in different degrees in different areas."
This is what makes me wonder about a single WHG - seems to me there ought to be two, the main one and a minority one mostly restricted to refuge zones mostly in the west.
What about r1b and bell beaker? Perhaps the implication is that only corded ware is from a yamnaya like source but r1b is from western anatolia via the aegean. Both finally clashing around 2500 BC in germany. I made a guess that r1a-L664 ancestors moved with r1b.
ReplyDeleteI doubt L664 moved into Europe with R1b from Anatolia. L664 very likely originated near the North Sea after its ancestral mutation CTS4385 moved west into Europe from the steppe, along with R1a-Z282.
ReplyDeleteThe Corded Ware R1a haplotype from Eulau looks like CTS4385+ (L664-). In other words, its potentially ancestral to L664.
" if I recall correctly the divergence between R1b and R1a dates to ~10,000 years before the present"
ReplyDeleteMakes no sense whatsoever. The divergence is necessarily much much older (see Underhill 2014, which uses full sequence, the only way to make some "molecular clock" sense). It is R1a coalescence (node) which seems so recent (10-6 Ka) but R1 as such is like 3-4 times that age and the complex branching or R1b is much older than the very recent coalescence of R1a.
For practical purposes R1b and R1a are two different things, much like I is different from J. While R1a has low complexity it has some (particularly a sharp distinction between European and Asian main branches). R1b has a great complexity that reflects its likely very ancient migrations.
"Zarzian Migration towards the Urals"
Uh? Can you document that claim Nirjhar? I'm unaware of that "migration" or even that Zarzian existed with that label for so long (6000 years ago is fully in the midst of the Sumerian period and right at the very end of the Ubaid period, arguably the first "empire" ever) but I'm willing to learn.
@Krefter: "Theoretical Yamna is 25% ANE, 40% Near eastern, and 35% WHG" (from the linked forum thread)
ReplyDeleteWhere do you get such high figures of "Near Eastern" component? Also, does that "Near Eastern" component includes or not the so called "Basal Eurasian" (African like) element found in EEFs?
Anyhow, please notice that Yamna culture as such had no impact in Europe west of the Odessa-Kaliningrad isthmus, being contemporary of Baalberge and Globular Amphorae, as well as other scattered nodes of early IE expansion in the Balcans, Caucasus and Central Asia. Yamna is what remained behind in the steppes after the first IE expansion waves and is likely at the ultimate origin of the Indo-Iranian branch (explained with nice maps: HERE).
For the record, because a lot of people imagine Yamna at the origin of IE expansion and it is just not correct. Although they were probably still good representatives of the original steppe IE stock (depending where the sample comes from).
"... north Euros can't be explained as Gok2+Yamna, there has to be significant Loschbour and Motala12-like ancestry".
That sounds like loosely correct to me but I'd like to see ancient samples (other than Motala, for example from Britain) to be safer. Also, in the case of Britain, Neolithic (Megalithic) colonization seems very strong and original from NW France. Why don't we see a Lochsbour IBD signal in them as we do among (North) French?
"Corded ware, Bell Beaker, and Unetice from Germany were practically modern northern Europeans"...
Maybe in the autosomal aspect (want clear data anyhow - all kind of doubts) but not in mtDNA, unless you pool them (i.e. unless you complete the process of admixture that was still incipient in the Late Chalcolithic).
"Southwest Europeans, even non-IE Basque, show the same signal of Yamna ancestry but much less".
Again I need clear data and a discernment of what is actually "Yamna" and what is "Motala". It's obvious that Basques have double ANE than Sardinians but it was unclear if that was caused by greater ANE affinity of WHGs relative to EEFs or by later waves, as you propose. And, if it is later waves, what part is attributable to actual IE flows (Yamna-like) and what part to earlier contacts with NW Europe (Motala-like)? 20% IE genetic infiltration seems a very high figure for peoples who were still speaking pre-IE languages (presumably Neolithic languages) in Roman times and where IE cultural (and demic) flows had only began arriving c. 1300 BCE (Urnfields wedge via Catalonia). If real it needs an explanation but, so far lacking detailed data, I cast a big question mark on this claim. Also notice how the paleohistorical Basque mtDNA pool (rather well documented nowadays) seems stable since at least Chalcolithic (possibly even Neolithic proper), what strongly suggests a continuity that does not allow for 20-25% IE immigration, much less if we understand that whatever IEs arriving so far West were already heavily mixed, so it'd be like 50%, something I deem impossible because it would have meant the indoeuropeanization of Basques (and Aquitanians, Iberians, etc.) necessarily.
Therefore there's some sort of substantial error in your estimates, Krefter. But, considering they are a draft, all is good.
@Ryukendo: "the process of spread of ANE across Europe is not 100% correlated with the spread of IE culture across history"
ReplyDeleteAlmost certainly right. The ANE signal is very interesting but should not be taken as automatically meaning IE genetic flow from the steppes in every case. Not at this (oh-so-preliminary) stage of the research certainly.
I already mentioned above that NW Europe (particularly the islands, where the population seems largely stable since Neolithic) has way too high ANE to have been caused by IE flows alone. The Motala mystery must be addressed properly before we jump to conclusions, likely to be too hasty.
My very tentative opinion is that, in the context of Atlantic Neolithic (~ Megalithic phenomenon) there were significant demic flows between SW and NW Europe that altered the simple WHG-EEF landscape described by Lazaridis. Extra complexity as you suggest may have also existed, of course, but even in the simplified scenarios there are way too many things that do not fit.
@Maju
ReplyDelete''I'm willing to learn.''
Please do that in the NI Blog as it is discussed and should be discussed more there:).
BTW Maju What is the age for R1a-M417 mutation??
@Nirjhar:
ReplyDeleteLink, please! I can't recall all the comments, articles or whatever other references that I'm sure you remember perfectly but I do not.
Okay, But please answer the question first:)
ReplyDeleteAbout the link in the Article there are various references for Zarzian migrations to the Urals the important one is from Bernard Sergent's data from his book (Les Indo-Européens. Histoire, langues, mythes, Paris 1995).
ReplyDeleteRe-Read it make your points....
@Grey:
ReplyDelete"This is what makes me wonder about a single WHG - seems to me there ought to be two, the main one and a minority one mostly restricted to refuge zones mostly in the west".
Yes... but what about the LGM and post-LGM rearrangements? They'd be internally European (no new waves from West Asia in any case, it seems) but they could be very complex. Also we don't understand well the expansion of Gravettian, how much it mixed with the previous layer, etc. In the particular case of Iberia (excluded Basque Country probably) it'd seem that Gravettian had more impact than Aurignacian or any other wave. On the other hand Magdalenian is like a "neo-Aurignacian" with harpoons (and it's been argued that it arose from epi-Aurignacian survival pockets, maybe in North France or Dordogne or wherever). And there are other complexities.
Well, the case is that do not understand well enough the demic-cultural processes affecting Europe and how they unfolded in the more than 30 millennia of Upper Paleolithic and Epipaleolithic. But one thing seems most likely: at least two different waves from West Asia seem to have happened prior to the LGM. And in this context it can be argued that one could be related (with all cautions) to yDNA I and mtDNA U (U5 particularly) and the other to yDNA R1b and mtDNA HV (H particularly).
My tentative idea is that it'd be Gravettian the one associated to R1b/HV and Aurignacian (and later Magdalenian) to I/U5. But this looks like it could be extremely oversimplifying, so there should be room to greater complexity, admixture and patchiness.
@Postneo: "r1b is from western anatolia via the aegean."
ReplyDeleteThis has two common misunderstandings:
1. Western Anatolian Neolithic is more recent than Thessalian Neolithic, so all chances are that West Anatolia as such did not have a genetic nor cultural impact in the European Neolithic (whose known origin is Thessaly) but rather than the immigrant seed of EEFs actually arrived via a coastal migration (they were good sailors judging on Cardial culture and Cyprus may well have been an intermediate step).
2. European R1b does not clearly derive from Anatolian R1b but both derive from a common source that could well be Balcanic (or alternatively West Asian, but in this case Iran seems a more likely option).
In any case, European R1b is in essence (M412) a distinctive subclade with its own patterns: a relatively rare core R1b-M412* that suggests origins somewhere between Hungary and the Pyrenees and then two subclades: R1b-S116 (the most common one) originating almost certainly in SW Europe and R1b-U106 originating probably near the Netherlands. Hence R1b in Europe (S116 and U106) most likely expanded from Western original cores, regardless of timeline (although my best estimates make it happen between the LGM and Epipaleolithic, for the record).
See:
→ http://leherensuge.blogspot.com/2010/08/r1b1b2a1-is-almost-unique-of-west.html
→ http://leherensuge.blogspot.com/2010/05/finally-some-good-research-on-r1b1b2.html
I doubt R1b is from Anatola. i guess that R1b expanded deeper in Europe from the Southwestern steppes.
ReplyDelete@Maju
ReplyDeletePerhaps you can enlighten the rest of us why R1b succeeded in all habitable regions of Europe during the stone age, but common haplogroups like I2, I1, now defunct branches of I* (pre-I1?), C-V20 died out or nearly died out.
I'm curious what theory or technological advantage the "stone age" R1b guys might have had over equivalents of their time.
Don't counter with giving me statistics on tiny outlier populations in the Balkans or NE Europe where R1b is low. My statement is clear that "Habitable" is referring to locations which would have drawn people and encouraged outbreeding. Not only that but we're aware that I2/I1 were at the very latest present in stone age Europe.
