search this blog

Thursday, February 4, 2016

Ancient European mito genomes suggest single major dispersal of non-Africans + Late Glacial population turnover in Europe


There's an important new paper at Current Biology on the peopling of Europe. The big question left open by the authors is where exactly did Western European Hunter-Gatherers (WHG) rich in Y-HG I2 and mtDNA U5 come from if, as the authors suggest, they weren't native to Western Europe. According to them it was "another, separate LGM refugium". Balkans?

Summary: How modern humans dispersed into Eurasia and Australasia, including the number of separate expansions and their timings, is highly debated [ 1, 2 ]. Two categories of models are proposed for the dispersal of non-Africans: (1) single dispersal, i.e., a single major diffusion of modern humans across Eurasia and Australasia [ 3–5 ]; and (2) multiple dispersal, i.e., additional earlier population expansions that may have contributed to the genetic diversity of some present-day humans outside of Africa [ 6–9 ]. Many variants of these models focus largely on Asia and Australasia, neglecting human dispersal into Europe, thus explaining only a subset of the entire colonization process outside of Africa [ 3–5, 8, 9 ]. The genetic diversity of the first modern humans who spread into Europe during the Late Pleistocene and the impact of subsequent climatic events on their demography are largely unknown. Here we analyze 55 complete human mitochondrial genomes (mtDNAs) of hunter-gatherers spanning ∼35,000 years of European prehistory. We unexpectedly find mtDNA lineage M in individuals prior to the Last Glacial Maximum (LGM). This lineage is absent in contemporary Europeans, although it is found at high frequency in modern Asians, Australasians, and Native Americans. Dating the most recent common ancestor of each of the modern non-African mtDNA clades reveals their single, late, and rapid dispersal less than 55,000 years ago. Demographic modeling not only indicates an LGM genetic bottleneck, but also provides surprising evidence of a major population turnover in Europe around 14,500 years ago during the Late Glacial, a period of climatic instability at the end of the Pleistocene.

Posth et al., Pleistocene Mitochondrial Genomes Suggest a Single Major Dispersal of Non-Africans and a Late Glacial Population Turnover in Europe, Current Biology (2016), http://dx.doi.org/10.1016/j.cub.2016.01.037

192 comments:

Krefter said...

More prove Palaeo-Europeans were essentially 100% mtDNA U(xK). There's lots of new samples, I think the only non Us and are a few Ms.

Ryan said...

I guess this makes sense given La Brana's Y chromosome and autosomal DNA affinities to SE Asia.

Tone said...

My bet: the Paleo Europeans before 14k bc are probably Y-HG "C" and MtDNA "M" and then WHG comes in with I2 and U5. But that's just a guess.

Krefter said...

@Ryan,
"I guess this makes sense given La Brana's Y chromosome and autosomal DNA affinities to SE Asia."

All the mtDNA Ms are 27,000+ years old. They also had extinct forms of U, like U2* and U8a*. mtDNA M itself might be 60,000+ years old. The mtDNA M in Paleo-Europe and modern SE Asia, split probably over 60,000 years ago. So, this mtDNA U doesn't suggest any relation to SE Asians.

Krefter said...

@Tone,
"My bet: the Paleo Europeans before 14k bc are probably Y-HG "C" and MtDNA "M" and then WHG comes in with I2 and U5. But that's just a guess."

This study has loads of mtDNA U(U5, U2, U8a) from Paleo Europe. Some even close to 40,000 years(Kostinki U2*, Romania U*). Just because C and M are exotic to West Eurasia, doesn't mean an extinct Eurasian population had them. That's the logic your using.

Romulus said...

I think the change in mtDNA fro pre to post LGM is simply the result of a population bottleneck which is well documented. As far as Y-HG C goes, it probably arrived with the Solutrean culture which was a discontinuity with the previous Aurignacian, hence the similarity between Solutrean tool technology and North American Clovis Points (where Y-HG C is also found). Continuity from Aurignacian to Gravettian to Magdalenian is documented by their creation of art which was not found in Solutrean. The authors mention a 3rd LGM European refuge on the Eastern European plain which is probably the origin of I1/U5a and how U5a ended up in EHG.

terryt said...

Very interesting, especially regarding M.

"We unexpectedly find mtDNA lineage M in individuals prior to the Last Glacial Maximum (LGM). This lineage is absent in contemporary Europeans, although it is found at high frequency in modern Asians, Australasians, and Native Americans".

That actually makes sense now of M's apparent sudden appearance in SE Asia with no apparent route there.

"The different spatial distributions and TMRCA estimates of these two ancestral clades have been interpreted as evidence of an early spread of modern humans carrying hg M into Asia, perhaps via a southern route"

The paper virtually destroys this idea. I have come to agree with Maju on the matter and now see 'Indian' M1 as having originated with all the other M haplogroups somewhere near the Burma/South China border region. How it reached there may still be a mystery although now a possibility presents itself.

"The potential impact of climatic events on the demography,
and thus the genetic diversity of early Europeans, has previously
been difficult to quantify, but it likely had consequences for the
relative components of ancient ancestry in modern-day populations".

In other words climate shift pushed humans south, driving many lineages to extinction. Remnant haplotypes from an original movement east through Central Asia have become extinct. There has certainly been a 'southern migration' through India but it was primarily from east to west, apart from that of Y-DNA F.

"My bet: the Paleo Europeans before 14k bc are probably Y-HG 'C' and MtDNA 'M' and then WHG comes in with I2 and U5"

The article tends to indicate that U was present along with M. But I tend to agree with your first point. Y-DNA C, related to 'Indian' C1b1a1 was found at Kostenki at 37,000 years ago, but the lineage is now extinct in Europe. As is mt-DNA M.

Krefter said...

Here are the brand new samples this study tested.

https://docs.google.com/spreadsheets/d/1FAuTvCIb4YuBED--zf0G5ITYUpqUgN_TYmqy2FqDfT0/edit#gid=0

Pre-LGM Europe(older than 25,000 years old) so far is mostly U2* and U5*. Otrher extinct forms of U have popped up, and then there's also N*, R*, and M*.

14,000-7,000yo (West)Europe mtDNA all looks the same. Almost 100% have U5b. mtDNA from Russia during the same period is U5a+U4+U2e.

Rob said...

What a great paper.

@ Dave

" Balkans?"

I'd bet, (a) Italy to SW Europe then north, (b) the eastern Danube region to the NW Europe, Scandinavia and the Dnieper basin.

@ Terry

I thought you'd like the mtDNA M :)
Let's also not forget that Oase was possibly some form of Y haplogroup K.

capra internetensis said...

Very interesting!

The longest sequence in any one place is the Belgian one, which goes from Aurignacian M, to Gravettian U2 (looks like some kind of pre-U2e maybe?) with one (pre-)U5*, and finally a Magdalenian U8a. That's the end of that site, but then in Mesolithic Northern France we have U5b1 and U5a2.

The transition at 14 000 years ago is in southwestern Germany and southeastern France, where it suddenly goes from a bunch of closely related U8a (plus one U2'3'4'7'8'9*) to all of U5b1 and U5b2 (and one U5a2). But there is nothing earlier than that.

In central Italy it goes from Gravettian U2'3'4'7'8'9* and U8c to Epigravettian U5b2b (earlier than the Magdalenian) to Mesolithic U5b2b1. So the transition seems to have been earlier here, whatever caused it.

I'm not convinced that the U5b wave had to come from somewhere exotic though, the Magdalenian U8a wave could just be from one area of the SW European refugium and the U5b wave from another. Or not, of course.

Aren't modern U2e and U8a European? I don't know much about mtDNA.

Ryan said...

Krefter - if his small SE Asian affinity comes from the male side of his lineage, his mitochondrial DNA being "normal" WHG isn't really a problem. La Brana still forms a pretty robust clade with the rest of WHG IIRC? La-Brana is someone with a small amount of SE Asian-like admixture, and a Y-chromosome lineage from coastal East Asia. Now we have S/SE Asian mitochondrial DNA showing up. I think it's unlikely the 3 aren't connected. Some sort of relic population population? A sort of dead clade walking that left an imprint on a limited number of hunter gatherers?

Romulus - Wrong haplogroup C I'm afraid. La-Brana's C-V20 is C1a2 and it's closest relatives are in Japan - presumably a Jomon lineage. In fact, other than a few C-V20 men in Europe, C1a is almost completely restricted to Japan.

The indigenous North American lineage is C-P39 - C2b1a1a. That's the complete opposite side of the haplogroup C tree. There's no relationship between the two lineages any more recent than the breakup of C itself, at least as far as we know.

Krefter said...

@Ryan,

C1a2 has been found in Neolithic Hungary and Turkey. None of those genomes with C1a2 showed an affinity to SE Asia.

Ryan said...

@Krefter - my R1b ultimately came from SE or S Asia (as haplogroup P), yet I don't show an affinity with SE Asia. Presumably some distant ancestor of mine did though.

Chad Rohlfsen said...

Y-C pre-dates the split of East and West Eurasians. There is no affinity to South Asians. The people that first entered Europe were the same as those that went everywhere else, like Ust-Ishim.

My bet is that WHG is simply Kostenki with a shit-ton of their own drift and going to their own pole, just as Native Americans do. 30k years in isolation and a severe population bottle-neck are to blame. There could still be some EHG in WHG, but nothing is certain yet.

Rob said...

Chad

'There could still be some EHG in WHG, but nothing is certain yet. "

But isn't 'EHG' somewhat of a fusion ? I.e. there is WHG in EHG, not vice-versa

Chad Rohlfsen said...

EHG rejects WHG admixture, but not by a lot. It looks like a drifted MA1, with minor ENA. EHG and MA1 form a clade in qpAdm.

Rob said...

Yes I recall the modelling in Haak 2015 showed 3 potential models for WHG - EHG- ANE. I guess we'll need to wait a little longer - for some Palaeolithic autosomal data from Europe & Russia - to clarify it further.,

Davidski said...

R1b ultimately came from SE or S Asia (as haplogroup P).

I don't believe this.

Chad Rohlfsen said...

I'm going to bet that if aDNA is taken, these samples will all sit between Kostenki and Loschbour, getting closer to Loschbour as time goes by. Europe was nearly depopulated. It's going to be a story similar to Native Americans. A group extremely drifted over time. The loss of M and near loss of C are due to the bottleneck. That is my bet.

Rob said...

Chad
I agree . That's what I thought quite a while ago; WHG possibly "evolved" from K14-like groups, "losing" some of the basalness in their isolation.

By the Clade models don't really suggest that tho, so they ? Rather they suggest K14 is proto-west Eurasian plus some other basal -stuff. It doesn't really cater to a "substractive" evolution

Chad Rohlfsen said...

Kostenki is close to basal West Eurasian. Also, close to Ust_Ishim. Kostenki does look like a cross between Loschbour and Ust-Ishim, solely due to age and not some ghost Loschbour pop that replaces them, IMO.

Rob said...

Possible. But the migrant group could have come during the Gravettian (30 kya).
I asked Ryu a while ago about the possible of WHG being a "drifted K14": if i understood him correctly, he argues a definite no.

If the WHG -ANE connection is true , then proto-WHG might have come ~ 30 kya, from the east (with proto-Gravettian industry), largely replacing the Aurignacionoid types such as K14, and their even more basal predecessors

terryt said...

"I thought you'd like the mtDNA M :)
Let's also not forget that Oase was possibly some form of Y haplogroup K".

Yes. I'm fairly sure the K was some sort of K2a, which relates it to NO rather than to P. I presume the Oase K was the product of a movement east to west through Central Asia.

"Wrong haplogroup C I'm afraid. La-Brana's C-V20 is C1a2 and it's closest relatives are in Japan - presumably a Jomon lineage"

Yes, even C1a and C1b are probably as different from each other as are O and R. Let alone the relationship between C1 (C1 is a southern haplotype apart from C1a) and C2 (American, but primarily East Asian). They just happen to have the same letter.

"Y-C pre-dates the split of East and West Eurasians. There is no affinity to South Asians. The people that first entered Europe were the same as those that went everywhere else, like Ust-Ishim".

But the earliest C we have in Europe is quite downstream and related to the South Asian version. And Ust-Ishim K2a is also very much a downstream branch. Unless you're going to postulate a single expansion of modern humans from a single region you have to admit that there were several movements over a very long period. Maybe only once out of Africa but then things get complicated.

"I don't believe this [R1b ultimately came from SE or S Asia (as haplogroup P)].

Pretty difficult to make a strong case against K2 having originated in SE Asia. All surviving clades have representatives there, some only there. And P is K2b2 and all K2b1 clades are Southeast Asian and most are east of Wallace's line. The only K2s not SE Asian are K2e in India, K2a in East Asia (or NO) and K2b2 in northern Eurasia and America (P). That's pretty convincing to me. On the other hand haploid genes are not necessarily a close reflection of diploid genes.

Krefter said...

The Paleo-U2 from Belgium(27,000 years old) is pre-U2e.

Davidski said...

It was difficult to make a strong case for a few sensible ideas using present-day DNA, but, as we now know, they were sensible because they were correct.

Let's wait for more Upper Paleolithic Y-DNA from Europe and Asia.

terryt said...

"It was difficult to make a strong case for a few sensible ideas using present-day DNA, but, as we now know, they were sensible because they were correct".

Do you not accept the recent shuffling of Y-DNA K?

Davidski said...

The present-day data from Southeast Asia is irrelevant to the origins of West Eurasian R. You can't look that far back in time with modern DNA with any certainty.

Romulus said...

@Chad

How else do you explain U5a in EHG? At some point before the Q-R split the ANE group wasn't travelling around with mtDNA U5, it is barely found outside of Europe. The qpAdm test probably doesn't work because I2 WHG split from I1 during the Gravettian, which also conveniently fits the Y Full date for the origin of I2 and U5a.

IMO here is where I1 and mtdna U5a come from:

https://en.wikipedia.org/wiki/Pavlovian_culture

In regards to a possible "Later" arrival of WHG, I don't think it is possible. There is demonstrated archeological continuity from the Aurignacian, to the Gravettian, to the Magdalenian, to the Azilian. They were all Cro-Magnon types who made the same kind of art and used the same tools. Bichon dates to the Azilian period and is a Cro-Magnon as stated in the paper.

Krefter said...

@Romulus,

I1 and U5a don't have a special relationship.

But, yeah I agree the U5a in EHG confirms they have sometype of common ancestry with WHG that post-dates MA1. Plus, EHG has a very close relationship with WHG that MA1 doesn't. I don't why Chad is saying EHG is a drifted version of MA1.

Romulus said...

@Krefter

You're forgetting about Stora Forvar, he was Pre-I1 and had U5a1 mtdna.

https://genetiker.wordpress.com/2014/05/05/analyses-of-the-stora-forvar-11-genome/
https://genetiker.wordpress.com/y-snp-calls-for-stora-forvar-11/

Chad Rohlfsen said...

Let me put it this way. WHG, as we understand it, did not exist at the time U5a developed. Even Bichon, at 11kya is barely in the clade of WHG. I do not count out the possibility that something like Kostenki could be in EHG, but it is not Loschbour, LaBrana, or KO1. This could be why it only marginally fails. I can test this on qpAdm. As far as EHG and MA1, EHG does fit pretty decently as fully MA1, while any WHG admixture into EHG is a failure. Now, modeling MA1 does work when making him a mixture of EHG and Ust_Ishim. I think this supports my idea that the movement from MA1 to EHG, and Kostenki to WHG is more about drift than admixture.

Chad Rohlfsen said...

Bichon is at 13kya, excuse me.

Chad Rohlfsen said...

BTW, I view this "basalness" in Kostenki, not as admixture, but a reflexion of his place in time in relation to the split of East and West Eurasians. It is not Basal Eurasian, or anything like that.

Krefter said...

@Chad,
"Even Bichon, at 11kya is barely in the clade of WHG."

In D-stats Loschbour and K01 aren't closer to each other than to Bichon. Why do you say he is barely WHG?

I agree WHG was not around with U5a formed. However something very close to WHG was. U5a is a good explanation for the WHG-like stuff in EHG.

Nirjhar007 said...

Its a shame that Eastern Europe was not sampled and of course no Autosomal components. But still a great paper ! .
Personally i'm not a fan of connecting SNP mutations with components..

capra internetensis said...

@Romulus

Stora Forvar had pre-I1 and U5a. BAB5 had pre-I1 and H. Nordic Bronze Age guy had I1 and J1c. 3 Motala men had I2a and U5a. Samara_HG had R1b and U5a. And then a crapload of Copper Age people had U5a and R1a or R1b of course.

I'm not seeing an I1 + U5 pattern here.

terryt said...

"The present-day data from Southeast Asia is irrelevant to the origins of West Eurasian R".

I disagree completely. You cannot understand any one region without considering the whole picture. I realise most here are only interested in Europe but Australia was settled around the same time as, if not before, Europe. Human migration has always been influenced by geography and in SE Asia we have one of the main geographic boundaries on earth. It even separates placental mammals and marsupials. Yet modern humans evidently crossed it fairly soon after leaving Africa.

"You can't look that far back in time with modern DNA with any certainty".

But many haplotypes show a marked division at Wallace's line which, I believe, tells us a great deal about the far distant time. Y-DNA C1b2 is the only C to cross, probably the first across. C1b1 is found in Borneo and South Asia while C1b2 is found across Wallace's Line in Timor and neighbouring islands. To me that indicates a fundamental division within C. Y-DNA K's movement across the line is more complicated with several K2b1 clades across the line and others not so clear-cut. Like C1b1a2, K2b1b is found in Borneo, west of the line, and K2b1c in the Philippines, perhaps across the line, or not, depending on where you draw the line. Similarly mt-DNA partitions into east and west of the line. N13, N14, S and O to the east, along with most of the Ps. And M14, M15, M29'Q, M27 and M28 also east of the line. They have to be ancient divisions in the respective haplotype lines. Nothing comparable has so far turned up in Europe. Virtually only downstream clades.

I believe we need to widen our perspective if we are to make any realistic conclusions about ancient human movement.

Chad Rohlfsen said...

I will go over various stats tomorrow evening.

mooreisbetter said...