@AWood:
ReplyDeleteNo idea. I only know the data and what the data says is that.
My best idea, as the PC model seems inviable, is that populations dominated by those R1b subclades (and lots of mtDNA H) re-expanded within that expansion from the Atlantic in Megalithic times. But only further evidence can confirm this hypothesis (so far nothing clear).
@Nirjhar
ReplyDeleteI'm not Maju, but I'd say R1a1a1-M417 is about 4-7 thousand years old, based on average branch lengths from Hallast et al 2014, with K at 50 kya. We know from aDNA that it is more than 3400 years old, but that doesn't help much. We should of course have solid aDNA calibration very soon for a much better estimate.
So I guess more or less Chalcolithic, in the same time frame as the spread of wheeled vehicles, Caucasus copper metallurgy, extended trade networks stretching from the Pontic Steppe and E Mediterranean to Tajikistan and India, the Kurgan expansion, Kura-Araxes expansion, beginning of the Early Harappan, cultural influence from South Asia into the Kopet-Dag region.... And umpteen other things, which may or may not be connected.
R1, using the same method, should be around 24 kyo, give or take a few millennia. R1b1a2-M269 looks to be around 6-8 kyo.
Maju,
ReplyDelete"Near eastern" aka "ENF" is a proxy for early Neolithic near eastern farmers that Davidski created. Stuttgart is 68-72% ENF and 28-32% WHG.
http://eurogenes.blogspot.com.au/2014/12/crowdfunding-for-2015-new-test.html
Here's my thread agian:(http://www.anthrogenica.com/showthread.php?3694-ANE-K8-oracles-for-Euros-using-ancient-Euros-and-modern-Middle-easterns).
My oracles reveal that Northwest Europeans can't be explained as European hunter gatherer(WHG+ANE) and EEF. So at some point there was heavy gene flow from somewhere east(more ANE-like) of west Europe.
We know it wasn't from west Asian-like populations, so what does that leave us with? East Europe.
It looks like to me Southwest Europeans simply were less effect by this admixture event.
In a basic PCA of west Eurasia you can see a cline starting at Sardinians and going all the way to Lithuanians. Northwest and southwest Europeans are in-between the two.
I forgot to mention Maju, that theoretical Yamna beats out Motala12 and modern middle easterns almost every time, as the source of ANE in Europeans. The Gaussian method is best.
ReplyDeleteThere's no way Hunter gatherers are the main source of ANE for any Europeans. Lithuanians are the best proxy for most, but It is possible a Lithuanian-like pop was created from a EEF+SHG/WHG+Yamna admixture event.
Cernavoda matches the accepted age of ancestral L23(ancestor of L51 and Z2103/2105) at 4200BCE.
ReplyDelete@Krefter: I don't understand well what is your basis for your "theoretical Yamna", personally I'd advocate for a much smaller "Near Eastern" apportion in that group but better let's just wait for a sample or two, OK? In the end, our opinions and models are pretty much rendered worthless if direct evidence says otherwise.
ReplyDeleteAnyway, using David's "fateful triangle" model, an a ruler and a calculator, and assuming the early Kurgan people were somewhere between NE European and Volga-Ural, I get that their WHG was around x2.6 their ANE (top axis), while you instead claim that it's just x1.4. Then the WHG+ANE is like x2.5 the ENF, so my "theoretical Yamna" is:
→ 29% ENF (<40%)
→ 20% ANE (<25%)
→ 51% WHG (>35%)
We are not that far from each other but I feel you're minimizing the WHG component and aggrandizing the others.
But then again, lacking a direct sample we can't but speculate. What if the real Yamna was (as is very possible) a more simple WHG+ANE composite with almost no ENF, for example? There's nothing that argues against a, say mere 5% ENF, all the rest coming from EEFs and other later admixtures maybe. Only actual ancient DNA data can tell us how they actually were.
"My oracles reveal that Northwest Europeans can't be explained as European hunter gatherer(WHG+ANE) and EEF".
Can't they? West Scots:
→ 1 pop.: Gokheim
→ 2 pops.: Armenian+Motala
→ 4 pops: Cyprus+Georgia+2xMotala
What do we see here? That (barring Gokheim) Motala is key. And what is Motala? Aboriginal Euro with more ANE than WHG and less than what I'd expect from Yamna. Where does the West Asian element (not ENF) come from? EEFs, naturally. And why is the West Asian element "displaced north"? Not sure but it's possible that it is because those pops have more WHG+ANE affinity than Palestinians (close to ENF). What happens if you force ENF or Palestinians as one of the references (is it even in your list?) that the Paleo-Euro and the Paleo-Siberian components grow. It's all cladistics and the choice of pops is key.
Really: why don't you have a Palestinian sample or a "theoretical ENF" one? You're artificially enhancing the West Asian element in those "oracles" by not having the right references.
Of course if you told me that "I do have a Palestinian/ENF sample but just does not shows up", then I would have to reconsider but my impression is that you're artificially displacing the West Asian immigrant pop. to the North by using the wrong reference pops.
The French example instead works well (quite as expected from Lazaridis): Motala+(Neolithic_Hungary+Stuttgart=2x EEF)+"Yamna", what may need a correction because we do not how exactly were the real Yamna people. Also apportions do not need to be exact quarters nor halves, so it'd be cool if your "oracle" algorithm could calculate other percentiles, because that may be a cause of error and confusion in itself.
What I retain of your attempt is that in 1st position for all Western European populations (and some not so western nor so modern) in the single best approximate, it's invariably Gokhem. That alone speaks volumes!
Maju,
ReplyDeleteThe abstract above says that Yamnaya had its ANE lowered by ~5000 years ago after mixing with people of Near Eastern ancestry.
So it's highly unlikely that Yamnaya had almost no ENF (aka Near Eastern admixture).
Patterson said that linguistic evidence has tracked the ancestral language, called “late proto-Indo-European” to about 3,500 years ago in the Caucasus, among a people who had wheeled vehicles at a time when they were just being put into use.
ReplyDeleteGenetic evidence ruled out one likely related group in the region, the Yamnaya, because their DNA showed the group had hunter-gatherer ancestry, which is inconsistent with the fact that two Indo-European groups, Armenians and Indians, don’t share it, Patterson said. That made Patterson look south, to the Maikop civilization, which likely had significant contact with the Yamnaya, as a plausible culture where Indo-European languages originated. Samples have been obtained from Maikop burial sites, but the DNA work to test that proposal is pending, Patterson said.
3,500 years ago is the late Bronze Age. So the date is wrong.
ReplyDeleteThere are other problems with that article too.
The abstract you posted has a date of 12-12-2014, the article I quoted is 4-12-2014, I think that abstract was probably written before Patterson gave the interview. Do you think Patterson didn't read the article and wouldn't have had it corrected if they misquoted him?
ReplyDeleteYes, I think the reporter made a mess of Patterson's presentation, and it wasn't corrected.
ReplyDeletePresume you are incorrect and the article is valid, why would they completely rule out the Yamna for having hunter gatherer ancestry? Would they rule them out if it was only hunter gatherer autosomal they contained, would it be logical to?
ReplyDeleteMaju, and others:
ReplyDeleteOnly when you realize the problem of making statements like "Corded ware was northern Indo-European", or 'pre-Germanic and Balto-Slavic', or the Catacomb culture was a "Graeco-Indo-Iranian' block, then you can *only begin to develop a more sophisticated understanding of European prehistory.
Its a very specific thing to be incorrect about...
ReplyDelete
ReplyDeletePeoples of the Baltic circle of cultures (Ertebo/lle) were hunters and fishermen, and produced only one or two pottery forms. This was a rather tall, broad-faced population of the North Europoid type, who buried their dead in the extended position on the back.
Certain cultures of syncretic appearance involving Northern and Southern features were formed in Central Europe and the Baltic during the 4th-3rd millennium BC, e.g. Comb Ware, TRB, and Globular Amphora cultures. During the Early Bronze Age these cultures were displaced by the Corded Ware (Battle Axe) culture characterized by flexed inhumation on the back or side under a barrow. Specifically, this culture embraces both bottle-like vessels and bowls with funnel-like neck of the Northern circle of cultures, and also vessels of the Danubian type. In an anthropological sense this population combines traits of southern gracile and northern massive types, in particular bearers of TRB culture (Schwidetzky 1978).
...
The Yamna culture of the Pontic-Caspian steppe is recorded for an enormous territory between the North-Western Pontic area and Trans-Uralia. Its sites are known here in the basin of the Emba and Tobol rivers, the Karaganda region and further eastward (Merpert 1974). The Yamna population generally belongs to the European race. It was tall (175.5cm), dolichocephalic, with broad faces of medium height. Among them there were, however, more robust elements with high and wide faces of the proto-Europoid type, and also more gracile individuals with narrow and high faces, probably reflecting contacts with the East Mediterranean type (Kurts 1984: 90).
http://dienekes.blogspot.ca/2006/05/anthropological-types-of-corded-ware.html
Maju, I didn't explain how theoretical Yamna came about.
ReplyDeleteTheortical Yamna is based on leaks Reich and Laz have given about Yamna genomes over the last few months.
Here's a hand drawn PCA someone gave Davidski. LNE/EBA=German Bell Beaker and Unetice. MNE is middle Neolithic(Germany) and ENE is early Neolithic(Germany). CWC is German Corded ware, and EHG are Mesolithic Russians.
http://imageshack.com/f/540/blmKW4.png
It was said that Yamna(from Samara Russia) could be modeled as 50% Aremenian and 50% EHG. Making them as Near eastern as Central Europeans, so and making it impossible for north Europeans to be Yamna+Gok2.
Another leak said CWC could be fit as 73% Yamna, and no Europeans today could be fit so high, and Lithuanians were the highest today. We assumed they predicted Yamna ancestry purely based on ANE, and so if Lithuanians with 18-20% ANE are close to 70% Yamna(assumption), maybe Yamna had something like 25% ANE. So 40% near eastern, 25% ANE, leaves us with 35% WHG.
If anything 35% WHG is a high-bound estimate for Yamna.
Romulus,
ReplyDeleteI can't presume I'm incorrect, because obviously the date in that article ought to be 3,500 BCE, which is the early phase of the Maikop culture. That's just basic knowledge and the only thing that makes sense in this context.
And it also seems to me that the proposal Patterson is testing is that the earliest Proto-Indo-Europeans had very little WHG ancestry and expanded from the Maikop culture both north into Yamnaya and south into Armenia and India.
If so, this would get around the problem that Armenians and Indians supposedly don't harbor any WHG ancestry.
However, since we're dealing with a language spread, then it might not be a problem in the first place, because we can't expect to find specific autosomal signals in every group that adopted Indo-European speech.
And I'd like to see how Patterson worked out that Armenians don't have any hunter-gatherer ancestry. I'm very skeptical of this claim.
I'd hate to break this to you, but reporters make mistakes all the time, particularly in science articles, so second hand sources like these aren't very useful, especially when the mistakes are very obvious, like that 3,500 YBP date. Also, Patterson put out a paper not long ago saying that all Europeans had significant East Asian ancestry: Ancient Admixture in Human History. That was a mistake. He makes mistakes too, like we all do.
Romulus,
ReplyDeleteWhat point are you making with those outdated anthro quotes?
We already know that Corded Ware and Yamnaya fit into modern North and East European genetic variation.
lol 3500 years ago? That has to be a typo or the reporter misunderstood what was being said.
ReplyDeleteWheeled vehicles came into existence during the second half of the 4th millenium BC, so 3500 BCE has to be what was meant.
That is a really complicated answer, the simple answer is that they didn't find R1a in the Yamna, or that they found a majority I.
ReplyDeleteRomulus,
ReplyDeleteOK, here's a much less complicated answer that you might actually understand: 3,500 years ago? LOL
Did you get that? Just wanna make sure.
3500 BC fits though. That kind of typo I can see. Getting one of the core points wrong, especially a follow up quote from Patterson about having to look south to the Maikop for the PIE people seems unlikely.
ReplyDeleteYou're missing the point. If they got that wrong, then what else did they get wrong?
ReplyDeleteActually, the other key point you're missing is that this has nothing to do with Y-DNA, but rather with inflated hunter-gatherer ancestry among the Yamnaya, which Armenians and Indians supposedly lack.
That's what's causing the models not to fit. These models they're testing are based on autosomal ancestry.
Maju,
ReplyDeleteIf we assume ANE K8, west Eurasian PCAs based on genptype, and the Basal Eurasian+WHG+ANE model are true, there's a very low chance east Europe isn't the main source for west European ANE.
There's absolutely no way around this, so I don't understand why you keep pushing for local ANE in west Europe.
Maju,
ReplyDeleteIn Oracles the Gaussian method is much more accurate than the Least-squares method.
Rankings don't matter, Oracle scores are what matter.
For west Scots this is a very good fit.
1 50% Yamna +25% Loschbour +25% Neolithic_Hungary @ 1.878936
I think you are grasping at straws :). They wouldn't ignore uniparental markers. You can tell by the language they use in the abstract you posted. Two words are important imo, "descended" and "affinity".
ReplyDeleteIf the Yamna don't show any Y DNA I I'll donate $50 to your blog.
I think you misunderstood what David was saying, Romulus. He isn't saying they did not recover Y-DNA from Yamnaya.
ReplyDeleteWhat he is saying is that the claim made by Patterson that Armenians and Indians lack the hunter-gatherer ancestry found in Yamnaya peoples, was made based on autosomal results and not based on sex-specific markers.
Yamnaya is modeled as 50% Armenian and 50% Karelian. You can't get 51% WHG with that. Our model follows the outline of the study. It could be that Yamnaya is 23%ANE, Corded at 16%, and Bell Beaker at 13%. Either way, they have to be close to 40% Near Eastern, as that is about half of what Armenians are.
ReplyDelete2/3 Yamnaya makes sense for Corded, as that is what I figured, just based on how they deviate from the 73% mark when a line a drawn from Gok2. About 60% Yamnaya is my guess. Corded Ware likely evolved out of the northern area around Samara and to the west. More than 75% of Yamnaya is found below this mark, so I would bet that there could certainly be IEs living in Yamnaya that are close to 60% Near Eastern with little WHG. Probably less than 10-15%. This would solve the WHG issue. They need samples from the Pontic region to get a full picture.
ReplyDeleteRomulus,
ReplyDeleteThe quotes attributed to Patterson in that article you posted are related to the autosomal DNA of the Yamnaya samples and the mixture/migration models based on it.
Whether the Yamnaya Y-chromosomes were tested or not is completely irrelevant to these mixture/migration models and the statements attributed to Patterson.
Do you understand what I'm saying?
I would bet my money rather on lineages such as J* and R1b and maybe some R1a (not meaning there wasn't already R1a in the Steppes) beeing brought by Caucasian like pastoralists from Kurdistan-North Iran into the Yamna horizone.
ReplyDelete@Krefter 40-50% WHG is far too high Not even modern Ukrain and South Russia, which has been heavily influenced by the Slavic expansion has that high WHG admixture. IF Western Europea was 50/50 farmer WHG like, and the Steppes differed in having much more affinity to ANE than West Europans. How can WHG reach levels of 50% in Steppes?
ReplyDeleteIf we take into account that your average North Caucasian has levels of 7-10% WHG, and that WHG is a minor component in all Indo Europeans outside Europe, I think an average of 20% WHG in Yamna is a good guess.
Kurti,
ReplyDeleteWe already know that Yamnaya and Corded Ware genomes cluster much further north than early and middle Neolithic farmers from Western and Central Europe, the latter of which are around 50/50 WHG/ENF.
So you need to take that into account. The estimate of 40% or more WHG among the Yamnaya might well be correct, and indeed, this is what might be causing the problems for Patterson.
I think that our Yamnaya was 36% WHG, correct?
ReplyDeleteYes, which put it close to the Erzya people of the middle Volga region on the K8 PCA.
ReplyDelete"So you need to take that into account. The estimate of 40% or more WHG among the Yamnaya might well be correct, and indeed, this is what might be causing the problems for Patterson."
ReplyDeleteI never thought of that. Middle Neolithic Germans must have been 35% or more WHG, and Yamna is much farther north, not just east of them.
CWC or BB+EEF might not work for many modern Europeans, but the similarity between bronze age Central Euros(In Hungary and German) with modern Euros, is no coincidence(I guess a similar genetic makeup can be created multiple times), and since the bronze age not a whole lot has changed.
The only people excluded from this even are Sardinians.
The Neolithic Spanish genomes are probably what caused this paper to be later than expected.
ReplyDeleteIt was a horrible idea of Reich to do this in my opinion, because everyone knows Neolithic Spain was the same as Neolithic Germany, Sweden, Hungary, etc. If anything he should have gotten a bronze age Spaniard, or a bronze age genome west of Germany in general.
Yeah, this paper would really kick ass if it also had:
ReplyDelete- Near Eastern Neolithic farmers
- Bell Beakers from Iberia
- Sintashta samples
@David: yes, it says so but which are the figures? Is it 10%, 20% or 40% (as K. claims a bit happily)?
ReplyDeleteAlso what kind of Near Eastern input is that: ENF/proto-Palestinian as in EEF or some other sort of more Northernly kind, as in K's "oracles"? I ask this because in some ADMIXTURE analysis it's possible to spot that dual kind of West Asian genetic influence in Europeans in an apparently irregular manner, so wondering.
@Mike: you come again arrogant and jerky, so I pass of even discussing whatever may be worth in your comment. Suffice to say that we disagree almost totally.
My personal guess is
ReplyDelete20% WHG, 50% NE farmer and 30% ANE.
@Davidski
ReplyDeleteThere are simply too many points speaking against this.
1. There was some statement made by Lazaridis and co that Yamna can be explained as 50/50 "Armenian like farmer and Karelian.
Average Near East for Armenians is 78%. Average of Near East for pure NE Europeans is 25%. This is roughly 50% Near East.
2. WHG is very minor in every Indo European group outside of Europe. Which again speaks against WHG as a strong elements.
3. We know that Yamna and every ther late and post neolithic sample in Europe cluster further NorthEAST. How does Yamna clustering further Northeast indicate more WHG in them, if the ANE element on itself can explain this as well?
So ~30% ANE in combination with 20% WHG can is much more of a possibility as 50% WHG.
Corded Ware was 1/3 different from Yamna this adds more WHG into the game for them.
Kurti & Maju,
ReplyDeleteWe don't know how much EEF or ENF admixture the Yamnaya genomes have. But we do know that they cluster just southeast of Corded Ware genomes, which themselves cluster in present-day North Central Europe.
It seems that the reason Yamnaya cluster there is because of a greater pull towards MA-1, which probably translates to more ANE ancestry, rather than a higher level of EEF/ENF/Near Eastern admixture.