Just a couple months ago, in the context of the peopling of Ireland, I emphasized how important it is to put all conclusions based on modern uniparental distributions through a model based on sound logic and population demographics.

Specifically, I wrote that ancient population sizes were minuscule compared to modern ones, and that if a population started a long long time ago, with a size that was way way small -- compared to subsequent waves -- that it would give a false signal that the original population was "conquered" or "outcompeted" or "never existed" or "originated elsewhere." I cautioned against those four errors.

I explained how ancient demographic models work here:

http://snplogic.blogspot.com/2015/12/the-spread-of-haplogroups-in-europe.html

Many "Interwebz Scientistz" failed to grasp this.

It is WONDERFUL to see another peer-reviewed, scholarly paper making this exact same point, and backing it up with newfound data.

Had I posted that mtDNA Hg M existed in Europe, I would have been ridiculed. But it's true. As the paper indicates:

-They started with a TINY initial population size.

-There is a loss every generation of males having males or females having female offspring.

-I've calculated the approximate odds of a male not having a male child or a female not having a female child (i.e. looking like their uniparental marker was "conquered") at 12.5%, each generation, totally random.

-The longer a population has existed in a locale (and being free of mutations), the more generations go by, the greater the chance that random happenstance, chance, etc. will make it appear that a Hg either never existed or was slaughtered in a mass killing/enslavement/mate preference.

Now you have further proof of it.

I'm waiting to hear how Hg N died out because of some studly new more beautiful females who moved in. Oh woops, that would only appear here if women could be R1b.

Davidski said...

Are you, perchance, related to Marnie?

Tesmos said...

terryt, the origins of K2 is based on modern diversity which is misleading. we should wait for ancient dna from south East Asia to be certain.

Alberto said...

@Chad

The idea that MA1/EHG and WHG evolved completely separately, and that after 15,000 years of isolation from each other and each one with its own drift they suddenly became more related than they were before doesn't seem to make a lot of sense. That's like one in a million chances to happen.

We have to accept that at some point ANE and WHG mixed to form EHG. We might not have the right samples still to make it work, but I doubt there's any other possibility that is realistic.

Kostenki losing its "basalness" (it's really more basal than the older Ust-Ishim) by drift is not impossible, but again an unlikely bet. I think that (pre-proto-?)WHG had to enter Europe without any Basal Eurasian admixture (so Kostenki14 might be a dead line).

What is not so clear is if WHG developed after entering Europe or before. If it did so after, then at some point it had to migrate from Europe to the Near East, to mix with Basal Eurasian populations there to form ancient Near Easterners. Or else, WHG evolved in West Asia and moved to Europe, and then some Basal Eurasian population moved from the south and mixed in the Near East with the WHG that stayed there. Similar possibilities regarding ANE and CHG (either ANE moved from North to South at some point, of ANE evolved in the south before moving north).

Matt said...

Thoughts:

- I'd love to see the "Pinhasi group" comments on this one, since IRC they're the ones looking at the nuclear sequence dna from around this era - http://horizon-magazine.eu/article/ice-age-europeans-roamed-small-bands-fewer-30-brink-extinction_en.html. Running D-stats to see if pre-Late Glacial HG actually were more / less related to WHG / MA-1 / EHG / Kostenki / Anatolian Neolithic than Late Glacial and post.

Although that article isn't clear whether they have anything older than Bichon (14ka), who looks very WHG, but is only just on the other side of the transition between Late Glacial HG and pre-Late Glacial HG that they talk about.

The article linked above talks about very small group sizes, so perhaps during the LGM you have very small groups, with some ill health problems from that maybe, very specialised into glacial conditions, then vulnerable to improvement in climate (exposing their small groups to competition from other groups with larger long term population size, less specialised into glacial conditions, who could've expanded disproportionately).

(Excerpts - "In some cases, small bands of potentially as few as 20 to 30 people could have been moving over very large areas, over the whole of Europe as a single territory, according to Professor Ron Pinhasi, principal investigator on the EU-funded ADNABIOARC project....This demographic model is based on new evidence that suggests populations were much smaller than is generally thought to be a stable size for healthy reproduction, usually around 500 people. Such small groupings may have led to reduced fitness and even extinctions.

‘You see a real reduction in population numbers and diversity, so you see the few lineages that probably split or separated before the ice age, and then stayed isolated during the ice age,’ he said. ‘Some time after the ice age, they kind of re-emerge, or disperse, and get together, as we see new contributions to European lineages from Asia and in particular the Near East.’
"

- Not totally 100% sure why Iberia is out as a refugium source of the "U5 Late Glacial-> population", as they don't have much / any mtdna from there which would exclude it directly. Is this along the lines of "This is where the previous wave into Europe came from, so it can't be where a later, different wave would've come from?" or something else?

Rob said...

@ Alberto

Makes sense to me :)
If i were to choose, I'd bet the ANE-whG clade formed in northern West Asia, but was replaced after the LGM. But this is on no solid data, admittedly

Rob said...

@ Matt

One has to entertain the notion that the later, U5 guys also exposed from SW Europe (btw the refuge was in Franco-Cantabria- mostly in southern France , not Iberia).

So it's all up in the air. But maybe the authors already know something we don't ?

rozenfag said...

@ Matt

If I understand correctly, you are wondering whether Pinhasi's group has any samples older than 14kya? If so, they do have, look here:

https://sites.google.com/site/pinhasierc/home/samples

UP I guess means Upper Paleolithic. AFAIK Sunghir(Central Russia) is dated as 30kya.

Matt said...

@ Rozenfag If I understand correctly, you are wondering whether Pinhasi's group has any samples older than 14kya? If so, they do have, look here

More or less. More precisely I'm really wondering if the statements in the article I linked above were based on samples older than 14ka, so whether they had any sampled at that time and whether that was the basis of what. Also whether any of those older than 14ka were from Western-Central-non-Russia Europe. If they have sequenced those older samples, then their take on this would be particularly interesting (not that we'll probably see it til they're ready to publish).

Davidski said...

Matt, can you think of any D-stats that can be plotted to show whether modern and ancient samples acquired their CHG from the steppe and/or independently of the steppe?

Davidski said...

Here's a little something. f3 not D-stats though.

https://drive.google.com/file/d/0B9o3EYTdM8lQdUNfckItWFhGZXM/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQNGU5TGtLS1Y1VjA/view?usp=sharing

Chris Davies said...

Y DNA haplogroup C and various clades of mtDNA haplogroup U have been sampled in North Africans. Presumably a result of southward gene-flow from the Franco-Cantabrian refuge after LGM. La Brana appears to show North African affinities, so the flow was no doubt bi-directional.

Krefter said...

There's nothing special about La Brana-1 having Y DNA C1a2. It was a rare lineage among WHGs and EEFs. No big deal. C1a2 doesn't change La Brana-1's autosomal makeup. There's no reason to make connections to SE Asia or North Africa because of his C1a2. Amoung WHGs everyone is most distant from La Brana-1, I think there's something screwed about his genome and not something exotic about his ancestry.

Matt said...

Unfortunately not really any ideas; obviously it seems like you would need some method to isolate exactly the CHG related ancestry in Yamnaya, then run a D-stat on that. I suppose you could compare the correlation of D(Outgroup,Pop)(Yamnaya,EHG) to the f3(Pop,Kotias:Outgroup) and see what that does? But that would maybe only give a signal of what populations are like relative to one another. That's something you'll have thought of and tried though.

...

Btw, with this study, wonder why they didn't include mtdna R3 KO1, as an apparent autosomally Mesolithic WHG that's contemporary with La Brana? Possibly because of it being found in a Neolithic context.

Chad Rohlfsen said...

Kostenki is not more Basal than Ust-Ishim. In fact, unlike other WHG samples, KO1 is the HG that is pretty significantly further from Ust-Ishim than the others. I think that farmer ancestry in KO1 is a possibility. I've stated that the EHG being closer to WHG is likely to be some EHG into WHG, but both also share drift from East Eurasians to the exclusion of Kostenki and MA1. There's a few things going on and I will post stats to explain my views later on this evening.

Romulus said...

@Chad

EHG into WHG is a possibility, but it doesn't explain U5a in EHG. I can't see any possible explanation for U5a in EHG other than UP European DNA in EHG.

Maybe Kotenski admixture is what is pushing WHG away from EHG and what EHG is lacking, just a thought.

Also interesting that the mtDNA M from the study clusters close to the mtDNA M found in Australian Aborigines according to their phylogenetic tree, another pop high in Y HG C.

epoch2013 said...

@Chad

"Kostenki is not more Basal than Ust-Ishim."

It has more basal Eurasian than UI. The D-stats pointing to that are in the Oase 1 paper.

http://www.nature.com/nature/journal/v524/n7564/images/nature14558-st1.jpg

epoch2013 said...

Shouldn't we revisit Oase 1 and try to weasel out anything that can be weaseled out? I mean, it's the oldest autosomical DNA we have from Europe and it is very strange. Not UI at all, for one thing.

bellbeakerblogger said...

Allow me to pour blood in the river.

I think this calls into question this paper from de la Rua (2015)
http://www.sciencedirect.com/science/article/pii/S1040618215000580#
with the Magdalenian Haplogroup H from El Miron that already looked questionable.

As is everything from Chandler 2005, in which we have Haplogroup H from "Mesolithic shell-middens" which were in reality stratigraphically mixed with Maritime Impresso when it was shoveled out of the ground at some point in the last hundred years, before being put in a box mixed with Medieval human remains or before being chemically treated, but was thankfully recovered after the museum burned down.

Other than La Brana I and II, almost every dna study from Spain and Italy is not useable. The weight of studies from around Europe should cause everyone to seriously reconsider the demographic Neolithization of SW Europe.

epoch2013 said...

@bellbeakerblogger

"I think this calls into question this paper from de la Rua (2015)
http://www.sciencedirect.com/science/article/pii/S1040618215000580#
with the Magdalenian Haplogroup H from El Miron that already looked questionable."

Why did that look questionable?

Jean said...

>>As is everything from Chandler 2005, in which we have Haplogroup H from "Mesolithic shell-middens" which were in reality stratigraphically mixed with Maritime Impresso when it was shoveled out of the ground at some point in the last hundred years, before being put in a box mixed with Medieval human remains or before being chemically treated, but was thankfully recovered after the museum burned down.<<

:) Love it! Is it true? Where did you hear the story?

Hector said...

It's a lot of fun reading the comments here. I will be brief since it is likely to be deleted by davidsky and I do not want to hide my identity.

Oase's Y is very likely to belong to the same branch as Ust Ishim's, opposite to NO side but within K2a.

KO14's C belongs to a branch that includes South Asians and Southern Chinese, SE Asians. It is not as downstream as to justify to call it "Indian".

FrankN said...

@Romulus: "IMO here is where I1 and mtdna U5a come from:

https://en.wikipedia.org/wiki/Pavlovian_culture"


Pavlovian has been tested in the study: 4 U5 (plain), 1 U8.

In general, the conclusion on "population replacement" is based on pretty shaky grounds. The main evidence is 6 out of their 8 (75%) "post-LGM" samples being U8a, which tends to get rare afterwards (but is still found in today's Basques). A closer look shows that oout of these six U8 samples, five come from the same region, namely three nearby caves in the Suebian Jura. Moreover, two specimens, a right and a left femur, are either belonging to the same or two maternally related individuals. A third one from the same cave may possibly maternally related as well.

Furthemore, the delineation of "post-LGM" is questionable. Commonly, as they also spell out in the Annex, the beginning of the Late Glacial is set at the beginning of the Dryas, 13.7 ky. With that temporal cut-off, two other samples, Bichon and Oberkassel998, both U5b, should also be regarded as "post-LGM".
If we leave aside the Suebian Jura caves, but consider the a/m pre-Dryas samples as "post-LGM", we are left with three U5b and one U2/3/4/7/8/9 for that period. On that data, the conclusion would be a reduction of genetic diversity during the LGM, as they are stating correctly, with M being a point in case.

Archeology indicates depopulation of most of Central Europe during the Late LGM. Aside from climate deterioration during the Younger Dryas, the Laacher See volcanic eruption (12.9 ky) surely played its role. Assuming that U8(a) settled primarily in Central Europe (all the study's finds are from the Czech Republic, Suebia and Goyet/BE), its (relative) disappearance may well have had a natural cause.

https://en.wikipedia.org/wiki/Laacher_See

Romulus said...

Whether or not there are stray examples of H in mesolithic Europe I don't think really matters. We can see that West European mtDNA haplogroups were established in the Beaker Era, with partial contribution from the Neolithic. Obviously partial because Neolithics were high in K and other haplogroups uncommon today. Some small group of Sredny Stog men swapped wives with Funnelbeakers and then founded Western Europe, picking up additional Neolithic Autosomal and mtDNA in the process. The word swapped is important in that sentence otherwise we would see just as much Steppe as Neolithic mtDNA in West Europe, as opposed to Scandinavia and East Europe where both are common.

Grimaldi man was a Sub Saharan African, are we assuming all Black men in Europe are descended from him?

bellbeakerblogger said...

@epoch

Well it's clearly an outlier and the history of mtdna studies in Spain and Italy don't exactly have a good record to be kind. If it was a full nuclear genome or something different enough to be plausibly ancient, then we'd have some confidence. But instead we are only a few hundred miles from much later La Brana and it happens to be a haplogroup, that in theory (admittedly this is backward reasoning) should be mostly localized in upper Mespotamia if it had even brached into is most basal version at that point (which it probably had).

@Jean

Actually, there was a group of remains from several site on the Sado, Muge and Tagus. These sites are hugely problematic in the first place, one being the gap between the 'so-called' Mesolithic and the Late Early Neolithic Impresso immigrants. Racially, some of these Mesolithic folks are extremely short, possibly some of the shortest people in Europe. That points to something we already know of the Portuguese Neolithic and the Mediterranean farmers in general, is that some of them had selection for shortness, especially in Portugual. (Here on Eurogenes going back a few months there was a paper on Portuguese 'short genes'.) That's very different from European Natives that were, on average, quite tall. Of course, Impresso pottery in Mesolithic layers doesn't help.

Assuming actual Mesolithic persons were taken, you have to remember where the sequence is in comparison to the rest of coastal Europe. Most were heavily calcified and highly fragmented bones (which is how they sorted out those that had been mixed from the Sado collection)
Overall, these bones are from the delta areas of a fairly mild to warm place and other than taking their word for it, I have deep reservations that their results reflect the genetic reality of the pre-Neolithic.

Romulus said...

@FrankN

I'd suspect the extinction of the Mammoth also played a role, also 12kya. Pavlovians were Mammoth hunters. The Transition from Mammoth hunting to fishing was probably not without impact on population size.

epoch2013 said...

@bellbeakerblogger

So you don't have anything suspicious on the study apart from its outcome?

bellbeakerblogger said...

@epoch

That's right, but let me say this. The burden isn't for me to bury my preconceived notions and suspicion only because they presented an apparent new fact.

It's their job to present facts that can stand to scrutiny. But that's the problem isn't it, I can only take on faith that they have a real result.

When you present a haplogroup that is not dissimilar from the archaeologists and geneticists that handled the remains, then the confidence level goes down.
When you present a haplogroup that is also an outlier to Paleolithic Europe, but also to much of Neolithic Iberia, but is rather modern, it goes down further.

So basically, at this point it is a contest between faith and preconceived notions. Take your pick.

Rob said...

mtDNA H was found in Mesolithic Karelia was it not ?
(der Sarkissian/ Haak; 2013)

epoch2013 said...

@FrankN

There seems to be something on the web of some survival of U8a among Basques and a Spaniard and a Finn: http://www.ianlogan.co.uk/discussion/gifs/U6_gif.htm

epoch2013 said...

@bellbeakerblogger

I'm fine with throwing the shell middens samples out. But for this study you need to do more than say you don't trust is. You can't consider it suspicious simply because it doesn't fit your idea.

bellbeakerblogger said...

@Rob

Karelia, yes. Full genome, different climate and 13kyr later. Very comfortable with that.

@epoch

Yes, but again what's my recourse? The verifiability of this result will depend on 'something else happening'. IMV, this is a full genome or many similar pre-Neolithic mtdna results from the Atlantic or North Africa. I don't think that threshold has been met yet, at least not for me.

FrankN said...

@epoch: On U8a see here as well:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1523212/

epoch2013 said...

@bellbeakerblogger

But you'd consider this papers conclusion on M valid even if it also doesn't present a full genome. I don't think you can dismiss the el Miron case that easily.

Also, do you have a link for the story on the Chandler 2005 samples?

epoch2013 said...

@FrankN

One sample from the 211. One can not call that abundant. So more or less something similar like Y-DNA C6.

Jean said...

> I have deep reservations that their results reflect the genetic reality of the pre-Neolithic.<

@ bellbeakerblogger

Oh me too! In fact I feel the results are best ignored. In my online table of Mesolithic DNA, I refer to the probability of contamination, but I also suspected a mix-up in dating of the type you describe. It is essential in multi-layer sites to get radiocarbon-dating of the remains themselves.

Kristiina said...