Indeed, there's really only one type of Near Eastern admixture, which is what I call ENF on my K8 PCA. The others that show up in various Admixture runs form due to the variable levels of ANE and probably very recent drift among modern Near Eastern and Caucasian populations.
Kurti,
ReplyDeleteThe Karelian genomes are from the Mesolithic, so they probably have 0% Near Eastern admixture.
Thus, 50/50 Armenian/Karelian is much less than 50% Near Eastern.
The majority point however which speaks against WHG beeing present close to 40-50% is that no Indo European outside of Europe show enough of this ancestral component. If it were only some groups we could argue that WHG got deluded. But all Indo Europeans outside Europe (suprise suprise the place where WHG peaks before IE expansion) have very little of that.
ReplyDeleteAlso what I forget to add. No even nowadays, after the Slavic expansion which should increase WHG allot, does it reaches anywhere close to 50%. Even Ukrainians have around 44% despite beeing relative "new comers" from further North.
Simply too many aspects speaking against WHG as major component imo.
Davidski,
ReplyDeleteHas anyone on Reich's team commented on ANE K8, or anything people on this blog have theorized about Yamna? I'm afraid the BB-WHG-ANE model is going to fall apart after their study. We might not get numbers from their study we're familiar with, just new models for relatedness and admixture.
Kurti,
ReplyDeleteWhat's your rough estimate for WHG among the Karelian hunter-gatherers?
Keep in mind that the nearby Swedish hunter-gatherers were 80% + WHG.
Krefter,
I'm pretty sure all of these new samples will land within the K8 triangle.
@Davidski "so they probably have 0% Near Eastern admixture."
ReplyDeleteOK but probably is not sure enough though.
The Armenian/Karelian thing was a rough explanation anyways. But still that would make ~40% Near East. Which I still think is higher considering that Yamna was an agricultural society to begin with and I don't know of any agricultural society with below 50% Near East.
But than even if it was below 50%.
That would only increase my estimations for ANE, because the other arguments against strong WHG still stay.
@Davidski"What's your rough estimate for WHG among the Karelian hunter-gatherers?
ReplyDeleteKeep in mind that the nearby Swedish hunter-gatherers were 80% + WHG."
As already said I doubt that Yamna can perfectly be explained as 50/50 karelian/Armenian, it was probably a rough visualation. Otherwise the ANE scores would be around ~16%. That doesn't sound like very much to me as the recent speaches of Reich, and lazaridis would indiciate.
And I already gave other reasons why I don't believe WHG could be that strong. Simply because other Indo Europeans outside Europe are very weak in that. Heck even Semites from The Levant and North Africans have much more of it.
I would correct my estimations to from 20 to 30% WHG at max. But still I doubt for many reasons.
I said probably because they might have a couple per cent, who knows? But I know for a fact they cluster as far north as the Swedish foragers, so even a couple per cent is unlikely.
ReplyDeleteThese Karelian foragers have to be mostly WHG. That's probably why Armenians were chosen as the other 50% for Yamnaya, because they have some ANE. If the Karelians had extremely high ANE, like nearly 50%, then Stuttgart would suffice as the other 50%.
And what do you mean that the Yamnaya genomes have to be over 50% Near Eastern because they're from an agricultural society? Northeast Europeans were agricultural too until recently, when they started becoming urbanized. This argument can't be applied to post-Neolitic populations.
Kurti, we can only guess what Yamna was based on Leaks. The 50/50 Armenian/EHG thing might not be exactly correct. That is the basis of our estimate on what Yamna would score in ANE K8.
ReplyDeleteThere's no disputing now that Bronze age pops from east Europe made a big effect on all western Europeans, but we don't know whether Samara Yamna is the best proxy for that eastern pop(s?), and we are even further from knowing what type of people brought IE languages to Asia.
IEs may have made little genetic impact on Asia.
It would be nice to get a genome of a early Indo Iranian from Iran or India.
Maju
ReplyDeleteObviously I hit a nerve with the basques
Sorry that , for you , nationalism takes precedence over scholarship
Why would ANE in the Yamnaya genomes be only 16% with a 50/50 Karelian forager/Armenian model?
ReplyDeleteIf the ancient Karelians had something like <40% ANE, and the modern Armenians even just 12%, that makes around 25% ANE, which sounds reasonable.
Here's another problem, Maju mentioned.
ReplyDeleteCelts expanded recently and much of western Europe before Rome was non-IE. So, how do we explain the ANE? One way or another most of it must of come from east Europe.
Assuming all ANE in Basque is of east Euro origin, Basque have to have a big chunk of east Euro ancestry. To be honest this sounds wrong.
Where's the archaeological evidence of over 30% of west Europe's gene pool being replaced during the bronze age, before Celts expanded?
We always have to consider in the back of our minds the Laz 2013/14 model might be wrong.
Krefter,
ReplyDeleteYou would still need an explanation why Basques and in fact all modern Europeans except the HGDP Sardinians cluster east of all Neolithic and Copper Age genomes from Western and Central Europe.
I can't see a better explanation than the arrival of ANE in most of Europe after the final Neolithic.
David
DeleteI know it's sounds wushu washy, but we need to fully exclude the seeds of change to be *within* Europe itself ?
Are u content we have done this ?
Kurti,
ReplyDeleteArmenians are 15% ANE. Karelians are probably double that, or a bit more. Yamnaya will be near 25% ANE.
@Krefter: that plot is most interesting. On that data alone CW and derived groups were almost exactly 50% EEF + 50% EHG, maybe with a slight tendency towards ANE only. This is a very simple explanation for the admixture in late Chalcolithic Central Europe (and the main source of ANE further west).
ReplyDeleteIt seems apparent that the "Lazaridis triangle" (WHG-EEF-ANE) can and should be replaced here by a new triangle EHG-EEF/ENE-ANE in order to explain best CW and the rest of the Kurgan peoples. I annotated that here for our convenience.
If so CW/(...)/Unetice appear as almost exact 50-50 admixture of EHGs and EEFs (ENE, the proportions vary slightly using MNE but I stuck to ENE for consistence with Lazaridis' EEF). Yamna is about the same as CWC but with strongest tendency towards ANE (and maybe also slightly towards EHG).
There's no apparent need per the graphic to get West Asians (other than the fraction already present in EEF/ENE) in this. Also, as a side note, Armenians are surely not the best proxy for West Asianness, as they quite apparently have some European backmigrant ancestry (attributable to Phrygians) and hence may introduce confusing factors. It only becomes necessary, maybe, if you stick to the "Lazaridis triangle" but if there's already an EHG sample that explains things better...
I'm guessing that the EHG sample is Kostenki 14 and therefore not too reliable as representative (too old) and therefore Reich et al. feel the need to explain the displacement of Yamna towards the bottom of the graph, which indeed looks like pushed by some Caucasus-like population.
An alternative explanation could be that it is a mix of WHG+ANE (or something equivalent) with only a very slight West Asian tendency. I will have to accept that you know more of those leaks and that there are actual formal tests that actually suggest that Caucasus-like apportion - but I want to read it in depth before I make up my mind.
Thanks for sharing that information in any case. Very interesting and worth thinking about it.
"Rankings don't matter, Oracle scores are what matter."
I thought they were sorted by best score. My bad.
@Krefter
ReplyDelete"IEs may have made little genetic impact on Asia.''
Also archaeological,archaeotextual,anthropological,traditional.
Still no problems then....
''It would be nice to get a genome of a early Indo Iranian from Iran or India.''''
Rakhigarhi coming soon....
So if CW derived groups were almost 50% EEF (~40% ENF) like and 50% H&G (which already includes at least ~20% ANE). despite beeing 1/3 local substratum
ReplyDeleteHow can we assume that Yamna could have been anything over 20-30% WHG?
I also gave my other arguments against this.
And I will add another one. Lets assume.
Yamna was 40% WHG, they would have to be more ENF like anyways since they would be slightly further south than CW. But just for the heck of it let's assume their ENF was also just 40%.
Only 20% would remain for ANE. 20% ANE is almost the same amount as among modern Northeast Europeans. So why would Lazarides, Reich speak of substantial MORE ANE in Yamna if it was the same amount found in modern Northeast, Central-East Europeans.
Alone if we take into account that some of the ANE must have been deluded on it's way further into Europe.
Absolutely NO WAY is it possible that ANE could have been as low as 15-20%
Now at all the other arguments
weak WHG in Indo Europeans outside Europe and even higher in Afro_Asiatic speakers
All the close neighbors, who have not been affected by the Slavic expansion do not have anything close to 30%, let alone 40% WHG
Maju,
ReplyDeleteEHG = Mesolithic Karelians
Another "cheap analysis" on your leaked plot, Krefter: HERE. What that plot suggests to me is that the "Lazaridis triangle" WGH-EEF-ANE can be simplified and reduced to a new "Reich triangle" in which the ANE vortex is replaced by a much more real vortex in Yamna itself. This seems enough to explain modern European ancestry (barring maybe some places like Sicily) as WHG-EEF-Yamna, which seems archaeologically consistent.
ReplyDeleteMy logic to do that "shrinking" was that if K14 (EHG) is problematic for too old, the same can be said of Ma1 (ANE). They are not representative (neither) of early Holocene populations of Europe, even if they are possibly their partial ancestors (particularly of Yamna people).
So there is another problem about what was behind the Yamna genesis but, following your plot, it would seem that they can be explained (tentatively at least) as a mix of EHG-ANE-EEF, without obvious need to resort to further West Asian inputs.