@Romulus “We can see that West European mtDNA haplogroups were established in the Beaker Era, with partial contribution from the Neolithic.”
I do not understand that statement. Europe was full of U5 and other U lines since the Upper Paleolithic, and they look like having spread from Europe to the east, and U lines are still relatively frequent in Europe. IMO, most European mtDNAs have deep roots in Europe.
U5
U5 Goyet2878-21 Belgium 26.7 kya, Dolni Vestonice Czech 31 kya, DolniVestonice16RRBP Czech 30 kya, Gurgy Les Noisats France Neolithic U5, RRBP Neolithic Gurgy Les Noisats France U5, Late Neolithic Perdigões Portugal U5
U5a and U5a1
U5a Monte Canelas Portugal Late Neolithic U5a1, Bom Santo Portugal Late Neolithic U5a1, LBKT Neolithic Budakeszi 4/8 Hungary U5a1, Lengyel Neolithic Brześć Kujawski Poland U5a, MN Bom Santo Portugal U5a1, MN Treilles Aveyron France U5, Late NE Monte Canelas Portugal U5a1a1, Greece Mykene LBA U5a1/U5a1a, Mesolithic Sweden Stora Förvar cave U5a1, Motala PWC HG x2 U5a1, Srubnaya Samara U5a1, Srubnaya Samara outlier U5a1, Ajvide PWC U5a1a1a1
U5a2
LesCloseaux3 France 9.9 kya U5a2, MareuilLesMeaux1 France 9.3 kya U5a2e, Motala PWC U5a2, Hohlenstein-Stadel Germany U5a2a, Donau Eneolithic Durankulak Bulgaria U5a2a, Bell Beaker Damsbo Denmark U5a2a, Erd 4 Vatya BA Hungary U5a2a, Montenegro Velika Gruda LBA U5a2, Srubnaya Samara U5a2a1
U5b and U5b1
Oberkassel 12 kya U5b1, Erralla Magdalenian Spain 12 kya U5b, Iboussieres31-2 France 11.8 kya U5b1, Ranchot88 France 10 kya U5b1, BerryAuBac1 France 7.2 kya U5b1a, Megalithic MN La Mina Spain U5b1, MN Bom Santo Portugal U5b, Kunda Lithuania U5b, Narva Lithuania U5b, Zedmar Poland U5b, Bell Beaker Quedlinburg Germany U5b, Corded Ware Karsdorf Germany U5b
U5b1a
Loschbour U5b1a, MN Treilles Aveyron France U5b1a, Corded Ware Eulau Germany U5b1a’b, Bell Beaker Knezeves Czech U5b, Bell Beaker Karsdorf Germany U5b1a1,
U5b1b
CuiryLesChaudardes1 France 8.2 kya U5b1b, RRBP NE Gurgy Les Noisats France U5b1b*, RRBP NE Gurgy Les Noisats France U5b1b’c*, LNBA El Portalón Sierra de Atapuerca Spain U5b1b, Visby PWC U5b1, Janisławice U5b1b,
U5b1, U5b1i and U5b1c
Iberia MN La Mina pre-U5b1i, Regional TRB Bernburg Benzingerode U5b1c1, MN Treilles Aveyron France U5b1c, Corded Ware Tiefbrunn Germany U5b1c2, Corded Ware Bergrheinfeld LNBA Germany U5b1c2, Minoan BA Ayios Charalambos cave Crete U5b1*
U5b2a
Schöningen NE Salzmünde Germany U5b2a2c, Baalberge MN Quedlinburg VII Germany U5b2a2, Farmer MN Blätterhöhle Germany U5b2a2 x2, Fisher-gatherer MN Blätterhöhle Germany U5b2a2, Farmer MN Blätterhöhle Germany U5b2a5
U5b2b
Pagliacci71 Italy 18,6 kya U5b2b, Rochedane rance 13 kya U5b2b, Iboussieres39 France 12 kya U5b2b, Italy Late Epi-Gravettian 11 kya U5b2b1, Epicardial NE Avellaner cave Cataluña U5b2b2, NE Gurgy Les Noisats France U5b2b1a, Fisher-gatherer MN Blätterhöhle Germany U5b2b(2)*, Fisher-gatherer Blätterhöhle MN Germany U5b2b2*
U and U2’3’4’7’8’9
Cioclovina1 Romania U 33 kya, Paglicci 108 Italy 28 kya U2’3’4’7’8’9, Rigney1 France 15.5 kya U2’3’4’7’8’9, Spain Solutrian Nerja Málaga U, Germany Magdalenian Hohler Fels U, Paglicci Cave Italy 11 kya U2'3'4'7'8'9, LBK NE Vedrovice Czech U*, LBKT NE Budakeszi 4/8 Hungary U2b, NE Los Cascajos Navarra Spain U x3, RRBP NE Gurgy Les Noisats France U, MN Fuente Hoz Alava Spain U (CRS) x3, MN Treilles Aveyron France (CRS), Late NE Paimogo Portugal U (CRS), LNBA Pico Ramos Bizkaia Spain U x3, Manych Catacomb Temrta V Russia U (CRS)
U2
GoyetQ53-1 Belgium 28 kya, GoyetQ376-19 Belgium 27.5 kya U2, GoyetQ55-2 Belgium 27.5 kya U2, GoyetQ56-16 Belgium 26.3 kya U2, Ust Tartas Sopka West Siberia U2e x2, Blätterhöhle 9000 kya U2e and Motala HG U2e1, Yuzhnyy Oleni Ostrov U2e, Donau Yamnaya Golyamata Mogila Bulgaria U2e1a; Kytmanovo Andronovo (outlier) Russia U2e1, Chociwel Unetice LNBA U2e1f1, Corded Ware Esperstedt LNBA Germany U2e1a1, Corded Ware Spreitenbach-Moosweg Switzerland U2e1, Potapovka Samara U2e1h, Sintashta Bulanovo U2e1e, Sintashta Bolshegarakanski U2e1h

Kristiina said...

U4
Starčevo NE Lánycsók Hungary U4, NE Los Cascajos Navarra Spain U4, RRBP NE Gurgy Les Noisats France U4, MN Dólmen do Ansião Portugal U4, Late NE Paimogo Portugal U4, Pitted Ware Ajvidex3 and Ire8 Sweden U4d, Karelia Popovo Mesolithic U4d, Ust Tartas Sopka West Siberia U4d1b, Corded Ware Tiefbrunn Germany U4, Bell Beaker Damsbo Denmark U4d1b, Manych Catacomb Temrta III and Temrta V Russia U4, Donez Catacomb Ukraina U4, Donez Catacomb Novozvanovka II Ukraina U4, Manych Catacomb Peschanyi Russia U4, Kytmanovo Andronovo U4d1, BA Sumer Mari Syria U4
U8
Dolni Vestonice Czech U8 31 kya, HohleFels79 Germany 16 kya U8a, HohleFels49 Germany 15,5 kya U8a, HohleFels10 Germany 15,5 kya U8a, GoyetQ-2 Belgium 15 kya U8a, Brillenhohle Germany 14,8 kya U8a, Burkhardtshohle Germany 14,6 kya U8a, Baalberge MN Quedlinburg VII Germany U8a1a; Maikop and Novosvobodnaya Klady U8, Barcin Neolithic Turkey U8b1b1, Sopot proto-Lengyel Alsónyék Hungary U8b1b, Schöningen NE Salzmünde Germany U8b1b
K
Starčevo Alsónyék Hungary K* x3, Starčevo NE Vinkovci Jugobanka Croatia K*, Starčevo Lánycsók Hungary K*, LBKT NE Tolna-Mözs Hungary K*, Cardial NE Can Sadurnı Spain K*, LBK NE Vedrovice Czech K* x3, RRBP NE Gurgy Les Noisats France K*,LBKT NE Bölcske-Gyűrűsvölgy Hungary K, LBK NE Flomborn Germany K* x2, LBK NE Derenburg Meerenstieg K*, RRBP NE Gurgy Les Noisats France K x7, Rossen NE Oberweiderstedt K*, Schöningen NE Salzmünde Germany K, Late NE Araba Spain K x14, MN Forager FB Ostorf Germany K, Globular Amphora Kowal Poland K, Megalith Dolmen of La Pierre Fritte France K, El Mirador Chalcolitic Spain K, LNBA Pico Ramos Bizkaia Spain K x4, Corded Ware Spreitenbach-Moosweg Switzerland K, Minoan BA Ayios Charalambos cave Crete K x2
K1a
Schöningen NE Salzmünde Germany K1a, Baalberge MN Quedlinburg VII Germany K1a x2, Regional TRB Salzmünde Germany K1, Regional TRB Salzmünde Germany K1a, Regional TRB Bernburg Benzingerode K1a, Regional TRB Bernburg Benzingerode K1, MN Treilles Aveyron Framce K1a x2, LNBA Tres Montes Navarra K (224C), Bell Beaker Kromsdorf Germany K1a
K1a1 and K1a1b
Megalithic MN La Mina Spain K1a1b1, Corded Ware Esperstedt LNBA Germany K1a1b2a, CZU Globular Amphora culture K1a1
K1a2
Menteşe Neolithic Turkey K1a2, LBK NE Halberstadt Germany K1a2, Cardial NE Cova Bonica Spain K1a2, Els Trocs NE Spain K1a2a, MN Bom Santo Portugal K1a2a1, El Mirador Chalcolitic Spain K1a2a, NE El Portalón Spain ATP3 K1a2b, Bell Beaker Rothenschirmbac Germany K1a2c, Szöreg Maros BA Hungary K1a2a, Battle Axe Viby Sweden K1a2a
K1a4
Barcin Neolithic Turkey K1a4, Menteşe Neolithic Turkey K1a4, Cardial NE Cova de la Sarsa Spain K1a4a1, RRBP NE Gurgy Les Noisats France K1a3/ K1a4a x2, RRBP NE Gurgy Les Noisats France K1a4c, Schöningen NE Salzmünde Germany K1a3/ K1a4 x5, Regional TRB Salzmünde Germany K1a4a1a2, Szczepankowice Unetice LNBA Polamd K1a4a1
K1b
Barcin Neolithic Turkey K1b1b1, El Mirador Chalcolitic Spain K1b1a1, Els Trocs 10407 K1b1a1, Megalithic MN La Mina Spain K1b1a1, Corded Ware Oblaczkowo Puola K1b1a1, Corded Ware Eulau Germany K1b1a x2, Unetice Moravska Nova Ves Czech K1b1a, Bronze Age Denmark K1b1a1, Battle Axe Sweden K1b1a1, Srubnaya Samara K1b2a
HV0
Starčevo NE Vinkovci Nama Croatia HV0, Rossen NE Oberweiderstedt Germany HV0, Lengyel NE Brześć Kujawski Poland HV0, Rossen NE Wittmar Germany HV0, MN Bom Santo Portugal HV0, Late NE Calden Germany HV0, MN Odagsen Germany HV0, MN Treilles Aveyron Germany HV0 x2, Corded Ware Oberwiederstedt 2 Germany HV0e
V
Starčevo Alsónyék Hungary V, Starčevo NE Vinkovci Jugobanka Croatia V, Els Trocs NE Spain V, NE Gurgy Les Noisats France V x2, Rossen NE Halberstadt Germany V, Regional TRB Salzmünde Germany V, Regional TRB Bernburg Benzingerode V, MN Treilles Aveyron France, BA Malpaso cave Castellnovo Spain V, Pitted Ware Ajvide 52 Sweden V

I stop here as this gets really tiresome, but I add that, in addition to the above listed old European U haplotypes and HV/V haplotypes, many Neolithic H haplotypes as well as J and T haplotypes are deeply rooted in Neolithic Europe and still vigorous.

bellbeakerblogger said...

@epoch

Well yes, at least in the case of a WHG like La Brana, there's no similar modern population, so I'd have 100% confidence in his HG designations based on the full genome but also because the HGs themselves are less frequent or rare in Iberia.

As far as the remains from Chandler, there isn't a paper that is critical other than a disorganized rant on my blog, but the many papers that are available on the shellmiddens don't give me the confidence that haplogroup 'whatever' found in a bone fragment and reflects a profile of culture x. These layers seem very complex. I just don't have a lot of confidence in these.

Rob said...

I don't think the reservations aired by BBB and Jean are unreasonable. For science's sake, we can demand greater clarity of context of said samples. If mtDNA H is indeed to be found in pre-Neolithic Iberia, then we can be rest assured that it will be.

Rob said...

Krstiina

Good effort. Did you forget K1c from Mesolithic Greece (Theopetra) ?

Nomen Cognomen said...

It seems that the WHG Y-Dna I-carriers are different than the original Europeans. Which makes sense because I is related to J, and we know Europe was inhabited prior to IJ split. I probably came from South-East Europe/Middle East.

On the terminology,

I propose we call the earliest European HG, OHG, Original Hunter Gatherers.

Krefter said...

@Nomen,

There's no confirmation yet that Europe was inhabated solely by a dead-line(OHG). We can see plenty of mtDNA U(inclu. Euro-specific proto-U5, proto-U2e) in Europe 30,000-40,000 years ago along with mtDNA M*/N*/R* and OHGs. It's possible proto-WHG/ANEs and OHGs lived side by side.

This paper is not saying that before the Last Ice age Europe was inhabted by dead-ends(OHGs). We know in fact they were dominated by mtDNA U and we have one 38,000 year old genome; Kostinki, who is very close to WHG. They're saying that U5-dominated groups from Europe(or West Asia) replaced their close relatives who also had lots of mtDNA U and U5.

So, in conclusion OHG doesn't work as a term yet. Of course at some-point lots of Europe was inhabated by dead-ends(OHGs) as most of Eurasia probably was. This isn't very important, because very quickly most Europeans were apart of U-dominated WHG/ANE-like groups.

Nomen Cognomen said...

Yes you are right about the ubiquitousness of U* mtdna in OHGs. But they also had extinct mtDnas like M, meaning they are not the uniparental ancestors of modern day Europeans (unless there was an internal bottleneck but that's dubious considering the age of I).

The difference is that mtDna U is 55,000 years old, while y-Dna I is 25-30,000 years old. The OHGs were probably the y-DNA F* and C* people uniparentally linked to South-East Asian countries, with M, K, U2*, U* MtDna.

While the "modern" I2, U4/5 bearing HGs probably expanded 15-25 thousand years ago into Europe.

Rob said...

Nomen

It don't think you know how "old" hg I is, do you ? Eg YFull has it at 42 000 YBP. So it might be worth waiting until we get more Palaeolithic European Y DNA before we start the silly neologisms

Krefter said...

@Nomen,

Mal'ta boy and a 33,000 year old Romanian are the only U*s. ~30,000 mtDNA from Czech and Belgium is mostly proto-U5 and proto-U2e, both typical of Mesolithic Europeans. And we have Kostinki's genome(had U2*) and he has a very close relationship with WHG.

WHG bearing U5b and I2 replaced(integrated?) very close relatives who therefore don't fit the OHG profile.

Rob said...

'I'm da O'G OHG'

- Master P

G Horvat said...

Most of the Goyet U sequences had mutations in common with U2e but were missing several; hence "pre-U2e". One of those expected but not present mutations was 6045T which was, oddly, present in the Goyet "M" sequences instead. 6045T is also pretty rare.

Chad Rohlfsen said...

There is no Basal Eurasian in Kostenki. Not even close. Even WHG is much more significantly related to Anatolians than Kostenki is. That wouldn't be the case if he had Basal Eurasian. Kostenki is close to Proto-West Eurasian, but slightly towards WHG, which will show as a progression from 36kya to 14kya. Just wait for the aDNA. There is no phantom Loschbour pop coming out of the mist.

Ust_Ishim Loschbour Kostenki MA1 -0.0099 -1.089 16434 16762 344074
Yoruba Loschbour Kostenki MA1 -0.0097 -1.331 16947 17279 344753
Yoruba Ust_Ishim Iberia_HG Kostenki 0.0024 0.371 22951 22842 475892
Yoruba Ust_Ishim Iberia_HG MA1 0.0046 0.737 17860 17696 368797
Yoruba Ust_Ishim Iberia_HG Karelia_HG -0.0125 -1.973 22020 22578 484317
Yoruba Ust_Ishim Iberia_HG Motala_HG -0.0080 -1.588 21638 21985 494328
Yoruba Ust_Ishim Iberia_HG Loschbour 0.0021 0.338 17611 17538 434693
Yoruba Ust_Ishim Iberia_HG Hungary_HG -0.0195 -2.944 15309 15919 367582
Yoruba Ust_Ishim Iberia_HG Karitiana 0.0039 0.714 25738 25536 513632
Yoruba Ust_Ishim Iberia_HG Han 0.0033 0.643 26247 26071 513632
Yoruba Ust_Ishim Iberia_HG Onge 0.0057 1.063 26323 26026 513632
Yoruba Ust_Ishim Kostenki MA1 -0.0006 -0.091 19243 19265 395639
Yoruba Ust_Ishim Kostenki Karelia_HG -0.0166 -2.639 24695 25529 514166
Yoruba Ust_Ishim Kostenki Motala_HG -0.0100 -1.903 25268 25778 525140
Yoruba Ust_Ishim Kostenki Loschbour -0.0038 -0.574 22002 22169 465954
Yoruba Ust_Ishim Kostenki Hungary_HG -0.0191 -2.932 19097 19841 404262
Yoruba Ust_Ishim Kostenki Karitiana 0.0010 0.178 27762 27706 549474
Yoruba Ust_Ishim Kostenki Han -0.0002 -0.035 28288 28298 549474
Yoruba Ust_Ishim Kostenki Onge 0.0015 0.261 28238 28156 549474
Anatolia_Neolithic Ust_Ishim Iberia_HG Kostenki 0.0471 7.215 23924 21770 475525
Anatolia_Neolithic Ust_Ishim Iberia_HG MA1 0.0513 7.884 18609 16793 368525
Anatolia_Neolithic Ust_Ishim Iberia_HG Karelia_HG 0.0043 0.650 22451 22260 484158
Anatolia_Neolithic Ust_Ishim Iberia_HG Motala_HG -0.0103 -2.047 21721 22171 494162
Anatolia_Neolithic Ust_Ishim Iberia_HG Loschbour -0.0002 -0.029 17703 17709 434388
Anatolia_Neolithic Ust_Ishim Iberia_HG Hungary_HG -0.0328 -4.967 15355 16397 367533
Anatolia_Neolithic Ust_Ishim Iberia_HG Karitiana 0.0733 13.184 27358 23622 513188
Anatolia_Neolithic Ust_Ishim Iberia_HG Han 0.0874 16.422 28197 23665 513188
Anatolia_Neolithic Ust_Ishim Iberia_HG Onge 0.0956 17.511 28410 23450 513188
Anatolia_Neolithic Ust_Ishim Kostenki MA1 0.0002 0.030 19077 19070 395035
Anatolia_Neolithic Ust_Ishim Kostenki Karelia_HG -0.0448 -6.909 23851 26086 513844
Anatolia_Neolithic Ust_Ishim Kostenki Motala_HG -0.0556 -10.252 24133 26973 524839
Anatolia_Neolithic Ust_Ishim Kostenki Loschbour -0.0502 -7.747 20978 23197 465266
Anatolia_Neolithic Ust_Ishim Kostenki Hungary_HG -0.0754 -11.795 18087 21038 404174
Anatolia_Neolithic Ust_Ishim Kostenki Karitiana 0.0269 4.750 28053 26583 548488
Anatolia_Neolithic Ust_Ishim Kostenki Han 0.0415 7.852 28932 26625 548488
Anatolia_Neolithic Ust_Ishim Kostenki Onge 0.0491 8.680 29033 26314 548488

Ryan said...