All that is annotated in this last graph, for our easier visualization. Please tell me what you think.
Kurti,
ReplyDeleteWe have Yamnaya at 38% Near Eastern, 36% WHG, and 26% ANE. It works. Why wrestle with it?
26% ANE, then CW is at least 66% of that. We went by the original 73%, and had CW at 19% ANE. 26% ANE is quite a bit more than 19%. It causes a major eastward shift.
ReplyDelete@Chad
ReplyDeletewho were the EEHGs according to you then???
I think they're about 65% WHG, 35% ANE.
ReplyDeleteTo all
ReplyDeleteLet me ask a simple question:
Why does Ust-Ishm shared more derived alleles to the Mesolithic European sample from La Brana than modern Europeans ? How does this sit with a position for 'massive eastern Intrusions' c. 4000 BC ?
COuld it be the antiquity of Ust-Ishm itself ?
Btw,
ReplyDeleteI did share with Reich, where I thought everyone would plot. Plus, some other hypothesis' about the shift. He didn't tip his cap though. He just said thanks.
@Chad
ReplyDelete''I think they're about 65% WHG, 35% ANE.''
And their Y-DNA Structure?
Maju,
ReplyDeleteYou can't put a triangle on that plot because it's based on a genotype PCA. You can only put triangles on plots created with output from Admixture, Structure etc.
Mike,
There are no hints in any of the data we have seen to date or in the leaks coming through about the new data that expansions from within Western and/or Central Europe after the Neolithic could have produced the current European genetic structure.
All indications are that all Western and Central Europeans could be fit as pure WHG or mixtures between WHG and ancient Near Easterners (ENF) until the final Neolithic. Then, rather abruptly from the Copper Age onwards, no Europeans could be fit like that except some Sardinians.
Kurti,
Go up to 35% WHG and we have a deal.
I'm not sure... R1a is a common choice, I wouldn't be bad, but there is a good chunk of L23 and M73 among the Bashkir, today.
ReplyDeleteMaybe, a C and a Q scattered around as well.. It's just gonna have to wait for the paper.
ReplyDelete@ Mike Thomas
ReplyDeleteWell, I think many of the disagreements have sublimated into semantics at this point. But they are still valid.
I don't think people here ever made a biology=culture argument. ANE not 100% associated with IE is probably true, but ANE generally associated with the spread of IE culture is probably true as well.
'Culture is only a social construct'
We had a class recently where a group of history graduate students attempted to dispute the existence of the 'east Eurasia cultural complex' and 'west Eurasia cultural complex' in Niall Ferguson and Pomeranz and other world hist texts with precisely that argument. Some of the undergraduates incl. me--the 'less educated ones' if you would like--immediately disputed this, because this was obviously a misunderstanding of the authors' terms. The authors were obviously using these words to refer to a system of inheritance of practices and ideas--which is what culture is; and it makes no sense to trace the bulk of the culture of Ming China to Ancient Egypt, no matter how hard we try. Such things as curly-leaved plantlike border decorations and realistic portrayal of the human form in sculpture, or anthropomorphic sculptures of gods, would not be present in East Asia in the post-Qin period were it not for contact with West Eurasia, which was set apart from East Asia in having these early on and having them all over. So distinct traditions exist, and these terms obviously have methodological significance at the very least, even if the boundaries were permeable. Just because colors are arbitrarily defined doesn't mean they are a 'social construct' and don't exist.
The existence of an 'IE culture' is abundantly inferred by philologists and linguists. It has been noted for some time that IE-speaking areas tend to associate knowledge and religion with sight, and semitic areas assoc. same with hearing. Linguistics give Proto-IE *wid, 'see/know' to become Ind 'Vedas', Eng 'wise, wit', Lat 'videre', Gr 'idea', 'eidon'--a testament to the deep grooves such a culture has carved in myth, epic, and even philosophy in West Eurasia even today since the axial age.
In fact, Dumezil's scholarship traces a trifunctional separation of priest and warriors over husbandmen, and India and Europe maintained the intellectual separation of the position of priesthood and coercive power well into the historic period, while on the other hand the Midddle East and China/Japan have sacred and worldy power fused, are essentially ceasaropapist in political origins and retain strong vestiges today, as Fukuyama highlights.
So 'IE culture' is undeniable, no matter how complex the social milieu.
@David: Sure, I know it's a simplification and that more data is needed but anyhow the Lazaridis triangle at least is implicit in the PC graph, so I wonder if the other triangles I tentatively inferred are also valid. Particularly relevant is the case of the "Reich triangle" in which we can skip ANE (a poor reference) and replace it directly with Yamna (a good one). The Yamna genesis is surely a bit more complicated but still worth wondering about.
ReplyDelete@ Mike Thomas
ReplyDeleteLast of all, I think you can forgive us if we doubt the usefulness of paleoanthropology/'advanced anthropological models' in the understanding of human history. That entire field is absolutely notorious for its political bias from the time of Boas. It tried to whitewash human history of any sign of violence or migration, and had to be literally strongarmed into accepting theory after theory it had formerly anathemised/villified after evidence--most often from other fields--came to contradict it directly.
When evidence of widespread cannibalistic infanticide and domestic violence in Australian Aborigines surfaced, it was censored for three decades. When people raised the hunting hypothesis, they were villified. When Jane Goodall uncovered warfare in Chimps, she was castigated. When people investigated death rates and reprod. success of tribal warriors, they were ostracized. When the bonobo sex utopia was described in Sex at Dawn it was promptly refuted by primatologists. Now we know better.
The interpretation of the neolithic by anthropologists/archaeologists was literally forced to change to migrationist by ancient DNA, and, if you count Cavalli-Sforza and mtDNA, shocks to the discipline have been forced in this way no less than three times. The evidence in archaeology was always the same, but it was genetics that provided more solid information than the countless volumes of 'diffusionist' vs 'migrationist' debate/interpretations could ever give.
Until recently everyone thought that people all over north Eurasia descended mostly in a stream of paleolithic local continuity from 'interpretation', and this was completely overturned. It turned out that archaeological groupings reveal precious little about the autosomes of those concerned.
Since places like UCL are already fusing data crunching on pop gen, aDNA and archaeology in wholly new methodologies, I have no doubt that future publications in those esteemed journals of yours will be affected by the results here, in time.
Maju,
ReplyDeleteI didn't pick the ancestral populations that form the K8 triangle.
The ANE, WHG and ENF populations have left very strong signals in modern genomes, so much so that even after tens of thousands of years I can tell that the Karitiana are ~42% ANE, Lezgins ~27%, Lithuanians ~18%, and so on.
These results can be reproduced using different samples and different marker sets, and I don't even need ancient reference samples to get the correct ANE scores, more or less anyway.
But without such strong signals there's not much I can do. Kostenk14, for example, doesn't seem to want to trigger anything like ANE or WHG.
The Karelian/Samara EHG might well form a cluster in Admixture, but if I analyze them with the K8, I have no doubt they will just turn out to be a simple mix of WHG, ANE and maybe East Eurasian.
I don't think Yamnaya samples will form a cluster. They'll be very mixed, like modern Eastern Europeans, IMO.
Ryu, you have misunderstood my commentary.
ReplyDeleteI am not critiquing the study of genetic evidence per se, but I am highlighting the gap between advanced anthropological theory and that proposed by geneticists. Naturally, one cannot wholly blame them as their field, afterall, is genetics and not social anthropology, or the like.
But before you get on your high horse on the social sciences, it is of course obvious that it has been abused by nationalist goals or political agendas. But then again so was physical anthropology, arguably the forerunner of genetic anthropology, even more so. And of course, today, one could argue that the anti-migrationist (archaeological) bias is a lefteist reaction in the post-war period, and sure it is, to a certain extent. But today's post-processual archaeology and post-modernistic discourse on the nature of ethnicity, identity, etc; only exists because of the very recognition of previous flaws with the field. THis is a sign of a discipline in a healthy state of self-criticism.
You say "Until recently everyone thought that people all over north Eurasia descended mostly in a stream of paleolithic local continuity from 'interpretation', and this was completely overturned"
Clearly you're mistaken. Even before the advent of genetics people were assuming (rightly or wrongly) waves of invaders based on skull shapes, etc. Are you unaware of this ? ?
I can only hope ancient DNA will make more impact. All evidence is good evidence. However, this has much further to go. The geneticist role models, and their community blogger-followers, first of all need to cease being 1 dimensional analysts, and academically 'grow up'. Based on the level of knowledge I have seen here , that will take many years. So for now, you can keep making your nice little Indo -European packages of **Corded Ware = Indo -European ethnic-historical community = ANE = R1a**, but you will be shown that this is simplistic *at best* , or frankly incorrect at worst.
But don't take my word for it; a quote from but one of millions of academics: '"In the past generation or so, archaeologists have recognized that boundaries of
archaeological cultures, based on material culture traits, do not neatly correspond to
how the people themselves perceive social, cultural, and ethnic boundaries. To
complicate matters further, the old sense that identities are discrete and long-lived
has been seriously challenged. Instead, anthropologists now consider identity to be
situational and relational and in the constant process of making, unmaking, and, sometimes, disappearing" ( D Hu; 2013).
This has actually been known since the 1960s. Hopefully, you boys will join the 21st century on this matter.
As for your analysis of DNA, I cannot doubt that there are many bright sparks here and I commend you dedication to the topic. But imagine how much more brilliant you'd all be if you got other disciplines up to scratch ?!
As a final point, I m sure that aDNA will contribute even more than to questions of mass migrations. Looking at kin structures, relations, etc within settlements and burials,as has already been done, will certainly assure a happy marriage between the natural and social sciences.
If you do supervised ADMIXTURE analysis, you always end up with the analyzed pops. being expressed as mixture of the fixed reference ones, be them real or "zombies". I understand that you know that, right? So if you don't use Yamna as reference, the results won't be expressed in Yamna terms, and if you don't use Basques or Gokhem as references the results will not reflect that possible origin but something else. This is the main problem I see with your and Dienekes' fixed reference exercises, which I consider unreliable unless other data backs the results (which sometimes may happen but not always).
ReplyDeleteThe key problem is why do you choose this or that reference population? Not always clear nor, I'd dare say, intellectually honest, really.
That's why in studies like Lazaridis' they choose other methods, namely the f4 formal test, which seems much more reliable and unbiased, providing us with those fascinating results.
Anyhow, why not to try to replace ANE by Yamna? Ma1 is indeed still important to understand Native American origins, which have nothing to do with Yamna/Kurgans/Indoeuropeans, but European origins have a much more direct connection to Yamna than to ANE, which can only be explained by mediation of Kurgan expansion (plus whatever leftover that can be attributed to Motala-like pre-IE admixture). I'm pretty sure that ANE can be scrapped almost totally if we use Yamna instead as reference.
It'd be also interesting to use Yamna to evaluate South Asian populations (assuming other references are approximately correct) and see how well the results correspond with those using ANE. I say this because I don't have clear at all how reliable are your results using ANE and, if ANE affinity or part of it has to be attributed to IEs, then Yamna is no doubt the reference to use, because Yamna is closest to the original PIE and almost undoubtedly ancestral (at least culturally) to Andronovo and hence historical Indo-Iranians.
@Maju
ReplyDelete''It'd be also interesting to use Yamna to evaluate South Asian populations (assuming other references are approximately correct) and see how well the results correspond with those using ANE. I say this because I don't have clear at all how reliable are your results using ANE and, if ANE affinity or part of it has to be attributed to IEs, then Yamna is no doubt the reference to use, because Yamna is closest to the original PIE and almost undoubtedly ancestral (at least culturally) to Andronovo and hence historical Indo-Iranians.''
Andronovo is not Indo-Iranian its Just Pre-Scythian-Sarmatians likely mentions In Avesta As Tuirya Nomads fond of the Horse etc....
I will also look forward to see that how much Yamanaya type Ancestry The South Asians Will Score....
BTW aDNA will be trillion times more decisive.
Maju,
ReplyDeleteIt is possible to run the type of supervised test you describe, but to do that I'd have to limit the number of test samples to just a handful, and ideally also have very solid reference panels.
Using single ancient genomes in such tests doesn't work well at all, especially if at most one of these genomes is low coverage. The results never resemble formal mixture stats very closely. They usually just look like garbage.
The type of test that I run to flesh out ancient components like ANE and WHG is based on a few well chosen reference samples - be it ancient genomes, modern individuals or synthetics with certain admixtures removed - and a large dataset of test samples. But these reference samples are really nothing more than pointers and the algorithm decides whether they reflect the patterns in the given dataset.
If I mark a chimp as one of the reference samples in a dataset of humans, I can assure you that I won't get a chimp cluster. The algorithm will just find a more suitable signal in the context of the number of K specified and the dataset it has to work with and just completely ignore the chimp.
The fact that I can get reasonable estimates of ANE ancestry based on a dataset of well over 2,000 individuals and 220,000 SNPs is amazing. If there was no clear signal of ANE in the dataset I'd get something else that might be ANE related, like an Amerindian, Kalash or Caucasus cluster, but maybe not even that.
So I can try and run a very limited supervised test with the Yamnaya samples when they become available, especially if there's a decent number of them. But I doubt the result will be very useful. I'd say that running the Yamnaya with the K8, getting their ANE/WHG/ENF proportions, and then trying to fit modern Europeans as mixtures of Yamnaya and other ancient samples based on these proportions will be more informative.
@ MT
ReplyDeleteI look forward to that as well. Hopefully we can have a very civil discussion on social processes when societies interact once we have a panoply of aDNA across time in different archaeological contexts.
Greg clark, an economic historian who you might want to read, discovered that social status is highly heritable in ancient societies, and that out-reproduction of the lower classes by the elites of settled agricultural societies is common, and that elite overproduction led to downward social mobility and further downward displacement of the lower classes by scions of the elite. The overall effect was to result in genetic turnover over generations as the population comes to resemble the elites more and more. I wonder if such a process was in operation as the no doubt small number of IE-speakers created those 'chiefdoms' and hierarchical proto-polities over peoples that outnumber them.
The other possibility is something like the interaction between Comanches and American settlers, or the Mongols and the Chinese in the Ordos, or Japanese pirates and the Ming coast, where continual raids resulted in virtual depopulation.
@ Davidski
Do you have time now to do ADMIXTURE before the paper comes out? I know this is highly off-topic, but HUGO Pan-Asian finds a Hmong-mien admixture component that is intriguingly elevated in Japanese and Koreans to the exclusion of North Han, who have almost none. This is despite the fact that the Sinitic component and other SEAsian components are lower, and the Ryukyuan/NE Asian components are higher, in Japanese and koreans compared to Han.
In the multiple runs in the SI, certain runs have the Hmong component appear in a fraction in in North han as well, and when that happens that component literally explodes in Japanese and korean.
This is very interesting, because as we know O2b has an ultimately southern provenance, and modern Japanese and Koreans mostly descend from a population that introduced rice farming, which once again is of southern provenance. Northern China had millet, and rice was not important till later. Lastly the rice/agri vocabulary in Japanese and Korean cannot be traced to any language on the continent.
If possible, could we do an ADMIXTURE experiment on a subset of the HUGO-PanAsian incl. N+S Chinese, Korean, Japanese, Ryukyuan, Miao, Atayal, Dai, Mal, and Kensiu? This can also be preparation for when we get the Jomon genomes.
Thank you so much if you have time for this.
Where is the HUGO-Pan Asian dataset? I'm having trouble finding a live link.
ReplyDeleteThank You!
ReplyDeletehttp://www4a.biotec.or.th/PASNP
I've put in a request to access the PASNP database.
ReplyDelete@ Davidski
ReplyDelete:)
David,
ReplyDeleteFor what it's worth, the Kensiu and Jehai populations found in the PASNP database seem to be closely related to the Onge, so they should constitute good ASI references for South Asians. It would be pretty cool if you could run a ADMIXTURE test with this data-set that would include the K8 ANE cluster, but I'm not sure if the SNP overlap is good?
I can try and run them with the K8 if and when I get the dataset. Haven't got a reply yet.
ReplyDelete@ Sein
ReplyDeleteThe overlap is very bad with most datasets if I remember correctly.
Nevertheless Zack managed to run HUGO merged with some HarappaDNA South Asians on only 5400 snps, and got results consistent with others.
http://www.harappadna.org/2012/03/pan-asian-ref3-k11-admixture/
A synthetic Onge created from this would still be quite bad I suspect.
@Chad Rohlfsen
ReplyDeleteWho are we? And on what do you base it? The frequencies simply don't fit. A "why wrestle it" is not any answer to my concerns/arguments.
I also heard by other users who said Russian H&G were more like ~50/50 ANE, WHG.
@ btw David, if you do have any time to download any of the Jorde Lab's public data (http://jorde-lab.genetics.utah.edu/?page_id=23) on Tibetans, and test via K8, that would be appreciated, if there is overlap between the SNPs they have and you have.
ReplyDeleteE.g. the available data for "Shared and Unique Signals of High-Altitude Adaptation in Geographically Distinct Tibetan Populations" 2014 or "Genomic analysis of natural selection and phenotypic variation in high-altitude Mongolians" 2013, with Maduo and Qinghai Tibetans.
If ANE is beyond what would be expected for a population most similar to Han (North Han average 2.5% ANE, Mongolian average 10% ANE) that might give us an indicator of whether the "High Altitude Adapted" population that Tibetans get ancestry from via the Sherpa had any ANE ancestry (I think it's probably not but it would be interesting if it was).
Krefter said... "...much of western Europe before Rome was non-IE."
ReplyDeleteNo it wasn't. Outside of pockets in NEW Spain and Italy, most of Western Europe was already IE speaking long before the Romans.
Kurti,
ReplyDeleteThey can't be 50/50. If they were, mixing with an Armenian would make yamnaya over 30% ANE. This isn't possible with keeping corded ware at 66%+Yamnaya and in Northcentral Europe. It can't be done mathematically.
75% of Yamnaya is south of Samara. A hunter group at 65/35 make more sense, with incoming pastoralists picking up or having about 15% ANE on their way north.
ReplyDeleteIt is possible that southern hunters near the Caucasus had more ANE and less WHG if the northern Caucasus populations are an indication. It also leaves room for a near eastern and ANE mix coming up out of western Iran. Let's just wait for the paper.
ReplyDelete@Rocca:
ReplyDelete"Outside of pockets in NEW Spain and Italy, most of Western Europe was already IE speaking long before the Romans."