@Krefter - "There's nothing special about La Brana-1 having Y DNA C1a2. It was a rare lineage among WHGs and EEFs. No big deal. C1a2 doesn't change La Brana-1's autosomal makeup. There's no reason to make connections to SE Asia or North Africa because of his C1a2. Amoung WHGs everyone is most distant from La Brana-1, I think there's something screwed about his genome and not something exotic about his ancestry."

My understanding is that La Brana's genome is of pretty decently high coverage actually, so I don't know what you mean by "screwed" about his genome.

There used to be Davidski's 4 pop Oracle for La Brana available here: http://eurogenes.blogspot.ca/2014/03/4-ancestors-oracle-results-for-anzick-1.html

Apparently that post has been taken down though? Not sure why.

Having C1a in both La Brana and Jomon samples means that at some point, there is a special connection between those two groups at some point, no if ands or buts. That connection may be so old and distant that it becomes lost in the autosomal DNA, but at some point it must have been there too.

So at what point did it become lost? And what was that connection? Either way though, those later samples lacking this strange affinity doesn't tell us much.

I'd note that La Brana, along with East Eurasian groups, did show greater similarity with Ust-Ishim than modern Europeans. Presumably because of some history of basal Eurasian and/or African admixture into Europe. So I don't think David's Oracle results are complete nonsense. They may just be trying to tell us something a bit different. (And I'd be interested in seeing similar D stats comparing Lochsbour, La Brana and Ust-Ishim).

I'm not necessarily saying the Onge travelled across Eurasia to Europe. This could just be a relic of an older migration. Or not.

@Davidski - My point wasn't about Y-DNA P so much as just how over time uniparental markers can become decoupled from autosomal DNA. I figured that was an uncontroversial example here, but apparently I was wrong lol. Like the Hausa - they have a lot of mammoth-steppe derived R1b, but not much of an autosomal connection to the steppe. Clearly at some point some of their male-line ancestors did.

I'm not sure we'll get the ancient DNA you hope for in the case of P though - it critical areas may have been too hot, humid, or under water to preserve the relevant DNA. Just to be clear though, when I say South and East Asia for Haplogroup P, I mean that in the broadest sense possible.

Davidski said...

Those admixture results were based on an old sequence of La Brana 1. We now have a high coverage version of this sample, and it doesn't appear to have any special affinity to Southeast Asia or, as is often mentioned in the comments here, North Africa.

Rob said...

Chad

You might well be correct.
There is no evidence for migration into Europe after 35 kya, and before the Neolithic. So I bet wherever the "proto-WHG" came/ evolved from, it was probably within Europe (sensu latu).

Maju said...

Does anyone know how to access the full sequences? I'm intrigued about the exact downstream classification of mtDNA M (my initial guess is that it should be M1 or closely related but need confirmation, what is clear is that it won't be "basal M" or anything remotely like that).

Krefter said...

@Maju,

Here you can see all the results from Posth 2016.

https://docs.google.com/spreadsheets/d/1FAuTvCIb4YuBED--zf0G5ITYUpqUgN_TYmqy2FqDfT0/edit#gid=0

The Ms don't share any mutations with modern M clades. They're their own clade.

Maju said...

Never mind, I found it. However, please someone check if I'm doing something wrong because the only genome I began checking (La Rochette) does not look like being M at all.

→ http://www.ncbi.nlm.nih.gov/nuccore/KU534951.1
→ http://www.phylotree.org/tree/subtree_M.htm

T489C yes
C10400T no (G)
T14783C no (A)
G15043A no (G)

???

Ryan said...

Gotcha David. Still, interesting that they both show up in Iberia, even if so distant in time.

German Dziebel said...

The data from the paper proves that Europe was NOT colonized from Africa. But the paper's title still claims that humans came out of Africa. Apparently, even data can't change "scientists" minds.

Maju said...

Worse: the genome does not even match with L3 (nor I can find alternative L sublineages that match even a single mutation):

Comparison with L3:

A769G no (C)
A1018G no (T)
C16311T no (A)

How sure are we that they are not Neanderthal remains?

Krefter said...

@Maju,

You should email an author about that.

Maju said...

@Krefter: maybe. By the moment I'm asking for your opinion, as I may be committing some sort of error, so please double-check.

I compared with the Vindija Neanderthal (http://www.ncbi.nlm.nih.gov/nuccore/NC_011137) and there is only another match among the considered markers:

M?
T489C yes / Vind-T
C10400T no G / Vind-T
T14783C no A / Vind-T
G15043A no G (ancestral human) / Vind-G

L3?
A769G no C / Vind-T
A1018G no T / Vind-T
C16311T no A / Vind-T (match)

So, yeah, please, second opinions anyone?

Maju said...

Oops sorry, the match is the 1018T, my bad. Rewriting that part:

L3?
A769G no C / Vind-T
A1018G no T / Vind-T (MATCH!!!)
C16311T no A / Vind-T

capra internetensis said...

@Maju

The mitochondrial DNA mutations they give are only the ones that differ from the basic sequence for the haplogroup the sample is assigned to (using the RSRS). So the M samples have all the defining M mutations, *plus* the ones listed in Krefter's spreadsheet. It's in the supplementary info. Nothing super-weird there, I'm afraid.

Rob said...

German

ы можете остановиться ?

Kristiina said...

Rob, yes you are right, and I may have omitted many more, but however for the sake of completeness:
Theopetra Thessaly Greece 7.55 BC K1c x2, Paternanbidea Navarra NE (73G, 16092, 16224, 16311) K*/K1c1 (?), Bell Beaker Osterhofen Germany K1c1, Mako BA Kompolt-Kigyoser Hungary K1c1, La Tène IA Glauberg Germany (73G 146C 152C 263G 315.1C 16224C 16311C) K1c1a (?)

@Krefter “They're saying that U5-dominated groups from Europe (or West Asia) replaced their close relatives who also had lots of mtDNA U and U5.”
I do not understand this West Asia thing. We have some ancient mtDNA from the Near East, i.e. from Tell Halula and Tell Ramad (Syria) and Tell Kurdu, Menteşe and Barcın (Turkey), and there is not a single U5, U2 or U4 there to date. There is U* from Tell Halula and U3x2 from Barcın and U8b1b1 from Barcın.

@Nomen Cognomen “But they also had extinct mtDnas like M, meaning they are not the uniparental ancestors of modern day Europeans (unless there was an internal bottleneck but that's dubious considering the age of I”
I think that there obviously was a bottleneck in the Late Glacial Maximum Europe. I think it is stupid to claim that men only mate with women having a certain haplogroup. I am sure many men were very interested in exotic women, as they still are.

@G Horvat
I noticed that one of Oleni Ostrov samples have the same HVR1 mutations as GoyetQ56-16 from Belgium, so we could presume that a female line managed to find its way from Western Europe to Mesolithic Karelia:
Yuzhnyy Oleni Ostrov U2e 7.5 kya U2e (129c, 189C!, 362C)
GoyetQ56-16 Belgium 26.3 kya U2/pre-U2e (129c, 189C!, 362C)

Rob said...

Thanks Kristiina
Are you aware of any K1c in modern pops ?

Maju said...

@Capra: I went to the original full sequence at GenBank (and provided links). Would you care to double check what to my eyes seems to be a total fiasco? I am still hoping that the fiasco is my mind, that I committed errors when contrasting the sequences but so far nobody has double-checked.

Kristiina said...

Rob, here you go:
K1c1
K1c1a Western Europe,
K1c1b Western Europe, including Sweden,
K1c1c Finland,
K1c1d Germany,
K1c1e Russia, Poland, including Volga Chuvash, Maris and Tatars,
K1c1f Slovakia, Germany;
K1c1g Western Europe,
K1c1 (16224C!) Iberia,
K1c2 Western Europe;

http://www.nature.com/ncomms/2013/131008/ncomms3543/full/ncomms3543.html
http://www.ianlogan.co.uk/sequences_by_group/k1c1a-f_genbank_sequences.htm

Alberto said...

@Chad

Thanks for those stats. So it does seem rather clear from them that Kostenki doesn't have any Basal Eurasian admixture. Was that some myth then? So that does place him as an ancestral West Eurasian.

What kind of stats make you think that WHG might have EHG admixture?

Kristiina said...

Alberto, the similarity between Yuzhnyy Oleni Ostrov U2e 7.5 kya U2e (129c, 189C!, 362C) and GoyetQ56-16 Belgium 26.3 kya U2/pre-U2e (129c, 189C!, 362C) is an indication of geneflow from WHG to EHG.

Rob said...

Kristiina

Thanks, mtDNA Queen

@ Alberto

Wasn't the "Basal" aspect of K14 was reported in the paper itself (Willerslev et al)?? but that was before many genomes were available

Karl_K said...

@Alberto, Chad

"There is no Basal Eurasian in Kostenki. Not even close. Even WHG is much more significantly related to Anatolians than Kostenki is. That wouldn't be the case if he had Basal Eurasian. Kostenki is close to Proto-West Eurasian, but slightly towards WHG"

The Kostenki paper didn't really say or show that it had the SAME "basal eurasian" as in EEF.

It said that the statistic D(Mbuti, East Asia; WHG, K14) is less than 0. So, there's some link between WHG and East Asians that is not shared by Kostenki.

There is no reason that this couldn't be a totally seperate "basal eurasian" branch. But it could also be explained in a lot of other ways.

Alberto said...

@Kristiina

Yes, for now I find that WHG admixture into EHG is the most parsimonious explanation, but Chad runs many stats to investigate these things, so I'm curious to see why he thinks the opposite.

@Rob

Yes, I went back to the paper to check the stats there. Probably the most relevant are the ones:

MbutiPygmy Han Loschbour Kostenki ‐0.039 ‐5.3 201566
MbutiPygmy Han Stuttgart Kostenki 0.000 0.0 208439

Which do suggest some Basal Eurasian admixture in Kostenki. Or else East Asian admixture into WHG/EEF (or WHG admixture into East Asians, including Han and Dai). The first option looks geographically/historically more likely. But contrasting the first stat above with this other one posted by Chad the second options would seem a better genetic explanation (?):

Yoruba Ust_Ishim Kostenki Loschbour -0.0038 -0.574 22002 22169 465954

Not sure what to think. Maybe Kostenki being more directly descended from Ust-Ishim than Loschbour is makes the stat insignificant (Basal Eusasian in Kostenki counterbalanced by stronger Ust-Ishim-Kostenki affinity due to closer in time/evolution?).

Alberto said...

@Karl, Chad

The paper did mention quite explicitly that Kostenki had the Basal Eurasian that is present in EEFs (I just read it again), and had this tree:

http://i.imgur.com/F7iRJRC.png

I wonder if now that we have Kotias (which does not show the strong WHG affinity that Anatolian Neolithic does could help here. Chad, could you check:

Kotias Ust_Ishim Kostenki Loschbour

So we can compare it to:

Anatolia_Neolithic Ust_Ishim Kostenki Loschbour -0.0502 -7.747 20978 23197 465266

Matt said...

@ Alberto, yeah, this is all what we discussed back when the Kostenki paper came out and Davidski was the first to be able to run comparison stats with Ust Ishim. I did think that with Kostenki the idea of Basal Eurasian maybe was more tenuous than established by Laziridis, for the reason that sharing between the ENA+WHG-ANE clades seemed specific to them, and not to map to any relatedness to the Ust-Ishim outgroup so well.

One thing about trying to resolve this by adding a specific link between Ust-Ishim and Kostenki though (i.e. Kostenki has Basal Eurasian but this is balanced out by Ust-Ishim ancestry) is that Oase shows exactly the same behaviour as Ust-Ishim (as shown by the stats Epoch linked above). I.e. D (Outgroup Oase Loschbour Kostenki), D (Outgroup Oase MA1 Kostenki) and D (Outgroup Oase Han Kostenki) all more or less non significant, while D (Outgroup Oase Stuttgart Kostenki) shows Kostenki more related to Oase.

Tobus said...

@Alberto:

Kotias Ust_Ishim Kostenki14_UP Loschbour -0.0331 -4.339 18437 19698 433479

Onur said...

@Maju

From the Supplemental Information section of the Posth et al. paper:

"Finally, three pre-LGM individuals (LaRochette, GoyetQ116-1 and GoyetQ363 3) carry mtDNA hg M defined by four positions from the L3 split (T489C, C10400T, T14783C, G15043A). Looking more closely at the intragroup M phylogeny, these three pre-LGM individuals are all placed on a branch not yet seen in modern-day individuals, either in Europe or elsewhere. In particular GoyetQ116-1 and GoyetQ363-3 from Belgium dated ~34 35 ka, show 6 (G207A, C1556T, C6045T, C8619a, A11084G, T16297) and 8 (G207A, C1556T, A6040G, C6041T, C6045T, C8619a, A11084G, T16297) derived mutations from the M root, respectively, whereas LaRochette from South France dated ~27.5 ka presents 7 (G207A, C1556T, C6045T, C6164T, C8619a, A11084G, C12816T) derived mutations from the M root. LaRochette carries a possible additional back mutation at position C16297T! (16Ts and 4Cs), compared to the two Belgian sequences both carrying the T16297C mutation. The four Cs at this position cannot be explained by damage and are likely the result of contamination or heteroplasmy, therefore an unassigned base (N) was placed at that position in the final consensus sequence (see above). Despite their geographic and temporal separation all three pre-LGM individuals share five mutations (G207A, C1556T, C6045T, C8619a, A11084) from the MRCA of the basal M lineage. This finding suggests that this M branch arrived in Europe before 35 ka and survived at least until 27.5 ka, before the LGM started (~25 ka)."

So it is clear that GoyetQ363-3, GoyetQ116-1 and LaRochette firmly belong to mtDNA haplogroup M but their M subclades are not close to any one of the existing M subclades. For me it is not strange to find mtDNA M in pre-LGM Europeans because the formation time of mtDNA M predates the split between West and East Eurasians and the LGM (Last Glacial Maximum) is sufficient to explain the disappearance of some of the pre-LGM mtDNA and Y-DNA lines in Europe.

epoch2013 said...

@Chad Rohlfsen

Maybe I'm missing the point, but I don't see how those stats prove K14 has no basal Eurasian. These are the stats that shows it likes the basal in current day Europeans:

D(Non-African1, Non-African2; Early Modern Human, African)

European, Stuttgart; K14, African -0.0002 -0.3

Meaning that K14 likes present day Europeans as much as Stuttgart.

European, Loschbour; K14, African -0.0052 -9.1

Meaning K14 prefers pure WHG over present day Europeans. What else than the presumed basal Eurasian admixture in EEF can cause the first D-stat?

Maybe part of is *considered* basal Eurasian in EEF is actually part K14 heritage in Anatolian farmers which simply was lost by drift in WHG. EEF consistently shows a part which is not shared by Middle-Easterners if I understand it all correctly.

Alberto said...

@Tobus

Thanks for the stat. So it seems that even removing the strong WHG preference by Anatolians/EEFs, Kostenki still doesn't show any affinity to Basal Eurasian.

@Matt

So what would you tentatively conclude then? Kostenki having no Basal Eurasian, and the stats with Han and Dai being because of post-Kostenki gene flow between WHG and East Asians? Or is it all an artifact of comparing genomes that are very old and/or distant in time from each other?

epoch2013 said...

@Alberto

But Kotias isn't PURE basal Eurasian if I recall correctly (Weren't there d-stats showing it shunned Anatolian EF?) and K14 showed a lot of affinity to a West-Asian part. This would skew that result as you are not sure what part of Kotias is related to what part of K14. It could very well be that K14 indeed does have BEU admixture but even has more Teal.

FrankN said...

@Kristina: "we could presume that a female line managed to find its way from Western Europe to Mesolithic Karelia"

Probably more than one, considering that Sami mtDNA is dominated by U5b. While details remain to be sorted out, there is general agreement about epi-paleolithic coastal colonalisation of Norway to well SE of the Northern Cape by the Ahrensburg-Bromme complex, which in itself is regarded as Magdalenian-derived.
https://en.wikipedia.org/wiki/Ahrensburg_culture

[The eponymous sites for the Ahrensburg and Hamburg cultures, both only a few hundred meters distant, are currently undergoing re-excavation in preparing for extension of the Hamburg-Ahrensburg suburban railway. Initial finds are described as "spectacular". The site seems to have been a semi-permanent settlement, and there is hope for uncovering some burials.
http://www.abendblatt.de/region/stormarn/article205796993/Archaeologen-machen-Sensationsfund-im-Tunneltal.html ]


@Alberto: "..or WHG admixture into East Asians, including Han and Dai"
For Han I am quite certain. Bronze technology was obviously spread into northern China by Sintashta/ Andronovo, plus we have Tocharians, Silk Road etc. Note also thie study on dentition:
http://www.academia.edu/3735477/Two-rooted_lower_canines-A_European_trait_and_sensitive_indicator_of_admixture_across_Eurasia
In European samples, two-rooted lower canines consistently exhibit frequencies of 5–8%. (..) In contrast, in Sub-Saharan Africans the trait is virtually unknown while in Asian and Asian-derived populations, it varies between 0.0 and 1.0%. Here we show that two-rooted canine frequencies for new migrants along the western frontiers of China and Mongolia ranged from 0–4%. These data suggest European-derived populations migrated into western China (Xinjiang Province) and Mongolia (Bayan Olgii Aimag)sometime during the late Bronze age (1000–400 BCE). (..) The highest regional frequency of two-rooted lower canines among the Asian samples is from Ordos China [Inner Mongolia] at 4.0%.".
Among "Dian" (non-Han speakers in Yunnan), the share was 1.1& (1/92). I had already elsewhere mentionned that tin-rich Yunnan might have served as independent entry point of bronze-making into SEA.