No, not "most" and and certainly "long before the Romans". At the arrival of written History most of Iberia and much of what is now France were not IE-speaking and all IE known there was Celtic, which in most cases had arrived only in the few previous centuries (La Tène) and in the last millennium at most (Urnfields-Hallstatt, whose greatest impact was precisely in Iberia, not elsewhere). There were also important pockets of likely pre-IE areas in Italy and the nearby islands. The exact ethno-linguistic situation in Britain and even "Gaul" itself in some cases is not well known (at least not before Caesar) but the usual archaeologically based consensus is that Celtic only arrived with La Téne, which pre-dates Hannibal by 200 years at the most and Caesar by 350 years or less. So barring Celtiberians, Lusitanians and possibly Arvernians, Indoeuropean (Celtic in essence) only arrived to West Europe c. 400-300 BCE.
Don't doubt the full Beaker package being IE. Bell Beaker will plot by Corded Ware. It certainly could've been pre-Celtic. Basques needn't be IE speaking with 80% R1b. There more than a few Native American tribes that are just like that, yet retain their own culture and language.
ReplyDeleteBronze age warrior from Poland
ReplyDeletehttp://dienekes.blogspot.com/2015/01/bronze-age-warrior-from-poland.html
They tested his DNA, but i heard nothing about a paper or genetic info. Would he be Corded ware, or a descendant of Corded ware?
@Chad: that's nonsense. It's not just R1b (whose aDNA origins and spread still remain utterly mysterious) but mtDNA. Basques are a stable population since the Neolithic.
ReplyDeleteAlso, what do you say about Iberians, Tartessians, paleo-Sardinians, Ligures, etc.? It's a wide area and we don't know so well the details of what was spoken in other regions further north but nothing point to IE before La Tène.
I was just reading on how, per the accounts of the time and lack of Celtic survivals in English (there may be more Vasconisms than Celticisms, oddly enough) the language shift from Celto-Roman to Anglo-Saxon in Britain must have been radical, partly because of colonization and partly probably because of brutal subjugation. Why can't you accept that the same happened as the Celts moved westwards with La Tène (and the previous more tentative migrations towards Iberia)? 300-400 years are enough to totally change the language and identity of a people, you can find many historical examples.
Anyways, what we see here with all this aDNA data is that there was a quite impacting Kurgan migration (in various phases) and that Indoeuropean languages in Europe must be attributed to it. Bell Beaker is now quite clearly demonstrated NOT to originate in the Kurgan (Corded Ware) area but in Iberia (several centuries earlier). The Central European branch of it may mean whatever (I'm not sure exactly of its implications) but it's not the expression of an East-to-West early IE migration in any case. Actually I don't think Bell Beaker is that important, that's another of Jean Manco's fetishes that have been debunked (but are still oddly popular in these forums, so populated by "Indoeuropeans" who feel their ethnic identity threatened by these discoveries): (1) archaeo-statistical data does not suggest major demic impacts with BB (except in Ireland) and (2) the BB symbolical flow was from SW to NE and not the other way around.
Maju,
ReplyDeleteMaybe Celtic for Iron age west Europe is like Slavic for medieval east Europe. Meaning Celts expanded in territory where the natives already had significant Yamna-type ancestry, and some could have been Indo European speakers.
Southwest Europeans(which makes sense with what you said about non-IE languages), are similar to Neolithic west Euros, but they are also shifted to the east towards MA1.
British and Irish are about as Yamna-like as Corded ware descendants, and so there's no way before Celts British-Irish were unaffected by eastern migrations, unless La Tene exterminated the natives.
@Krefter: but how would IE arrive there? I understand that you are archaeologist, which cultural flow could have brought anything IE/Kurgan westwards before the Urnfields and La Tène expansions? There is NOTHING!
ReplyDeleteMany have argued in the past that Bell Beaker could be that but nowadays it seems very apparent that BB symbolism flowed in the opposite direction. There's nothing else unless you want to go for what I call "Indoeuropean pan-spermia", by which IE was spoken simultaneously from Ireland and Portugal to Nagaland and Sri Lanka since "always". Believe it or not there are some scholars tempted by this nonsense but it's utter nonsense nonetheless. Indoeuropean must have expanded and diverged from some common origin and the most parsimonious theory by far is the Kurgan model. The genetic data we are seeing here and before that in Lazaridis, etc. only confirms this.
Krefter,
ReplyDeleteThat Bronze Age Pole might be from the Unetice Culture, but I guess we'll find out soon when the full results are released. I'll write up a blog post when that happens.
@Chad Rohlfsen
ReplyDeleteSorry but the way you underline your argument doesn't convince me. Actualy you simply ignored my whole arguments and didn't even go into one of them and all you did was say "why wrestle it" Certanly a half Armenian wouldn't turn Yamna to 30% ANE but an half Armenian would no way turn an Yamna individual into only 35% ENF. You also seem to go down with your ENF estimation down from one post to another.
An half Armenian, Half Karelian (if the Karelian was indeed 80% WHG which I doubt) would turn Yamna ANE to less than 17%. And this is less amount in modern Northeast Europeans. So how are you going to explain that? No what you are going to do is simply ignore this.
I am not going to sum up all my arguments which make your estimation unlikely. So I only tell to reread my comment and please only answer my post if you are willing to answer my questions and not put a phrase like "why wrestle it".
Thanks and Good Day.
@Maju - My understanding was that Ma-1 was more closely related to the European ANE ancestry than Amerindian ANE ancestry, no?
ReplyDeleteI'd think he'd be especially relevant for studying the later migration to the Americas too. Mitochondrial haplogroup C4a'b'c is present in early kurgan burials, and also common specifically in northern North American aboriginal populations. While Ma-1 belonged to haplogroup U, I wouldn't be surprised if he had some strong affinities to the the people belonging to haplogroup C4a'b'c both in kurgan burials in Eastern Europe and in present day North American aboriginal populations like the Dene and the Algonquins.
http://www.pnas.org/content/110/35/14308.full.pdf
@Davidski Re: the spread of R1a, I'm not sure why you're still so skeptical of this ancestry passing through the Iran/Turkey region on its way to the Ukraine. It's supported by archaeology (the spread of micro blade technology from Lake Baikal to the Zagros - something I only found out about because of you referring to it on this blog) and genetics (both this paper and the Underhill paper). If R2 made it to South Asia somehow, I don't see why it's implausible that R1 made it to the Zagros mountains. That doesn't exclude it from passing through Siberia earlier.
Re: the Basques' ANE - let me throw out a few hypotheticals here that I'd love to hear people's thoughts on.
Suppose there are a group of nomadic hunters living on the mammoth steppe ~20-30 kya. Suppose they are HIGHLY mobile, and are dominated by haplogroups C-M217, P, Q and R (as the latter two emerge), and are mostly genetically homogeneous, with a bit of an East-West cline, and with C-M217 entering the population somewhat late from the south east. Suppose they follow migrating megafauna - their major food source - where those megafauna migrate.
Now, as we get later and later in the Pleistocene, the megafauna begins to die out. The mammoth steppe changes from one large continuous expanse to disconnected patches of surviving megafauna. The range of these mammoth steppe hunters collapses into isolated pockets as well, before eventually shifting to other species (ie bison, gazelles) as the mammoths and others go extinct. This collapse is fairly rapid, and changes the population structure from one that is highly homogeneous to one that is highly structured between each of these surviving pockets.
Here's a map of the different ecozones at at the LGM and at the start of the Holocene - http://a.foucault.free.fr/HotSprings_col.jpg
So suppose one such pocket is in and around the savanna-like grasslands in and around the Zagros at the start of the Holocene, hunting bison and gazelle and the like, and that this group is mostly R1a. This population may also be present in the grasslands around the Caspian.
Another pocket is along the northern fringe of Eurasia, following the retreating steppe, and where mammoth populations briefly expanded before crashing around ~11-12 kya. This group is mostly haplogroup Q with some C-M217 along its south eastern edge.
Another pocket is in the grasslands of southern Spain and the Morroco. This group is mostly R1b.
Another would be around Mongolia, with high levels of C-M217, and India could have received R2 this way too.
Is this totally implausible? It strikes me that if we look at the plausible distribution of haplogroups Q and R at 9 kya, it corresponds to places that are 1) savana and 2) were either tundra/steppe at 21 kya or adjacent to tundra/steppe at 21kya, and haplogroups N, I and O correspond to places that were forested and hence not conducive to nomadic hunters of megafauna.
Am I out to lunch?
ReplyDelete" Bell Beaker is now quite clearly demonstrated NOT to originate in the Kurgan (Corded Ware) area but in Iberia"
Only half true. Certainly BB has little to do with Kurgan, but the BB did not "begin" in Iberia. Apart from Iberian archaeologists who emphasize the early C14 dates for Sites in the Targus, there is still advocates of the influences from the low Countries. Whatever, the case, the BB was a pan-West European phenomenon from the outset, and not a spreading wave from Iberia.
Maju, whilst I agree in that most of West Europe was likely non-iE until the 1st mill BC, I fear that your rationale for this might derive from Vennamans' hogwash.
Forgot to include this link in my previous post: http://www.nature.com/ncomms/journal/v3/n6/images/ncomms1881-f1.jpg
ReplyDelete@Mike - R1b does seem to have swept over western Europe around the late Neolithic / Chalcolithic though. Where did it come from?
Adding to the mystery is R1b's presence in Cameroon too.
It must have come from somewhere. I do think the pre-IE ANE signal being discussed here is real, and I think it would make sense that both it and R1b are related. Perhaps there was reservoir of ANE heritage isolated somewhere that we haven't yet discovered. Iberia seems as likely a place as anywhere.
I'd think the appearance of vigesimalism in Celtic languages is pretty good evidence for some sort of Basque-like substrate.
Ryan,
ReplyDeleteThere's no evidence at the moment that Near Eastern R1a is ancestral to European R1a.
The last Underhill paper doesn't provide such evidence, even though the authors think that it does, and many people believe them.
The data in that paper show the presence of a rare form of R1a-M420* in the Near East, which, judging by its haplotypes, is the result of a recent founder effect in the region.
We don't know when and how that R1a-M420* ended up in the Near East, and what its origins are. But it's not ancestral to R1a-M417, and thus to European R1a.
Also, like I said, why would a population that lowered the level of ANE on the steppe bring R1a to the steppe? Mal'ta boy, the main ANE proxy, belonged to R* and lived on the steppe.
The Near East is not a plausible source of Y-DNA R for Europe. To me it looks like a sink.
Also, Iberia could not have been a reservoir of ANE, because if it was, then we'd hear about it in this last Reich abstract, because it features samples from Spain.
Kurti, it doesn't go down with each post. 38% Near Rastern is what we settled on a few weeks ago. Think and say what you wish. It makes no difference to me.
ReplyDeleteMaju, you will see with the Beaker genomes. NW European like with different uni-parental markers from Corded Ware. R1b is not from some Atlantic refuge.
Davidski,
ReplyDeleteThe bronze age Pole can't be Unetice because he was from the eastern edge of Poland.
http://www.anthrogenica.com/showthread.php?3732-Reconstruction-and-possibly-a-genetic-analysis-in-the-future-of-4-000YBP-Pole&p=65530#post65530
There are Unetice burial sites as far east as present-day western Ukraine. There was a study done on them recently.
ReplyDeleteHas everyone forgotten about the 4000YBP guy from Portalon Spain?
ReplyDeleteRemember he clustered east of Basque, and by Tuscan? There's a good chance he had ANE ancestry, and more near eastern than modern Iberians.
He could have been part near eastern or east European.
@ Ryan "@Mike - R1b does seem to have swept over western Europe around the late Neolithic / Chalcolithic though. Where did it come from?"
ReplyDeleteI honestly can say I don't know any over anything that I've read from the commentary by academics and bloggers alike. My guess would be via West Asian highlands, Southeastern Europe, etc.
but its difficult to say with virtually 0 ancient DNA from SEE (also w.r.t. R1a)
ReplyDeleteKrefter, this was also posted by Davidski via his post on the UISPP Burgos 2014 abstracts:
ReplyDelete"Five of the 10 individuals analysed have not yielded sufficient coverage for genomic analysis. We present low coverage genomic sequences (average depth between 0.2 and 1%) of five early Iberian farmers dated to between 4,000 and 5,000 years old, from El Portalón. These individuals display a similar pattern to that observed for central and northern European farmers and all show genetic similarities to modern-day southern Europeans, particularly to Sardinians, in contrast to the recently published 7,000 year old hunter-gatherer from near-by La Brana in Spain."
No ANE in Spain at 2000BCE... Probably no R1b either.
ReplyDeleteDavid, where do you think R1b "came from"? If it is shown to have come from the north of the Black Sea, then I'll readily accept defeat.
ReplyDeleteIf R1b was from West Asia that late, it wouldn't be able to create NW Europeans by mixing with a Gok2 type.
ReplyDeleteR1b came from Central Asia. It somehow made it to present-day Bulgaria by the Copper Age, from where it expanded into Western Europe via the Carpathian Basin and the Near East via Anatolia. However, some of the R1b in the Near East arrived there directly from Central Asia, even as late as the Turkic migrations.
ReplyDeleteAgree . Is there actual day from copper age Bulgaria?
ReplyDelete*data*
ReplyDeleteRK,
ReplyDeleteThanks for the link!
I wish there was some way to just use those Onge samples. In Zack's Reference 3 K=11 run, the Onge cluster pretty much acted as a perfect ASI reference. East Asian and Papuan noise disappeared from all South Asian populations, even tribal South Indians. Without the Onge, ASI tends to leak into East Asian and Oceanian percentages for South Asians, even with an ASI-like cluster present (as is the case with West Eurasia K8's "South Eurasian" cluster).
Then again, perhaps it's not a big deal, since we already know how various populations breakdown in terms of actual ASI. In my general neighborhood, the average Non-Punjabi Jatt was around 20% ASI, the average Punjabi Jatt was around 15% ASI, the average HGDP Pashtun was around 11% ASI, the Baloch were around 7% ASI, and a Persian individual from the Iranian province of Khorasan was 3% ASI. Based on how these things correlate, Tajikistani Tajiks are probably 4% ASI, in this scheme. HGDP00244 is almost a carbon copy of myself in genetic terms, something which has been demonstrated again in our K8 results (we are pretty much identical). HGDP00244 is 7.87% ASI, so I guess I'm 8% ASI in this context. Most western Afghan Pashtuns are probably around 7% ASI, ranging from 5% to 10%. Scheduled caste South Indians were pretty much 40%-45% ASI.
Although we didn't realize this at the time, but the "South Asian" cluster in K=11 was strongly ANE-shifted. It was very high among Lezgins (and other Caucasians), and peaked among the Kalash. In terms of Fst, it behaved just like David's K8 ANE cluster (similar distances to other comparable components). At the time though, we were really confused as to the nature of this component. Ancient DNA has clarified so much which was previously muddled and confusing.
Davidski - "We don't know when and how that R1a-M420* ended up in the Near East, and what its origins are. But it's not ancestral to R1a-M417, and thus to European R1a."
ReplyDeleteFair enough. They're operating under the assumption that the greatest diversity will be found close to the origin, but the diversity in the Zagros may just be because it is from one of the earliest migrations from the Urheimat, with that diversity subsequently being lost in the steppe due to the various demographic turnovers there.
It really won't be settled until we get many more ancient DNA samples from the steppe, the Zagros and Siberia that predate the expansion of R1a.
I'd like to point you back to this post of yours from last year again though: http://eurogenes.blogspot.ca/2014/01/another-look-at-lazaridis-et-al-ancient.html
You say you don't really know where the Near Eastern R1a comes in. The microblades figure from that posts suggests it could also be a parallel migration from Siberia that occurred simultaneous to a migration to the Pontic steppe. Just throwing that out there.
We're not really talking about the Near East there though so much as we're talking about the interface between the Caucasus, the Near East and Central Asia.
"Also, like I said, why would a population that lowered the level of ANE on the steppe bring R1a to the steppe?"
A counter intuitive suggestion I suppose. The introduction of agriculture to Europe seemed to bias the genomes of the population in favour of the migrants in terms of the Y DNA, and in favour of the indigenous people in terms of the mitochondrial DNA. Why would we expect something different in the Ukraine?
But yah, I am arguing that the surviving R1 lineage came from the group that had less ANE heritage (though not none), which I suppose isn't all that logical.
Re: R1b in Spain - the coverage in terms of samples isn't exactly impressive. There doesn't seem to be evidence of ancient R1b anywhere else that I'm aware of either, so it must be "hiding" somewhere unsampled so far.
So we just haven't been lucky enough to find ancestral R1b yet, or it was present in an area that is now inaccessible - under water or in the Sahara or something, though I think that's unlikely.
Definitely a big mystery so far. It seems to predate IE expansion to the region at least, so while IE is probably the largest source of ANE in Europe, it may not be the only one.
Just to clarify, I meant the "average non-Jatt Punjabi". Although, Punjabi Brahmins were 16% ASI.
ReplyDeleteMike,
ReplyDeleteI'm not aware of any Copper Age genomes from Bulgaria. From memory Chad was in contact with Reich recently trying to talk him into testing some samples from Ezero and related cultures. I'm not sure how that went.
But as I mentioned before, I have seen PCA with two good quality Bulgarian genomes, one of which was from the Bronze Age. They both clustered between Spain/Tuscany and Sardinia on a PCA that basically looked like my own, so my impression was that they were largely derived from a Neolithic Sardinian-like population that was shifted east by an influx of ANE into the Balkans during the Bronze Age or a little earlier.
Ryan,
There are some basal lineages of Q in Iran and of course the ancestral mutation of R1b-V88 probably arrived in the Near East fairly early too, so these would be my candidates for the markers of the groups that introduced microblade technology to the Near East.
Sein,
ReplyDeleteI think the Onge component underestimates the true native South Asian input into present day South and South Central Asian populations, because it's based on a tiny and extremely isolated group located a fair way east of India.
I suspect that ancient DNA will show indigenous South Asian foragers to be Onge-like, but shifted closer to ANE. In other words, I'm expecting the native South Asian component to be almost intermediate between the Onge component and ANE, albeit much closer to the former.
David, do you think that high ANE in those pops is due to the lack of an ancient genome there?
ReplyDeleteBtw, paper on Dagestan..
http://www.nature.com/ejhg/journal/vaop/ncurrent/full/ejhg2014299a.html
Chad,
ReplyDeleteNo, I think South Asians do carry ANE at levels of up to 30%, and South Central Asians at around 30%.
I just think that the real levels of ASI are higher in these regions than what the Onge component suggests, because actual ASI isn't the Onge component.
It seems to me that what happened in the ANI/ASI Reich paper is that the real ASI was split between ANI/ASI because of the use of the Onge references, with some of it, the most ANE-like part, going into ANI. So ANI was actually a mix of ENF, ANE and ASI.
South Central Asia might have been a contact zone between ANE and ASI for a very long time, or perhaps these two components, and even WHG, derive from the same basal branch?
Ancient genomes from South and Central Asia are needed to clarify this. Using modern isolates like the Onge only confuses the issue.