Further back in time, we have this:
http://www.nature.com/cr/journal/v26/n1/full/cr2015147a.html

"Around 15 000 years ago, a subset of ancestral dogs started migrating [from southern EA] to the Middle East, Africa and Europe, arriving in Europe at about 10 000 years ago. One of the out of Asia lineages also migrated back to the east, creating a series of admixed populations with the endemic Asian lineages in northern China before migrating to the New World."

I have some issues with that paper (insufficient wolf sampling, no ancient DNA), and the timeline is clearly off (NE/Eur dog evidence 15 kyBP, NAmer 10 kyBP). Nevertheless, while to me the original place where the ancestors of current dogs were domesticated is still open, latest by the epi-paleolithic the domesticated dog was spread all the way from Britain to China, which should imply corresponding human migration.

Tobus said...

@epoch:
In a direct comparison K14 forms a clade with WHG, MA1, East Asian and Native American relative to UI, indicating no significant "Basal" ancestry:

Chimp Ust_Ishim Karitiana Kostenki14_UP -0.0061 -0.94
Chimp Ust_Ishim Han Kostenki14_UP -0.0064 -1.055
Chimp Ust_Ishim MA1 Kostenki14_UP -0.0012 -0.152
Chimp Ust_Ishim Loschbour Kostenki14_UP -0.003 -0.396


Alberto said...

@epoch

Yes, Kotias is not pure Basal Eurasian, but it does not have WHG admixture like Stuttgart or Anatolia Neolithic. It might have ANE on top of Basal Eurasian, but that wouldn't favour Loschbour over Kostenki (in theory).

@FrankN

Thanks. The relevant gene flow in this case should be Upper Paleolithic, and related equally to Dai and Han in East Eurasia, and to WHG and ANE in West Eurasia, but postdating Kostenki 14 (most likely, from the genetic point of view, from East Asians to West Eurasians, since that would make the difference significant as it is between Loschbour and Kostenki, but the opposite probably wouldn't).

Ok, so I'm not going to try to figure that out. For now I'll just take that Kostenki has no Basal Eurasian admixture and place him as an ancestral West Eurasian in the tree.

Kristiina said...

Frank, U5b is by far the oldest mtDNA in Finns. On Ian Logan site, Saamis seem to have at least two different U5b1b1a lines which are both shared with Finns. The Saami motif seems to be a haplotype with 16335. Finn and Saami specific U5b1b1a with 16144C is shared also with Maris, Bashkirs, Mordvins, Chuvash, Komi-Zyrjans, Yakuts, Belarussians, Russians, Poles and Slovaks. I see that among the new samples there is one U5b1b, i.e. CuiryLesChaudardes1 France 8.2 kya with mutations at T16092C, C16327T which are both absent in the Finn and Saami specific U5b1b1a which instead has 16144. The distribution of U5b1b1a in Sweden and Norway is Northern and it is especially frequent in former Saami areas. IMO, Finn and Saami specific U5b1b1a may very well have originated in Belarus area.

At the moment, on the basis of the ancient mtDNA there is no trail from Western Europe to Norway and Karelia. The closest we get is Karelia Popovo Mesolithic U4 (U4d1?) as PWC Gotland (Ajvide53, Ajvide58, Ajvide70 and Ire8) has plenty of U4d and PWC Hangvar Gotland (Ire 3) may be U4d1. Also Bad Dürrenberg forager in Germany (7000 BC) may have been U4d as well as Kunda Culture Spiginas Lithuania, but it is very difficult to distinguish between U4b and U4d. Yuzhnyy Oleni Ostrov U4a and Kola Peninsula Bolshoy U4a and Karelia Yuzhnyy Oleni Ostrov U5a1 and Kuola Peninsula Bolshoy U5a1 are Eastern (EHG), as well as Yuzhnyy Oleni Ostrov H2a2b.

epoch2013 said...

@Alberto

But it shares with K14 a non-WHG component. It also pops up among MA-1.

@Tobus

But that doesn't mean it has no basal Eurasian. It simply means they share a lot of ancestry.

@Both of you

The fact that EEF is composed from a WHG component and another component, plus the fact that it is absolutely certain that the French have MORE WHG component than Stuttgart make the D-stats I posted - They are from the Oase 1 paper, mind you - very clear. Even if the French have additional admixture next to EEF and WHG they are more likely to share it with K14, which would skew towards the French. Yet despite that, K14 doesn't prefer the French over Stuttgart.

European, Stuttgart; K14, African -0.0002 -0.3

epoch2013 said...

@Tobus and Alberto

The only thing that could possibly skew things is that WHG is composed of more than one component, a component which is more abundant in Stuttgart than in French, even if the French have more WHG. But why then the affiliation K14 has with Loschbour? The paper on the Cardium pottery paper had some D-stats clearly showing Stuttgart prefered KO1 over Loschbour.

epoch2013 said...

So that would be a nice D-stat: Does K14 prefer KO1 over Loschbour?

Chimp K14; Loschbour K14

capra internetensis said...

@Maju

OK, the La Rochette sequence (571 bp) seems to be 5 bp longer than RSRS (nominally 569 bp). 523, 524, and 3107 are dummies in RCRS. So there are insertions and maybe deletions.

T489C for M
RSRS
481 ctcatcaata caacccccgc
La Rochette
481 ctcatcaaca caacccccgc

C10400T for M
RSRS
10381 aaaaaggatt agactgagcc gaattggtat atagtttaaa caaaacgaat gatttcgact
La Rochette
10381 acaaaaagga ttagactgag ctgaattggt atatagttta aacaaaacga atgatttcga

The sequences align
--aaaaagga-tt agactgag-ccgaattggt-at atagttta-aa caaaacga-at gatttcgact
acaaaaagga tt-agactgag ctgaattggt at-atagttta aa-caaaacga at-gatttcga--

T14783C for M
RSRS
14761 caaaattaac cccctaataa aattaattaa ccactcattc atcgacctcc ccaccccatc
La Rochette
14761 cgcaaaatta accccctaat aaaactaatt aaccactcat tcatcgacct ccccacccca

--caaaatta-ac cccctaat-aa aattaatt-aa ccactcat-tc atcgacct-cc ccaccccatc
cgcaaaatta ac-cccctaat aa-aactaatt aa-ccactcat tc-atcgacct cc-ccacccca--

Looks good to me

epoch2013 said...

@epoch

"So that would be a nice D-stat: Does K14 prefer KO1 over Loschbour?

Chimp K14; Loschbour K14"

O, bloody hell. That obviously was a mistake. Any D-stat that will use the entire available genome of K14 and checks if Loschbour or KO1 is preferred.

Chad Rohlfsen said...

I think KO1 has Basal Eurasian. They are the only hunter that is nearly significantly further from Ust-Ishim and also shares the most drift with farmers and Bronze Age groups. KO1 is almost as far from Ust-Ishim as the early farmers.

What I am wondering is if there is Basal Eurasian in WHG, to the exclusion of Kostenki, making them closer to East Asians that way(meaning Gravettians have BE, but not SSA in farmers), then the farmers have Basal Eurasian, plus SSA, slightly neutralizing that relationship. This is what I think makes the most sense, as saying that a WHG group went in and replaced 40% of the aDNA in the Near East during the LGM, making them closer to Loschbour than K14.The same in North Africa. I think that it make more sense, as there is no evidence of Euro HG introgression into the Near East during the LGM.

Chad Rohlfsen said...

There are a few things which I am trying to work out here. Here are a few groups of Dstats.

Chimp Yoruba Loschbour Ust_Ishim -0.0216 -4.727 21532 22481 501362
Chimp Yoruba Loschbour Kostenki -0.0154 -3.091 18892 19485 466911
Chimp Yoruba Loschbour MA1 -0.0134 -2.647 15153 15563 370832
Chimp Yoruba Loschbour Karelia_HG -0.0033 -0.716 18529 18652 471295
Chimp Yoruba Loschbour Motala_HG -0.0016 -0.418 17744 17802 480720
Chimp Yoruba Loschbour Iberia_HG -0.0006 -0.121 15231 15248 435586
Chimp Yoruba Loschbour Hungary_HG 0.0017 0.312 12924 12881 368602
Chimp Yoruba Loschbour Karitiana -0.0016 -0.429 21470 21540 503209
Chimp Yoruba Loschbour Han -0.0014 -0.404 21912 21972 503209
Chimp Yoruba Loschbour Onge -0.0002 -0.068 21925 21936 503207
Chimp Yoruba Loschbour Iberia_MN 0.0034 0.900 18671 18543 471925
Chimp Yoruba Loschbour Iberia_EN 0.0073 1.859 19687 19403 483258
Chimp Yoruba Loschbour LBK_EN 0.0039 1.136 20610 20451 500583

Chimp Yoruba Han Onge -0.0011 -0.515 24795 24851 594922
Chimp Yoruba Han Iberia_MN 0.0041 1.596 24596 24393 554482
Chimp Yoruba Han Iberia_EN 0.0067 2.560 25411 25073 568947
Chimp Yoruba Han LBK_EN 0.0033 1.520 26490 26318 590988
Chimp Yoruba Han Anatolia_Neolithic 0.0010 0.463 26528 26476 592420
Chimp Yoruba Han Kotias 0.0037 1.150 21970 21809 510071

It's pretty complicated. I am trying a few more things at the moment.

Chad Rohlfsen said...

Maybe, scratch all that. I think I figured it out. Hold on a bit, so I am able to run a few more things.

Alberto said...

@epoch

The stats comparing to French are not relevant because French have basal Eurasian. It's already complicated enough as it is to try to add modern Europeans to the mix. What's relevant is comparing preferably ancient samples without Basal Eurasian Admixture (BEA) and other ancient ones with BEA and see how Kostenki behaves. And in those stats it looks clear that Kostenki doesn't have BEA.

@Chad

The only thing I can make out of this is that Han and Yoruba behave similarly, so I think that everything else equal, age matters (meaning that very ancient samples have lower affinity to modern ones). So I'm willing to ignore the stats with Han/Dai in favour of the ones with Ust-Ishim et al.

But yes, the question of how the ancient Near East got WHG ancestry is still completely open.

Kristiina said...

Chad, KO1 (Tiszaszölös-Domaha´za, 5780-5650 BC) has an unusual mtDNA R3/R1b. MtDNA R1 has been found in Manych Catacomb Temrta V Ukraina R1 x2, BA Takhilgat Uzuur-5 Mongolia R1b1 and maybe in Early Xiaohe China R1 (189-192-311, 183-189-192-311). Today, R1 reaches highest frequencies in Adyge and Kabardians (R1a), as well as in Yakuts (R1b). It is also found in India (R1a, R1b), Russia (R1a), China (R1a) and Finland (R1b). The sister branch of R1 is R2 which is 21–31 kya old in Iran.

epoch2013 said...

@Alberto

To be honest, they are a mix of 2 French and 2 Sardinians. But nevertheless that mix should have:

1) Far More WHG
2) Far more other admixtures (e.g. West Asian) shared with K14
3) Substantially less EEF = BEU.

Nevertheless, K14 is equally related to both that mix and Stuttgart. To me it makes it clear. There is no way around this. Mind you, in the K-14 paper itself there was an F3 graph also pointing to something likely. And ADMIXTURE shows it as well, even if the tool has its limits (RK explained that a while ago).

epoch2013 said...

@Alberto

"What's relevant is comparing preferably ancient samples without Basal Eurasian Admixture (BEA) and other ancient ones with BEA and see how Kostenki behaves. And in those stats it looks clear that Kostenki doesn't have BEA."

No, obviously not if these non-BEA samples have *other* affinities with K-14 than the BEA part. Such as Kotias has, or Loschbour. K-14 could have a small BEA part and a large WHG part. In your model this D-stat would prove the absence of BEA in K-14:

Stuttgart Loschbour; K14 African -0.0050 -7.5

But that is not true. It just shows that K-14 has a larger WHG part than a BEA part. It does not show absence of the BEA part.

epoch2013 said...

@Alberto

Take, as an exmaple, African-Americans. If given the choice in D-stats they will greatly favour Africans over Europeans. But they still ahve substantial European admixture.

@Chad

Chimp Yoruba Loschbour Ust_Ishim -0.0216 -4.727 21532 22481 501362
Chimp Yoruba Loschbour Kostenki -0.0154 -3.091 18892 19485 466911
Chimp Yoruba Loschbour MA1 -0.0134 -2.647 15153 15563 370832
Chimp Yoruba Loschbour Karelia_HG -0.0033 -0.716 18529 18652 471295

That may be due to age. The older the sample the more ancestral it shares with Africans which drifted in the more than 35.000 years between Karelia and Ust'Ishim.

Alberto said...

@epoch

It's complicated, so nothing is really certain here. The stats on the Kostenki paper that led the authors to conclude that it had Basal Eurasian admixture are the ones I posted above and similar ones with Dai. They used Han and Dai as populations that don't have Basal Eurasian Admixture and otherwise no relation to West Eurasians (they didn't have Ust-Ishim yet, apparently). And both Dai and Han preferred Loschbour over Kostenki. This was a clear indication that Kostenki had BEA (the other possible explanation is that Loschbour has East Asian admixture, but the authors probably discarded that possibility).

However, when using Ust-Ishim and Oase, this difference disappears. You might say that that's because Kostenki has a closer relationship to both Ust-Ishim and Oase than Loschbour has, which compensates for it having BEA. Yes, that's possible, but not too parsimonious (see later).

What rather seals the deal for me is that when you place Yoruba in the position of Han/Dai, the stats are similar: Yoruba prefers Loschbour over Kostenki.

So we have different possibilities:

- After Kostenki, but before Loschbour, Europe got an influx of East Asians *and* Sub-Saharan Africans.
- Kostenki has Basal Eurasian Admixture over Ust-Ishim, Oase *and* Loschbour, but is otherwise more closely related to Ust-Ishim and Oase than Loschbour is, while being more related to Loschbour than the other two are. In other words, Ust-Ishim/Oase + BEA -> Kostenki. And then Kostenki - BEA -> Loschbour. Somehow BEA entered in the chain temporarily and then left.
- Kostenki doesn't have BEA, Loschbour doesn't have SSA and East Asian admixture, and the stats with Han/Dai and the ones with Yoruba are biased by Kostenki's old age.

And probably a few more explanations. Given the choice I take the last one, but all are possible.

epoch2013 said...

@Alberto

I don't know if Oase 1 is a good reference. The stats in the Oase 1 paper make very clear Oase 1 is related to nobody, not Ust'-Ishim either. It doesn't prefer Ust'Ishim over K14:

Ust'-Ishim K14, Oase 1 African -0.0005 -0.6

Oase 1 doesn't prefer Ust'-Ishim over Loschbour:

Ust'-Ishim Loschbour, Oase 1 African -0.0002 0.3

And so on. Oase 1 is alsp not a good reference for the fact that is has far more neanderthal admixture as well.

But Ust'-Ishim indeed does share non-drift with K14 that can be explained by age.

epoch2013 said...

@Alberto

Anothet thing. Doesn't this D-stat Chad Rohlfsen just posted mean Yoruba prefers K14 over Loschbour?

Chimp Yoruba Loschbour Kostenki -0.0154 -3.091 18892 19485 466911

Alberto said...

@epoch

Precisely because Oase is not related to Ust-Ishim or Kostenki, it helps to discard that option (Matt rightly pointed me to this above when I suggested the possibility of stronger affinity between Ust-Ishim and K14).

The stat with Yoruba says the opposite: Yoruba prefers Loschbour over Kostenki.

FrankN said...

@Alberto: Dog-related migrations should clearly post-date Kostenki. However, the Aurignacien (Kostenki) and the Mal'ta-Buret Complex are linked by Venus figurines, and also similar tools and dwellings (e.g. "cold storage pits"), indicating that Andronovo possibly wasn't the first AMH expansion out of the European Steppe into the Altai.

As concerns dog-related migrations, it might be useful to recall the earliest dated finds of safely-identified domesticated dogs in human burial contexts:
- 14.5 ky calBP La Morin (F), Magdalenien
- 14.0 ky calBP Bonn-Oberkassel (D), late Magdalenein/ Federmesser
- 14.0 ky calBP Ain Malhala (ISR), Natufian

Apparently OHG and Pre-Proto EEF were linked during the UP - whatever the direction of spread. I personally tend to domestication starting in the Magdalenian for the following reasons:
a) Greater value added in a peri-glacial environment (hunting assistance, fur, feed conversion, i.e. fattening on excess prey, slaughtering outside hunting/gathering season);
b) Cultural impact, namely hunting with bow & arrow, which requires dogs to retrieve wounded prey, first evidenced for Ahrensburgian;
c) NE had fox domestication (PPN Jericho) as alternative model;
d) the presumptive "Urwort" *khu(r), reflexes of which are widespread around the globe (c.f. Shar Pei, Dingo, Itzquintli, Georg. "kali", Hausa "kare", Sumer. "ur-gi", Quech. "alkho" etc.) looks better preserved in Basque "(za)khur", Ir. "cu" than in Semitic "kalbu".

@Kristina: There already was a 2013 analysis on a restricted set of Paleolithic mtDNA. Their phylogenetic tree (Fig. S2) clustered Sami U5b1 together with Italian, Berber and Fulbe samples. Right next to that clade's root sits Oberkassel_999 (U5b1). It only yielded 16192, which seems to be basal. That paper estimated the TMRCA for U5 at 30ky. I am anything but good in all that TMRCA arithmetic, but it doesn't look impropable to me that a 14 ky old sample could still be basal to U5b1.
http://www.cell.com/current-biology/abstract/S0960-9822(13)00215-7?_returnURL=http%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0960982213002157%3Fshowall%3Dtrue

While migration via Belarus can't be excluded, a maritime route looks much more plausible for better and less seasonal food supply. Quick colonalisatian along the Norvegian coast, starting directly at the end of the Younger Dryas, and reaching Finnmark by 11 ky BP, is archeologically well documented.
http://www.academia.edu/9402828/Bjerck_H._B._2007_Mesolithic_coastal_settlements_and_shell_middens_in_Norway._In_Milner_N._O._E._Craig_and_G._N._Bailey_Shell_Middens_in_Atlantic_Europe_5-31._Oxford_Oxbow

If that is the source of Finnish U5b1b1a, it should have parted from the Doggerland sisters around 11.5 kyBP, which would provide sufficient time for genetic differentiation, and leave little hope for finding direct EHG parallels. While Fig. S1 of the current study places CulryLesChaudarches, and also Bichon and Loschbaur on seperate U5b subclades, Iboussieres31_2 sits right next to Oberkassel_999, and might be worth a check.

epoch2013 said...

@Alberto

So all those northern WHG/EHG types grew more and more African over time?

Chimp Yoruba Loschbour Ust_Ishim -0.0216 -4.727 21532 22481 501362
Chimp Yoruba Loschbour Kostenki -0.0154 -3.091 18892 19485 466911
Chimp Yoruba Loschbour MA1 -0.0134 -2.647 15153 15563 370832
Chimp Yoruba Loschbour Karelia_HG -0.0033 -0.716 18529 18652 471295
Chimp Yoruba Loschbour Motala_HG -0.0016 -0.418 17744 17802 480720
Chimp Yoruba Loschbour Iberia_HG -0.0006 -0.121 15231 15248 435586
Chimp Yoruba Loschbour Hungary_HG 0.0017 0.312 12924 12881 368602

Rob said...

About Ko1- we should remember that he is a Neolithic sample which "looks Mesolithic "
That might explain why Chad is finding some basal in him

epoch2013 said...

Oase 1 is important as it seems utterly non-related to anything in Europe. How can that be? How can they have come into Europe without picking up BEA as well as WHG? It even seems not related to Ust'-Ishim.

Can some one do the following D(Non-African1, Non-African2; Early Modern Human, African)?


Suruí, Mixe; K-14, Yoruba
Suruí, Mixe; Oase-1, Yoruba
Suruí, Mixe; Ust'-Ishim, Yoruba
Suruí, Mixe; Ma-1, Yoruba

So, to make clear what I want. I want to see if the remainder of the genome of K14, Oase, Malta as well Ust-Ishim when compared with an African fits Suruii or Mixte.

Just to see how much of the mystery population that added to American Indians also added to European (-ish) paleolithics. Mixe apparently has no such admixture, Suruí the most.

Chad Rohlfsen said...

Oase1 has too much Neandertal to show relation to anyone. If you take it the excess Neandertal, he is probably identical to Ust-Ishim. There is the possibility of age being a factor here. The other two are Gravettians come off Ust Ishim after Kostenki, so they are slightly closer to Asians. Number two, East Asians and WHG share extra Neandertal admixture, post Kostenki. That doesn't really jive for the African and Near East preference, so Gravettians might just be closer to East Asians, if age is not the culprit. Looking at how MA1 is also acting, it may be age. It might be a question for Nick, if two pops are drifted more from Ust Ishim and OoA if they'll look artificially close.

Kristiina said...

Frank, this paper is about U5: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0010285. It is not very recent, but, for example, the phylogenetic tree (Figure S2) is interesting. In this paper, the age of U5b1b1 is between 3.7 and 10.8 kya. In their phylogenetic tree, the Finnish and Saami haplotypes are close to Belarusian haplotypes. When you look at the phylogenetic tree, you see that the Berber haplotype is U5b1b1e and the closest European haplotype to the Berber haplotype is found in a Czech! On Ian Logan site(http://www.ianlogan.co.uk/sequences_by_group/u5b1b1_genbank_sequences.htm), there are other U5b1b haplotypes that are found in Africa but none of them have Saami correspondences but, instead, are related to West Slavic or German haplotypes ( C16192T!), or Spanish haplotypes ( U5b1b1b), or Scandinavian haplotype (U5b1b1e).

Of course, it is possible that U5b1b1a1 took the coastal route via Norway. We need ancient DNA to prove or disprove it. However, it is not insignificant that that this new paper only found one U5b1b sample which is not close to the Finnish U5b1b. To date, we have very few ancient U5b1b1 specimens. Apart from this new CuiryLesChaudardes1 U5b1b sample, there is only one ancient U5b1b sample, Janisławice forager (6000 BC), and Janisławice is in the middle of Poland and Poland is close to Belarus.

Chad Rohlfsen said...

Right now, I'm leaning towards either age or more Neandertal admixture. I don't see Gravettians being different from Kostenki at the beginning. People need to ignore admixture components on things that old. Ust Ishim didn't have Near Eastern or African ancestry. Neither does Kostenki. I think drift has done a number on WHG. That is why they drifted to their own pole. I think that even with 10k years of separation, hunters in Ukraine can be identical to those in Iberia. Just look at Native Americans. Some tribes have been isolated from others for almost 15k years, yet are still almost identical. Native Americans are not mixed with a ghost pop to sit way out on their own pole. It's isolation. Just as will be shown for WHG in isolation for 30k years.

Tobus said...

@epoch:
But that doesn't mean it has no basal Eurasian. It simply means they share a lot of ancestry.

"Basal Eurasian" is defined as being non-African and non-UI - ie a Eurasian population that is basal to (diverged before) UI. It was invented to explain why all non-Africans form a clade under UI except for EEF (and pops with EEF ancestry). If Kostenki contained admixture from a population NOT descended from UI we'd expect a non-zero D-stat against the "pure-UI" clade, and we don't see that.

K14 doesn't prefer the French over Stuttgart.

But that doesn't mean it *has* basal Eurasian. Indeed, we see the exact same thing with UI in the same table:
European, Loschbour; UI, African -0.0027 -5.7
European, Stuttgart; UI, African -0.0001 -0.2

So all those northern WHG/EHG types grew more and more African over time?

Or vice versa. The Mota paper indicated European admixture into Yoruba post-Mota which would pull Africans towards modern West Eurasians more than older ones.

Oase 1 is important as it seems utterly non-related to anything in Europe.

I'm not sure where you got that idea - it's just not *closer* to anything in Europe than what we already had. It shows a strong affinity to K14 and prefers the UI branch over Basal-affected moderns, eg:

European Oase K14 African: -0.0049 -7.3
European Loschbour Oase African: -0.0021 -4.5
Stuttgart Loschbour Oase African: -0.0015 -2.4

I suspect it was a "Basal Eurasian" that has no surviving descendants and was closer to UI than the Basal Eurasian that left it's mark in EEF.


Here are the D-stats you requested, sorry no Oase as I don't have it:
Chimp Kostenki14_UP Loschbour HungaryGamba_HG -0.0126 -1.325

Surui Mixe Kostenki14_UP Yoruba 0.0038 1.002
Surui Mixe Ust_Ishim Yoruba 0.0048 1.492
Surui Mixe MA1 Yoruba 0.002 0.489




FrankN said...

So all those northern WHG/EHG types grew more and more African over time?

Or they gained a component that is also present in today's Africans.

The Raghavan e.a. 2015 paper on the peopling of the Americas has tree-mix diagrams in the SuppMat. The 4th admixture edge showing up there is from Dai (directly, not the Dai-Han branching) to Yoruba. This edge remains strong and constant, in spite of further edges wildly shaking around the tree geometry. [The first edges are (1) MA1->Athabascan, (2) French/Sard. branch to MA1->Altai, (3) MA1->W.Greenlanders.]
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4733658/

I first thought about some kind of artefact that results from a lack of NE populations, until the Mota study came out. While we are probably not talking direct Dai->Yoruba admix here, the signal bypasses Han and is as such apparently Indian Ocean, ASI and/or "teal"-related. I have elsewhere already talked about linkages: Goats, bananas and Dravidian metalworkng terminology to WA, black-eyed peas, donkeys and sorghum to IVC, and yDNA T that is spread from the Caucasus/ Zagros to Northern Nigeria, Madagascar, and Indonesia.

T, possibly forming part of CHG/"teal", has been found in Elbe-Saale LBK aDNA, though AFAIK not yet in any other EN/MN contexts, which could mean it arrived indepently from EEF in Central Europe. Kristina has just informed us about KO1´s mtDNA, today centered on the North Caucasus. La Brana's yDNA C is also not typically European... Seems like we all have forgotten that people were able to walk, and, more importantly, to use boats, already before the onset of the Neolithic.

Can one of you guys design some clever D-stats to figure out whether there is non-Kotias, but otherwise teal-related ancestry in Africa and EHG? Kurdish samples? Kalash? Dai-African runs with Gorilla and Greenlanders as outgroups?

Chad Rohlfsen said...

Here is a qpAdm run for Loschbour. I still don't find any actual East Asian or MA1 to be necessary. Loschbour is no closer to East Asians than MA1, who actually has roots in the southern half of Asia. It makes zero sense, at all, to have everyone be equally admixed with East Asians. It has to be age. Loschbour can even fit as 100% Kostenki, with a better fit than Karitiana as MA1 and Han.

left pops:
Loschbour
Kostenki
MA1
Han

right pops:
Mbuti
Ami
Ju_hoan_North
Japanese
Yoruba
Karitiana
Papuan

0 Loschbour 1
1 Kostenki 1
2 MA1 1
3 Han 33
4 Mbuti 10
5 Ami 10
6 Ju_hoan_North 5
7 Japanese 29
8 Yoruba 16
9 Karitiana 12
10 Papuan 14
jackknife block size: 0.050
snps: 133461 indivs: 132
number of blocks for block jackknife: 703
dof (jackknife): 568.113
numsnps used: 76829
codimension 1
f4info:
f4rank: 2 dof: 4 chisq: 5.992 tail: 0.199773413 dofdiff: 6 chisqdiff: -5.992 taildiff: 1
B:
scale 1.000 1.000
Ami 1.602 -0.609
Ju_hoan_North -0.012 0.020
Japanese 1.551 -0.583
Yoruba 0.020 0.096
Karitiana 0.856 2.298
Papuan 0.542 -0.043
A:
scale 112.201 575.618
Kostenki -0.167 -0.331
MA1 0.227 1.681
Han 1.709 -0.256


full rank 1
f4info:
f4rank: 3 dof: 0 chisq: 0.000 tail: 1 dofdiff: 4 chisqdiff: 5.992 taildiff: 0.199773413
B:
scale 1.000 1.000 1.000
Ami 1.596 -0.646 0.567
Ju_hoan_North -0.028 -0.038 1.205
Japanese 1.554 -0.586 -0.279
Yoruba -0.004 0.021 1.821
Karitiana 0.855 2.288 0.083
Papuan 0.552 -0.024 -0.910
A:
scale 113.035 576.052 3586.669
Kostenki -0.184 -0.326 -1.691
MA1 0.220 1.682 -0.348
Han 1.708 -0.251 -0.137


best coefficients: 0.783 0.162 0.055
ssres:
-0.000428401 -0.000462764 -0.000095222 -0.000703754 -0.000123750 0.000281402
-1.051196533 -1.135516243 -0.233651970 -1.726848974 -0.303653683 0.690494410

Jackknife mean: 0.780193492 0.164540455 0.055266053
std. errors: 0.077 0.075 0.024

error covariance (* 1000000)
5870 -5477 -393
-5477 5665 -188
-393 -188 581


fixed pat wt dof chisq tail prob
000 0 4 5.992 0.199773 0.783 0.162 0.055
001 1 5 10.966 0.0520544 0.832 0.168 0.000
010 1 5 10.216 0.0693506 0.941 0.000 0.059
100 1 5 104.331 6.45022e-21 -0.000 0.968 0.032
011 2 6 14.971 0.0204872 1.000 0.000 0.000
101 2 6 105.667 1.64442e-20 0.000 1.000 0.000
110 2 6 1971.212 0 0.000 0.000 1.000
best pat: 000 0.199773 - -
best pat: 010 0.0693506 chi(nested): 4.224 p-value for nested model: 0.039856
best pat: 011 0.0204872 chi(nested): 4.755 p-value for nested model: 0.0292153

Chad Rohlfsen said...

Also, Kostenki is about as good a fit as the admixing population into Karitiana, as MA1.

Maju said...

Thank you very very much, Capra. I was completely puzzled but your take, possibly a 2-bases insertion before 769 (which is the lowest numbered apparent mismatch), solves the problem.

I wanted to check for possible downstream M markers, because I really find hard to believe that the lineages can't be classified downstream of macro-haplogroup M (whose branches should by then be all well defined), but I logically began by testing the M-defining markers and almost none seemed to match. Now, thanks to you, I can proceed.

human443 said...

Chad, I think the reason why that fit is so low, is because any East Asian ancestry in Loschbour is going to be very ancient, i.e. not share the bias towards Karitiana/Atayal/Japanese that Han would have over say Onge. That would likely explain why East Asian input gets kicked out. Maybe trying Onge would actually work? Also, the ANE like ancestry might be closer to EHG, as suggested by their closeness in d-stats.

Krefter said...

@Chad, Tobus,

Here's a strategy of mine using D-stats to detect non-EEF Middle Eastern ancestry.

https://docs.google.com/document/d/1Ng6kCvs4OGFeuF3qjnDDWDsTZVPdQHNnoR9KBQA1ZYE/edit

Maybe you should do a few tests to see how they turn out, and if they make sense, do more tests.

It can be used on any humans. It could easily confirm ANE or Western admixture in Native Americans.

Chad Rohlfsen said...

I tried the Onge. There's not really any difference.

Chad Rohlfsen said...

Krefter,

If you want to look for an admixing population, you need to go about it a different way. Your Dstat requests will give obvious answers and some, not so much.

For Italians, you'll want to do something like;

Remedello_BA Italian_Tuscan German_Beaker Anatolia_Neolithic -0.0182 -8.235 15860 16450 363972
Remedello_BA Italian_Tuscan German_Beaker Kotias 0.0093 2.369 13766 13513 317778
Remedello_BA Italian_Tuscan German_Beaker Georgian 0.0104 4.678 16348 16012 364138
Remedello_BA Italian_Tuscan German_Beaker Armenian 0.0044 1.972 16297 16156 364138
Remedello_BA Italian_Tuscan German_Beaker Druze 0.0019 0.963 16277 16214 364138
Remedello_BA Italian_Tuscan German_Beaker BedouinB 0.0017 0.777 16353 16297 364138
Remedello_BA Italian_Bergamo German_Beaker Anatolia_Neolithic -0.0212 -9.662 15782 16465 363972
Remedello_BA Italian_Bergamo German_Beaker Kotias 0.0051 1.368 13667 13527 317778
Remedello_BA Italian_Bergamo German_Beaker Georgian 0.0061 2.828 16238 16042 364138
Remedello_BA Italian_Bergamo German_Beaker Armenian 0.0003 0.151 16195 16185 364138
Remedello_BA Italian_Bergamo German_Beaker Druze -0.0017 -0.855 16186 16241 364138
Remedello_BA Italian_Bergamo German_Beaker BedouinB -0.0028 -1.282 16243 16334 364138
Remedello_BA Italian_South German_Beaker Anatolia_Neolithic -0.0128 -4.241 15943 16355 362248
Remedello_BA Italian_South German_Beaker Kotias 0.0168 3.108 13835 13379 316298
Remedello_BA Italian_South German_Beaker Georgian 0.0153 5.161 16428 15931 362414
Remedello_BA Italian_South German_Beaker Armenian 0.0105 3.527 16398 16056 362414
Remedello_BA Italian_South German_Beaker Druze 0.0087 3.224 16375 16094 362414
Remedello_BA Italian_South German_Beaker BedouinB 0.0071 2.386 16441 16209 362414
Remedello_BA Italian_WestSicilian German_Beaker Anatolia_Neolithic -0.0132 -5.791 16017 16447 363971
Remedello_BA Italian_WestSicilian German_Beaker Kotias 0.0130 3.257 13868 13513 317777
Remedello_BA Italian_WestSicilian German_Beaker Georgian 0.0151 6.715 16503 16011 364137
Remedello_BA Italian_WestSicilian German_Beaker Armenian 0.0110 4.804 16485 16125 364137
Remedello_BA Italian_WestSicilian German_Beaker Druze 0.0081 3.901 16461 16196 364137
Remedello_BA Italian_WestSicilian German_Beaker BedouinB 0.0088 3.854 16557 16269 364137
Remedello_BA Italian_EastSicilian German_Beaker Anatolia_Neolithic -0.0152 -6.541 15999 16494 363971
Remedello_BA Italian_EastSicilian German_Beaker Kotias 0.0103 2.626 13838 13555 317777
Remedello_BA Italian_EastSicilian German_Beaker Georgian 0.0137 6.073 16495 16049 364137
Remedello_BA Italian_EastSicilian German_Beaker Armenian 0.0093 4.108 16473 16170 364137
Remedello_BA Italian_EastSicilian German_Beaker Druze 0.0062 2.958 16443 16241 364137
Remedello_BA Italian_EastSicilian German_Beaker BedouinB 0.0078 3.377 16563 16305 364137

Chad Rohlfsen said...

Basically, you have a starting point, an end point, and two possible admixing populations.

epoch2013 said...

@Tobus

"I'm not sure where you got that idea - it's just not *closer* to anything in Europe than what we already had. It shows a strong affinity to K14 and prefers the UI branch over Basal-affected moderns"

It shuns UI. But I already figured that could also be due to extra Neanderthal. I don't see the strong affinity to K14 in the papers D-stats, though.

European Oase K14 African: -0.0049 -7.3

is in the paper: 0.0049 7.3

However, you are right on UI behaving as K14. Damn, I missed that one :-/

OK. So, for the sake of the argument we call UI ancestral to K14. Then if K14 has additional BEA admixture we should be able to see it if we subtract K14 from UI, in other words use UI as outgroup.

Any ideas on what D-stat to use?

Kristiina said...

Frank, you are probably interested in this tree Shaikort recently posted on the thread Ancient Greeks and Romans may have imported a whole new genetic cline into Europe https://drive.google.com/file/d/0B3vEDdpZDjUpaEJGQUVScmtCeVk/edit?pref=2&pli=1
There are quite a few exciting migration edges:
American → Beringian
American → NE Euro
American → San
Caucasian → E African
San → Papuan
Root African, excluding Pygmy → Mediterranean
Pygmy → Root East Asian

Chad Rohlfsen said...

epic,

There is no Basal Eurasian in Kostenki. We've posted all you need to see.

Krefter said...

@Chad,

Thanks a lot for the stats. Here's me analysis of the results.

Why There are Regional Trends
More Negative: Italy is closer to German_Beaker as opposed to Middle East(inclu. EEF) than other Italians.

More Positive: Italy is less close to German_Beaker as opposed to Middle East(inclu. EEF) than other Italians.

The results tell us....

German Beaker-opposed to EEF, is higher in North Italy than South. It has risen in Italy since Copper age.
German Beaker-opposed to West Asian, is lower in all Italians than Copper age Italy except maybe in Bergamo. It has a North-South trend, with North having almost as close relation to Beaker-opposed to West Asian as does Copper age Italy.

These stats support German Beaker and West Asian-related ancestry in all Italians(except no suggesting of West Asian in Bergamo). But we need to treat the results as German Beaker-opposed to EEF and German Beaker-opposed to West Asian.

The *opposed* part is most important. Karelia_HG is closer to English opposed to Italians than Scottish are closer to English opposed to Italians. It doesn't mean Karelia_Hg is closer to England than Scottish are, it means Scottish has common ancestry with Italians that Karelia_HG lacks.

Kristiina said...

As U5 is at the core of this new paper, these conclusions from Malyarchuk et al 2010 are relevant:
“Tambets et al. have noted the considerable diversity of the U5b1b1 subcluster in western and southern Europe, suggesting that these regions, rather than eastern Europe, were the probable place of origin of U5b1b1. Analysis of complete mitochondrial U5b-genomes indicates that the origin of the whole U5b1 subcluster can be associated with southern and central parts of Europe, because each part of U5b1-phylogeny demonstrates the presence of earlier subclusters of Mediterranean prevalence.”
This new paper found U5b1 in France 12 and 10 kya (Iboussieres31-2, Ranchot88)

“The most ancient identified subhaplogroup, U5b2, requires further phylogeographic studies. However the data presented here allow us to suggest that at least subcluster U5b2a is characterized by a predominantly central European distribution, since a large number of U5b samples from Poland, Slovakia and the Czech Republic fall into this subcluster. For instance, subcluster U5b2a2 is frequent in central Europe (with the highest frequency of its subcluster U5b2a2a1 in Poles) and dated as arising between 12–18 kya, depending on the mutation rate used. It is also remarkable that within U5b2a1a, a Mediterranean branch precedes subcluster U5b2a1a1, which is characteristic of central and eastern European populations. Another U5b subcluster, U5b3, has its most likely homeland in the Italian Peninsula, from where it expanded in the Holocene along the Mediterranean coasts. Hence, in general one may conclude that an initial diversification of U5b occurred in southern and central Europe, followed by the spread of a particular U5b subclusters into eastern Europe.”

This new paper found U5b2b in Italy 18.6 kya (Paglicci71) and in France 13 kya (Rochedane) and 12 kya (Iboussieres39, Iboussieres25-1)
“Subhaplogroup U5a appears to be younger than U5b, and its two subclusters, U5a1 and U5a2, expanded soon after the LGM. Subcluster U5a2 is relatively frequent in central and eastern Europe, but some haplotypes within U5a2b were detected in Mediterranean populations as well (Italians, Tunisians). Subcluster U5a1 is also present in different populations of central and eastern Europe. It is noteworthy that one of its subclusters, U5a1d, demonstrates a coalescence ages varying from 19 to 21 ky (19.0±4.2 ky for complete mtDNA clock rate to 21.4±8.4 ky for synonymous clock rate), which are comparable with the age of the whole U5a subcluster. Subcluster U5a1d has been detected in eastern Europe (in Russians, Belorussians and Tatars) and southern Siberia (in Buryats). The presence of such old U5a1 lineages in eastern part of Europe may be indicative of eastern European origin of the whole U5a1 subcluster, however ancestral haplotype for U5a1 has been found in central Europe in Czechs.”
This new paper found U5a2 in France 10 kya (LesCloseaux3) and 9.3 kya (MareuilLesMeaux1) and U5a2c in Germany 8.7 kya (Felsdach).

Kristiina said...

U5a1 was not detected! Instead, U5a1 probably arose somewhere further east as we have these samples (not exhaustive):
Mesolithic Baikal Lokomotiv and Ust’ Ida U5a1a, Mesolithic Stora Förvar cave Sweden U5a1, Motala PWC HG x2 U5a1, Ajvide PWC [ajv 5, 29a] U5a1a’g, PWC Fridtorp Gotland [Fri 27] U5a1, Tianshan Beilu BA China U5a1a
Lebyazhinka Mesolithic Russia (c. 7500 BC) U5a1, Karelia Yuzhnyy Oleni Ostrov U5a1, Dnieper Eneolithic Molyukhov Bugor Ukraina U5a1, Manych Yamnaya Peschanyi U5a1, Nevskoye Catacomb Ukraina U5a1, PWC Ajvide Eksta Gotland U5a1, Kuola Peninsula Bolshoy U5a1, Ust Tartas Mesolithic Sopka West Siberia U5a1, Karakol Altai mound 3 Gorny Altai U5a1, Andronovo Oust-Abakansty Russia U5a1, Late Xiaohe China U5a1
Khvalynsk II Samara Eneolithic U5a1i, Yamnaya Peshany V Kalmykia U5a1i

However, U5 itself probably arose in Europe: (from Malyarchuk et al)
“It has been suggested that hg U5, or a genetically close ancestor to U5, arose among the first settlers of Europe, and subsequently spread all over the continent along with the Aurignacian industry. According to archaeological data, this Upper Paleolithic industry spread rapidly across western and central Europe roughly 42 to 40 kya, and there is evidence for a slightly earlier influx in southern and central Europe.
This new paper detected U5 in Czech Republic 30 kya (Goyet2878-21) and in Belgium 26.7 kya (DolniVestonice16 and 43)

Krefter said...

@Chad,

Your method is exactly the same as mine, except German_Beaker replaces what would be the second EEF. And the first set you did with Italians uses two EEFs like mine.

Remedello_BA Italian_Tuscan German_Beaker Anatolia_Neolithic

Can you explain how your method is differnt? Because I don't see how they're differnt. Do you mind if I throw some more D-stats at you today?

Tobus said...

@epoch: is in the paper: 0.0049 7.3

So it does, my bad.

Not really sure how to isolate Basal in K14 as opposed to K14/WHG in later pops, but here are some basic stats that might give you some ideas:

Ust_Ishim Kostenki14_UP French Kotias -0.0124 -2.329
Ust_Ishim Kostenki14_UP Han Kotias 0.0552 8.398
Ust_Ishim Kostenki14_UP Loschbour Kotias -0.039 -4.652
Ust_Ishim Kostenki14_UP Kotias Chimp -0.0552 -7.623
Ust_Ishim Kostenki14_UP French Chimp -0.0623 -10.733
Ust_Ishim Kostenki14_UP Han Chimp -0.0011 -0.168
Ust_Ishim Kostenki14_UP Loschbour Chimp -0.0886 -11.335

Krefter said...

@Chad,

Can you run these? I have a feeling EEF as close affinity to SouthWest Asians, which was diluted by CHG and ANE.

Remedello_BA Italian_Tuscan Saudi Kotais
Remedello_BA Italian_Tuscan BedouinB Kotais
Remedello_BA Italian_Tuscan Druze Kotais
Remedello_BA Italian_Tuscan Jordanian Kotais
Remedello_BA Italian_Tuscan Assyrian Kotais
Remedello_BA Italian_Tuscan Saudi Loschbour
Remedello_BA Italian_Tuscan BedouinB Loschbour
Remedello_BA Italian_Tuscan Druze Loschbour
Remedello_BA Italian_Tuscan Jordanian Loschbour
Remedello_BA Italian_Tuscan Assyrian Loschbour
Remedello_BA Italian_Tuscan Saudi MA1
Remedello_BA Italian_Tuscan BedouinB MA1
Remedello_BA Italian_Tuscan Druze MA1
Remedello_BA Italian_Tuscan Jordanian MA1
Remedello_BA Italian_Tuscan Assyrian MA1

FrankN said...

@Kristina, re U5:

Thanks for linking that older paper. I already knew it, but it was good to read it again in the light of the new findings. To add to your comparison, my key takeaways are the following:

1) I am with you in assuming a (Central) European late MP/ early UP origin of U5.

2) We now have confirmation on U5b2 in the Adriatic refuge (Paglicci). Assuming that Iboussieres (Provence) represents a Ligurian refuge, presence of U5b2 also there indicates the Appenines being less of a barrier during/after LGM than partly assumed.
For U5b1, we so far only have La Brana and two unreliable mesolithic Portuguese U5b1 (Chandler 2005). Nevertheless, it is fair to assume Iberia as primary U5b1 refuge. U5b1 in Iboussieres sugests communication with the Ligurian refuge during/after the LGM. Bichon and Oberkassel point to quick post-LGM expansion.

3) For U5a, a Carpathian Basin refuge (in interglacials was linked to the Adria) looks plausible. Post-LGM, U5a1 may then have expanded towards the Black Sea / NE Europe, while U5a2 moved into CE.

4) The U5 cousins appear to have intemixed after the LGM. In the French Jura (13 ky), U5b1 and U5b2 appear together, in the Swabian Jura (~9ky) I5a2, I5b1 and I5b2 have been found.

5) If I interprete Motala, chopped heads mounted on long piles, correctly, U5a1 wasn't welcone to her cousins' party at what was then still the western shore of Lake Ancylus. U5a1 possibly arrived by boat from the NE. You have nicely pointed the Motala connection via Gotland (Stora Förvar, PWC) to the Dniepr, Khvalinsk, Andronovo and ultimately Xinyang.
U5a1 mesolithic presence on Lake Baikal makes me wonder whether Motala shouldn't in fact be regarded as EHG. Anyway, a likely post-LGM origin of U5a1 in the Carpathian Basin, and of "Finnish" U5b1 in Iberia/ Southern France puts the whole EHG-WHG distinction on shaky grounds.

6.) U2 is an interesting case. Quite widespread in the UP (Kostenki, 3 out of 5 27ky Goyet Cave/BE finds), it gets lost from the CE mesolithic report, to reappear again with LBKT Hungary. As U2e and U5a1 appear in tandem in Motala (U2e1), Ust-Tartas and Yamnaya BG (U2e1a), it is fair to assume the partnership already commenced during the LGM, i.e. U2 also finding refuge in/near the Carpathian Basin (NW Black Sea?).
Unlike U5a1, U2(e) has quite a LN/CA presence: CW KAR28 U2e2, CW ESP28 U2e1a1, BB Kromsdorf U2e, CW Switzerland U2e1, Unetice EUL51 U2e1f, QLB34 U2e1f, QLB 37 U2e2 ROC4 U2e1, ROC10 U2e2, Chociwel/PL [RISE139] U2e1f1, further occurences in Poltava (U2d2), Poltapovka (U2e1h), Sintashta (U2e1h, U2e1e) and Andronovo(U2e, U2e2, U2e1). U2 (undiff.) has been furthermore found in Nuraghic Sardinia. We are talking several lines here, and it would be interesting to sort out to which extent they display Paleolithic continuity (which LBKT and Nuragic Sardinia most likely do), and how much of it is Steppe re-introduction. Note here that Poltava U2d2 is absent from the CE record, and Yamnaya BG U2e1a has but a single possible correspondence (CW ESP28, U2e1a1).

7.) What happened to U8? Completely exterminated, except for some Basque U8a holdout, und U8b in Italy and Jordan today? Not quite. There is a Maikop occurence (Klady, U8 undiff.), Unetice QLB38 (U8a1a), and ROC 2/3 (U8b1a1). Speculatively, I'd say some of the Swabian U8a cave girls made it into the Carpathian refuge and from there towards the Caucasus (Maikop). U8b might have survived on the Adria.

8.) Finally, I have noted the KO1's mtDNA R was already present 40ky ago in Fumane/I. Another possible R has been reported from Toledo, Louinha, POR (7 ky). So we are not necessarily talking about Caucasian immigration, KO1's mtDNA R3 might also represent Paleolithic continuity, and CE migration into the Caucasus as seems to have been the case with U8a.

Rob said...

Frank

"
5) If I interprete Motala, chopped heads mounted on long piles, correctly, U5a1 wasn't welcone to her cousins' party at what was then still the western shore of Lake Ancylus"

Ha ha. That's a bit of a tabloid-easque theorem. We don't know the haplotypes of those making the sacrifices, do we?

FrankN said...

@Kristina: "..this tree Shaikort recently posted.."
Thx. I had skipped that thread's comments, but, as quite common here, some OT gemstones should always be expected, and Shaikorth's tree is one of them!

To those edges:
-American → Beringian
Expected. Another confirmation on "out of America" (mesolithic), Dene-Yenissean etc.

-American → NE Euro
Somewhat expected. Ties in with my earlier post on strong genetic differentiation between Fennoscandian and Siberian reindeer herds, which suggests human Beringian DNA arriving in NE Europe rather by boat than with reindeer herders. Just that it apparently wasn't a Beringian, but an American arrival. Another tree run that includes Greenlanders could be interesting...
You, as a linguist, might want to dig a bit deeper into the proposed IE - Chukotko-Kamchatkan -Uralic macrofamily, which might extend to other Amerindian languages, e.g. Salish. Jaeger identified especially strong parallels between CK and Goidelic, which runs counterintuitively, except if one assumes sea-based contact.

-American → San
Unexpected, worth to keep an eye on. The other way round I would have understood, given the transfer of African Calabash to the American East Coast during the UP.

-Caucasian → E African
Expected. yDNA T, see above.

-San → Papuan
Unexpected. SEA to/from West Africa would have tied in with the tropical neolithic package (Banana, Taro, Yam), SEA->San could have been explained by "out of Madagascar" gene flow. But his has the wrong direction and endpoints for both.

-Root African, excluding Pygmy → Mediterranean
Some mix of EA yDNA E migration (whenever that took place), WestMed contacts, and otherwise a proxy for "Bedouins" that are missing in the structure?

-Pygmy → Root East Asian
Looks pretty basal and ancient. Possibly MP coastal route "out of Africa".

@Rob - on Motala: Heads mounted on stakes, well visible, on the only outlet of Lake Vättern, just a couple of kilometers away from what was then still the western shore of the Ancylus Lake (in the process of becoming the Baltic Sea) - I tend to call this "No Trespassing beyond this point !" rather than "sacrifices".
Maybe that interpretation is wrong. Strontium analysis would help to find out whether we are talking locals or foreigners. Such analysis has been announced, but I couldn't find any info on the result. In any case, the U5a1 aDNA trail is pretty clear:

- Motala ~7.8 ky
- Stora Förvar n. Gotland ~9.4 ky
- Lebyazhinka, Samara ~9.5 ky

"We don't know the haplotypes of those making the sacrifices, do we?"
No. We just can make interference from comparable sites of the Maglemose-Kongemose tradition. This gives us:
- Cough's Cage (Cheddar Man, ~10ky): U5 (unreliable)
- Blätterhöhle, Hagen, ~9ky: U5b2a2, U5a2c3, U/k, U5a, U2e [I forgot that case of paleolithic continuity in my list above]

Blätterhöhle is interesting, as it was re-used during neolithisation by Michelsberg & FB (6.0-5.2 ky). In that period, we have: 12 U5 (incl. 8 U5b, of which 6 U5b2), 3 H5, 2 H11, 1 J. When looking for places where EHG entered the MN/LN, NW Germany seems to be one of them.

terryt said...

@ Tesmos: "terryt, the origins of K2 is based on modern diversity which is misleading. we should wait for ancient dna from south East Asia to be certain".

To base it on anything other than the information we have available at present is to base it on what we want to believe. No matter how widespread K2 turns out to have been it is doubtful that K2b will ever be shown to have arisen within a Southeast Asian K2.

@ Ryan: "Having C1a in both La Brana and Jomon samples means that at some point, there is a special connection between those two groups at some point, no if ands or buts".

Exactly. But, of course, that doesn't necessarily mean they had any substantial autosomal genetics in common. As everyone presumably knows, once someone (A) carrying a particular haplotype has children with some carrying a different haplotype (B) the autosomal DNA in the offspring is reduced to half of A's. And so on. If the haplotype moves any great distance its 'original' autosomal DNA will be progressively reduced. Therefore to say any European carrying Y-DNA K2 should have some substantial proportion of SE Asian ancestry is ridiculous. In other words ENA, WHG and ANE are relatively independent of to any particular haploid DNAs.

@ Krefter: "The Ms don't share any mutations with modern M clades. They're their own clade".

That was my understanding of the figure in the paper. It makes sense it is unrelated to any surviving M lines. The early M lines outside SE Asia/East Asia have become extinct for one reason or another.

Krefter said...

@Davidski,

Actually the D(Chimp, Druze/Cypriot, Mbuti, European) stats show an interesting correlation and might be useful. Look at the ranking sheets of this spreadsheet.

https://docs.google.com/spreadsheets/d/1TsOTPEbJgHhvd3oenwcDPioRYVu4IKSICp2o2zsvsKU/edit#gid=1095468982

Whoever shares more with Druze shares more with Cypriot. Several SouthEast Europeans who are near the bottom in EEF sharing are near the top in Cypriot/Druze sharing.

Can you run the following stats to confirm the correlation.
Chimp Cypriot Mbuti Serbian
Chimp Cypriot Mbuti Montenegrin
Chimp Cypriot Mbuti South Italy
Chimp Cypriot Mbuti Croatian
Chimp Cypriot Mbuti Stuttgart
Chimp Cypriot Mbuti HungaryGamba_EN
Chimp Cypriot Mbuti Spain_EN
Chimp Cypriot Mbuti Spain_MN
Chimp Cypriot Mbuti Remedello_BA

Davidski said...

Krefter,

Chimp Cypriot Mbuti Serbian 0.4056 485205
Chimp Cypriot Mbuti Montenegrin 0.4059 485205
Chimp Cypriot Mbuti Italian_South 0.4035 485206
Chimp Cypriot Mbuti Croatian 0.4048 485206
Chimp Cypriot Mbuti Stuttgart 0.4118 476965
Chimp Cypriot Mbuti Hungary_EN 0.4138 432287
Chimp Cypriot Mbuti Iberia_EN 0.4122 413358
Chimp Cypriot Mbuti Iberia_MN 0.4094 387668
Chimp Cypriot Mbuti Remedello_BA 0.4147 293734

Kristiina said...

Frank, K comes out of U8b, and K is really very widespread in Europe and in the Middle East. Moreover, Satsurblia (13 kya) was K3. The exciting question is where U8b arose. This paper did not detect U8b but it may very well have arosen in the eastern Mediterranean. U8b has been detected in Barcin neolithic Turkey U8b1b1, Sopot proto-Lengyel Alsónyék Hungary U8b1b (yDNA J2), Schöningen NE Salzmünde Germany U8b1b, Unetice BA Rocken Germany U8b1a1. MtDNA K would be U8b2. K has been found in pre-pottery Syria, but it seems that we do not know how derived their K is.
You forgot that there are even three different U8a haplotypes in Finland (U8a1a1 C12135A, U8a1a1b G9777A, U8a1a1b1 A9338G). U8a is particularly common in Russian Karelia.
http://www.ianlogan.co.uk/sequences_by_group/u8a_genbank_sequences.htm

As for R1, it is true that R1 has not any closer relationship with R2, and R is a huge group covering Europe (mostly R0(HV) and R2 (J/T)), West Asia, India, East Asia and America

@Frank “U5a1 mesolithic presence on Lake Baikal makes me wonder whether Motala shouldn't in fact be regarded as EHG. Anyway, a likely post-LGM origin of U5a1 in the Carpathian Basin, and of "Finnish" U5b1 in Iberia/ Southern France puts the whole EHG-WHG distinction on shaky grounds.”

Apart from U5a1x2, Motala also has U5a2, U5a2d and U2e1. MN Bom Santo Portugal carries the same mutations at 16256 and 16270, as well as MN Treilles Aveyron France, but I do not know how reliable these results are. However, I agree that there should be some Siberian/ANE admixture in Motala.

Finns are mostly WHG, I have been found to be even 49% WHG in one of Davidski’s tests. Mesolithic EHG hunter gatherers in Karelia do not share mtDNA or yDNA with modern Finns, so there cannot be much genetic continuity.

@Frank “You, as a linguist, might want to dig a bit deeper into the proposed IE - Chukotko-Kamchatkan -Uralic macrofamily, which might extend to other Amerindian languages”

At the moment, I think that all these languages are modern language families that represent a modern way of life and are not Mesolithic survivors. Therefore, I think that it is the modernity that unites IE, Uralic and Chukotko-Kamchatkan and not the hunter gatherer past. However, older Mesolithic roots are there, and they could be recovered at least in part, but I do not know if those older roots correspond to modern family groupings.

Krefter said...

@Davidski,

Would you say most people in Russia's Far North have EHG/ANE admixture but lack West Asian admixture? That's what it seems like in D-stats and ADMIXTURE.

Davidski said...

I'm not sure, because even Mansis need Georgian as a minor third ancestry when being modeled as EHG/ENA. Here are those D-stats.

Remedello_BA Italian_Tuscan Karelia_HG Loschbour -0.0253 -4.471 296692
Remedello_BA Italian_Tuscan Kotias Bulgarian -0.0105 -2.889 323047
Chimp Karelia_HG Samara_HG Italian_Tuscan -0.1127 -17.087 285864
Chimp Karelia_HG Samara_HG Remedello_BA -0.1172 -13.013 179971
Chimp Satsurblia Kotias Italian_Tuscan -0.0906 -16.091 364176
Chimp Satsurblia Kotias Remedello_BA -0.0828 -10.452 232960

Kristiina said...

I have to correct what I said above! In Lazaridis paper, Motala U5a1 is identified with additional substitutions at G5460A (Motala 1) and G5460A + A9389G (Motala 3). I checked Ian Logan site for correspondences and found out that G5460A is found on two modern U5a1 haplotypes: U5a1 A15218G A16399G is found in Finland and USA (?); and U5a1a1b G15110A is found in Denmark and Belarus. So, Motala U5a1 may be an early mainly WHG offshoot of U5a1.

U5a1a1 T5495C A15924G is found in Portugal, and it may be related to U5a1 in MN Bom Santo Portugal and MN Treilles Aveyron France.

Kristiina said...

Frank, La Braña, 7960-7750 BP is not U5b1 but U5b2c1, so it is the same haplotype as that found in HohlensteinStadel Germany, Swabian Jura, 8.6 kya. La Braña was a foreigner from the Italian Ice Age refuge.

Rob said...

Kristiina

" I think that all these languages are modern language families that represent a modern "

I agree entirely. Can you elaborate perchance ?

FrankN said...

@Kristina: Thanks for reminding on the U8b/K relation. This implies that the study's conclusion of post-LGM popukation replacement becomes even more questionable, even though, as you correctly point out, we don't know to which extent European K represents paleolithic continuity, or Neolithic repopulation from the NE.
In any case, U8b/K is a potential connection between WHG, CHG and ENA. It may be tentatively linked to the spread of the domesticated dog around 15 ky BP across SW Euasia (Magdalenian - Natufian, possibly also Zarzian [Palegawra Cave, Iran] and Epi-Gravettian [Mezin/UKR, Elizevich/ RUS]).

In relation to R0/HV, this new paper may be of interest:
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0144391

"We also confirm a major lineage expansion that probably followed the Late Glacial Maximum and preceded Neolithic population movements. We finally observe that Italy harbors a reservoir of mtDNA diversity, with deep-rooting HV lineages often related to sequences present in the Caucasus and the Middle East. The resulting hypothesis of a glacial refugium in Southern Italy has implications for the understanding of late Paleolithic population movements. (..)
The major discovery of our analysis is the ancient divergence and isolation of the Italian HV* lineages. New sources of variation are found within HV4 (mostly in Southern Italy) and HV0 (mostly in Northern Italy), as well as within HV* and HV-16311*. Exclusive Italian lineages splitting at the root of the major branches can be assigned to new autochthonous haplogroups, such as the proposed HV4d, HV1e, HV0g. (..) If the Mesolithic specimen found in Favignana (Sicily) belonging to mtDNA haplogroup HV1 [18] will be confirmed, this may prove the antiquity of this haplogroup in Southern Italy."

FrankN said...

@Kristina: "I think that it is the modernity that unites IE, Uralic and Chukotko-Kamchatkan and not the hunter gatherer past.
I tend to agree. Can we specify "modernity", e.g. as CA, meaning processes that started in the late 4th mBC?

However, let ma also draw your attention to Jaeger's most recent publication, which now deals with the full World Language Tree:
http://evolang.org/neworleans/pdf/EVOLANG_11_paper_147.pdf

"The Eskimo-Aleut languages form a clade with the Chukotko-Kamchatkan languages (which could be interpreted as indicative of common descent by proponents of the Nostratic/Eurasiatic macro-family), but this clade is part of a larger clade comprising languages from the North-American Pacific Northwest, including Wakashan and the Salish. While the geographic proximity points to possible contact, there are no good reasons to assume that the involvement of Eskimo-Aleut in this larger clade should be of a genealogical nature."

The paper continues speaking about "chance similarities", which IMO may often be less of "chance", but owed to archeologically and/or genetically traceable movements:
"In a few cases — surprisingly few, we dare to say —, the observed patterns can only be interpreted as the result of the accumulation of chance similarities. For instance, Siouan (North America) and Alor-Pantar (Papunesia) are embedded in the SE Asia part of the tree, Ainu (East Asia) in the American part, and Dravidian (South Asia) in (or neighboring to) the Subsaharan African part. Finally, as mentioned earlier, the loci in the tree of two large clusters containing Papuan
languages are not meaningful, so they should probably be regarded as random."


I will post some of these "chance similarities" separately, my browser is in "reser mode" and has already killed my first attempt on this comment.

Krefter said...

@FrankN,

Early Neolithic Germany/Hungary had 19% K. 35%(9/26) Neolithic Anatolians have K. With over 100 mtDNA samples from Euro-HGs, 0% are K. So, IMO 99.99% of K in Europe is from West Asia/Balkans(2/2 Mesolithic Greeks had K1c).

Nirjhar007 said...

Frank and Rob,
If you can give a suggestion, between the period of 4000 to 3000 BC. Which Geographic location according to you in Eurasia had the highest amount of population?.

Rob said...
This comment has been removed by the author.
Rob said...

Nirjhar

Outside Mesopotamia - for Europe it mihjt be the C-T area (recently estimated at 700 K by Werninger et al). In Asia , one might presume the IVC ? But you might be more informed about the latter

epoch2013 said...

http://www.pbase.com/daveb/mtdnau
http://www.biomedcentral.com/content/pdf/1471-2164-7-124.pdf

Apparently the Var region in France has 8% U8a. South-West Europe 2%. North-East 0.6%. So locally there certainly was survival.

FrankN said...

Rob, Nijhat: A recent estimate for CT, based on systematic assessment of some 60 sites in the Southern Bug-Dniepr interfluve (SBDI), estimates
"during its time of peak population (Cucuteni B / Tripolye CI period, c. 3750 B.C.E.) the Cucuteni-Tripolye complex as a whole was composed of 27,000 people inhabiting roughly 15,000 km2, for an average density of 1.8 persons per km2. The SBDI, while accounting for only 10% of total inhabited territory, was home to 43% of the population and boasted an average density of 7.8 persons per km2."

http://www.acsu.buffalo.edu/~tkharper/Harper2013.pdf

This is well in line with, e.g. Rhinish LBK up to 8.5 p/km² in settlement chambers, but only 0.8 p/km² overall, as only 10% of the total area was exploited.
http://digitalcommons.wayne.edu/cgi/viewcontent.cgi?article=1060&context=humbiol

It has been established that even with MN innovations (dairying, ard), densities abvoe 10 p/km² could not be locally sustained. Where they exisrted, they required external food supply, typically from transhumaing pastoralists, e.g. by selling salt to them. The transhumance area, OTOH, would be far less densely populated, thus bringing the overall densities down.
Could you provide a link to the Werninger study? It seems like he applied a local peak density of 10 p/km² to the whole C-T territory of about 70 kkm².

On your original question, Nijhar: The most densely populated areas should be those combining (a) developed agriculture on good (quickly regenerating) soils, (b) a documented pattern of dense, long-term rural settlement, (c) possibly availability of additional aquatic ressources (rivers/ coasts), and (d) documented secondary activity (salt sooting, tool/ pottery manufacture) allowing to trade in agricultural surplus from the periphery.

In Europe come to mind here:
- N Catalonia (fish, saltmining)
- NE Italy (fish, stone processing, e.g.Remedello daggers)
- Maas/ Rhine (flint mining)
- Western Baltic (fish, amber, fur, ~2 P/km²)
- Elbe-Saale (copper, salt, possibly pottery, >2 P/km²)
- Moldova (Cucuteni - salt)
- Thrace (Varna Culture - fish, salt, gold, spondylus export)

Population hotspots, but probably too localised to bring about
sizeable concentration over a large scale, would include areas around today's Paris, Frankfurt/M, Prague, Cracow and Bratislava, Obsidian islands (Sardinia, Lipari, Melos), and lakes w. piledwelling (Balaton, Constance, Zurich, Gemeva, etc.).

In the NE, we have (West) Georgia (metal mining) and Mesopotamia (salt). At Babylonian times (2300 BC), S. Mesopotamia is estimated at 7.1 p/km²; so 2.5-3.5 p/km² might be a reasonable 4th mBC guess. The IVC population density has been estimated at 1.3 p/km² by 3500 BC (link), which would place it clearly behind Elbe-Saale, the W. Baltic and CT, behind Mesopotamia anyway.
http://arxiv.org/pdf/1110.1091.pdf

For the Yangshao culture in the N Chinese loess belt, estimates range up to 800k people, over maybe 500 kkm², which would yield a density of 1.6 p/km². Effectiv densities in the Wei Valley, the agricultural core, should have been substantially higher, and -in all fairness - it is the 21 kkm² Wei Valley to compare with the 20-25 kkm² Elbe-Saale region, not all of NC China. However, I haven't yet found a detailed assessment for the Wei Valley yet. Nevertheles, for the subsequent (3 mBC) Longshan Culture. a density of ~ 3 p/km² in the Rizhao area (22 kkm², Yellow Sea Coast) has been established.
https://journals.lib.washington.edu/index.php/BIPPA/article/viewFile/11873/10500

This leaves the big question mark - tropical SEA, especially Indonesia and the Phillipines with their fertile volcanic soils. At the time of first European penetration (1938), the New Guinea highlands had an average population density of 26.9 p/kim². Subsistence included species that clearly (sweet potatoe) or possibly (pigs) only reached there after 3 kBC. Nevertheless, it serves as illustration of the densities the "tropical neolithic package" (Banana, Yam, Taro, Coconut) may sustain.

Rob said...

Franks

Thanks for your more detailed reply

Sorry, I was horribly off in my recollection. You're correct, the figure estimated was 37 K

Link:

https://www.academia.edu/2320521/The_effect_of_climatic_variability_on_population_dynamics_of_the_Cucuteni-Tripolye_cultural_complex_and_the_rise_of_the_Western_Tripolye_giant-settlements

(Sorry, it s by Harper alone, not Werninger et al)

another similar
https://www.academia.edu/20139358/The_Geographic_Corridor_for_Rapid_Climate_Change_in_Southeast_Europe_and_Ukraine

Maju said...

@Epoch: For whatever is worth U2 is also common in Perigord-Limousin (4%) and areas of Normandy (5%), per Dubut 2004, However it was absent in Brittany and the Northeast (samples were small anyhow).

Nirjhar007 said...

Rob,Frank
Thank you guys :) good and important info!. The IVC had the population of 5,000,000 or more around the 2600-2200 BC period wen the civilization was at its peak. Roughly ( if i'm not mistaken) the 10 % of then global population.

Nirjhar007 said...

Maju,
whats your current position on N-Mtdna's origin? in light of the Paleolithic Mtdna in Europe.
Do you agree with Dienekes?
http://dienekes.blogspot.in/2016/02/mtdna-from-55-hunter-gatherers-across.html

I think you are also thinking on a post related to the study?.

Maju said...

@Nirjhar: IMO mtDNA N, which has 16 basal subclades and not just R, N1/I, N2/W and X (which are the more or less West Eurasian ones, a mere 1/4 of the total) must have coalesced in SE Asia (or Bengal as the furthest western possibility), maybe expanding in the open niches left by the Toba catastrophe (of course, I consider the first H. sapiens peopling of Asia quite older, but that is mostly related to mtDNA M). IMO some N clades and very notably mtDNA R, are associated with the expansion of Y-DNA K2, whose origin should also be in SE Asia, however others seem more related to Y-DNA C, as are also some M subclades, so probably mtDNA N is older than Y-DNA K2 but mtDNA R is roughly contemporary and strongly associated to it. The trail from SE Asia to the West, via North India, is most apparent in Y-DNA P1 and mtDNA R, but you can also see the K2+R association in East Asia and Papua: it's something very striking, really.

"I think you are also thinking on a post related to the study?"

I don't like enough the approach of the study, because it clearly uses only certain and not all the available ancient mtDNA of Europe, what leads to a very limited explanatory power (what about mtDNA H?), probably restricted to the Central European province. However I'm intrigued about that mtDNA M, because I'd expect it to be already under some of the more than forty basal sublineages of the macro-haplogroup. So I guess that, when I have time and willpower, I will explore the matter (no promises: I've been losing the passion and motivation that used to drive my blogging activity).

Nirjhar007 said...

Well don't worry Maju , you are living in the greatest era of blogging and things will get only better! :)and of course thanks for the detailed opinion!. I also saw your old post on it-
http://forwhattheywereweare.blogspot.in/2011/12/on-origin-of-mitochondrial-macro.html

terryt said...

"IMO mtDNA N, which has 16 basal subclades and not just R, N1/I, N2/W and X (which are the more or less West Eurasian ones, a mere 1/4 of the total) must have coalesced in SE Asia (or Bengal as the furthest western possibility)"

I doubt N coalesced in SE Asia. AS you point out it has separate eastern and western members. It appears to me that the present N haplotypes are remnants of a once widespread group. Just as the paper suggests is the case for M, I strongly suspect climate change was responsible for reducing that spread. Toba may have been responsible but I fail to see how Toba could have eliminated M in Europe yet left SE Asia untouched.

"IMO some N clades and very notably mtDNA R, are associated with the expansion of Y-DNA K2, whose origin should also be in SE Asia"

I agree, except I fail to see any N clades that may have been involved, at least until the movement was able to leave the subcontinent and push west and north. It then picked up some of the western N clades.

"probably mtDNA N is older than Y-DNA K2 but mtDNA R is roughly contemporary and strongly associated to it".

R's expansion looks to be totally independent from that of N. Presumably R coalesced in SE Asia soon after N arrived there.

"however others seem more related to Y-DNA C, as are also some M subclades"

C and N in Australia certainly indicate a close relationship between C and N when humans first entered that continent. However association between C and M probably post-date C's arrival in the East.

"The trail from SE Asia to the West, via North India, is most apparent in Y-DNA P1 and mtDNA R, but you can also see the K2+R association in East Asia and Papua: it's something very striking, really".

Interestingly the settlement of Papua looks totally independent from that of Australia. Very few haplotypes overlap. And of the ones that do it is quite easy to see which of the two regions they coalesced in.

Squad said...

The presence of haplogorup M is really puzzling because it's very weird that no M lines survived while C-V20 did given that we know maternal lines usually are in a much less intense competition than paternal lines and thus are normally not only able to survive but to avoid significant bottlenecks as well. In this case the fact that no kind of M that is not M1 or some gypsy M exists in Europe makes me doubt about these findings. And we have far enough modern samples today from all corners of Europe to be able to confirm this claim.

However, I agree with the claim that ''Y-DNA I+mtDNA U'' were not the first people in Europe and I fail to see how anyone would claim the opposit given what we know about pre-C-V20's very early divergence, in a time when haplogroup IJK itself hadn't yet bifurcated or just did so.