search this blog

Friday, October 3, 2014

Scratch the North Caucasus


I've just spotted a few interesting extras in the final draft of Lazaridis et al. that appeared in Nature last month, including this quote from page 126 of the freely available supp info:

The finding of high ANE ancestry in the North Caucasus might suggest that the Caucasus is a potential source of this type of ancestry in Europe. However, when we try to fit present-day Europeans as a 3-way mixture of a North Caucasian population+EEF+WHG in the structure of Fig. S14.20 this model is successful for only 5 populations (Bergamo, Bulgarian, Italian_South, Spanish_North, Tuscan using Lezgins as a sister group to the admixing population). Admixture from the Caucasus would need to be substantial to account for observed ANE levels in Europe (e.g., for a European population with ~15% ANE ancestry, almost half of its ancestry must come from a Lezgin-like population with ~29% ANE ancestry; this would account for the ANE ancestry but would greatly dilute its WHG-related ancestry, and yet present-day Europeans have increased affinity to WHG in Extended Data Fig. 4 relative to Stuttgart).

This was rather obvious anyway, but I know that there are a lot of people online who cherish the notion that Europe was invaded in a big way by groups from the Caucasus and/or Anatolia during the Bronze Age, and I'm guessing this paragraph was a response to the comments that the authors received from these people during the public review process.

Indeed, the updated supp info also has a couple of new Principal Component Analyses (PCA) of West Eurasian populations, with which Lazaridis et al. underline the point that most Europeans and Near Easterners form "two discontinuous clines" in such analyses (pages 76-80). I could be wrong, but the impression I get is that they're again communicating how very improbable it is for most Europeans to harbor any Near Eastern and Caucasian admixture that post dates the Neolithic transition.

This of course leaves pre-Turkic far Eastern Europe, Western Siberia and/or Central Asia as the source(s) of the ANE-rich population movements that apparently had such a profound impact on most of the European gene pool after the final Neolithic.

Citation...

Lazaridis et al., Ancient human genomes suggest three ancestral populations for present-day Europeans, Nature, 513, 409–413 (18 September 2014), doi:10.1038/nature13673

See also...

Coming soon: genome-wide data from more than forty 3-9K year-old humans from the ancient Russian steppe

Corded Ware Culture linked to the spread of ANE across Europe

120 comments:

Helgenes50 said...

If we take the West Asian and South Asian (K13) as a reference to the Indo European contribution, it does not represent a big percentage, at least in North West Europe. Since this is the only missing components before the age of metals.

Davidski said...

It seems to me as if Laz et al. are convinced that the Proto-IEs who moved into Europe weren't overly Near Eastern or Caucasian, but mostly ANE and WHG. Keep in mind also that when this last draft was being written up, at least some of the Samara and Corded Ware samples would've been sequenced by then.

Helgenes50 said...

So, the West Asian or Caucasian, as not present before the age of metals, would have come with PIE but would be only a part of them

Matt said...

If we take present day North Caucasians as the limit of ENF+ANE populations (which is sensible at least for the time being) and require all solutions to contain EEF, then it seems like an extra WHG+ANE population is the most parsimonious.

Although, other models that might work

- Three way mix of ENF+ANE in proportions beyond North Caucasians with WHG and EEF

- Two way mix of WHG and an ENF+ANE in proportions beyond North Caucasians

At the moment, we're treating the presence of EEF as the constraining variable, as is reasonably sensible as to do so would call for less population replacement - whatever fits have to contain EEF, so this constrains the upper limit of how much more ENF comes in with ANE, which is why we get EEF+"Hunter"(WHG+ANE), rather than WHG+(ENF+ANE).

But then if Samarra turn out mostly more like ENF+ANE* (or ENF+WHG+ANE in some proportion that excludes being a hunter+EEF mix), we might end up looking at models which are some form of two or three way admixture, possibly even without EEF in some parts of Europe (and the most parsimonious model might change).

(I treat this as pretty low probability that Samarra will turn out like this, but there are some parts of the ANE influenced world where a mix of ENF+Hunter (with WHG+ANE as fits Europe) might not work so well).

*we might already be able to exclude this possibility based on haploid / uniparental work, I dunno.

Balaji said...

WHG in the Near East and the Caucasus is O% though there is some UHG in these regions. The Near East and the Caucasus as well as Europe experienced an influx of ANE in the last few thousand years. If the same source contributed ANE both to Europe and the Caucasus, this source could not have had any WHG. By any reasonable definition of European this source cannot be called European.

barakobama said...

This is miss leading. If you made Europeans a mixture of Mari, EEF, and WHG it wouldn't make much sense either. We don't know if modern Caucasians are a good enough proxy for bronze age ones.

This is also a problem for your PIE zombie. Because it was pretty much a Caucasian with about 30% extra ANE. Modern Europeans have much more WHG than Neolithic western European farmers, so if Indo Europeans were the main changers of Europe after the Neolithic they all must of been mostly ANE and WHG.

barakobama said...

"By any reasonable definition of European this source cannot be called European."

They were who they were, no one is trying to put them into definitions which can generally be used today(Sardinians lack ANE, Greeks have post-Neolithic near eastern ancestry, etc. European is a general definition not a clean cut definition). Let's not continue past were PIE more European or south-west Asian battle, which is centered around race-competition, is agenda-driven, and makes people biased.

Balaji said...

What's in a name? that which we call a rose
By any other name would smell as sweet

My point is that ANE in the Caucasus has to be accounted at the same time as the ANE in Europe. This requires WHG in the source population to be 0%.

Davidski said...

Mari are Uralic-speakers with a high level of ENA, which was basically missing from the middle Volga region during the Copper Age. The nearby Chuvash are Turks, again with ENA.

ZeGrammarNazi said...

"K=9 reveals a split within West Eurasia, with one component maximized in Loschbour and Northern
European contemporary populations (both NE and NW) and one maximized in BedouinB. From the
ancient samples, only Stuttgart shows mixed membership in these two components, consistent with
the hypothesis that Early European Farmers represented a mixture of West European Hunter
Gatherers and Near Eastern farmers."

That quote is from the Laz supplemental data. Looks like you were on to something when you claimed that the Bedouin were the best proxy for ancient Middle Easterners.

Davidski said...

Balaji,

There's actually no evidence in Laz et al. that Lezgins don't have any WHG ancestry. There's only evidence that they don't have more WHG than Stuttgart. This isn't surprising, because Stuttgart is a Neolithic European, while Lezgins are West Asians.

So it's possible that Lezgins acquired their Near Eastern ancestry from a farming population that had less WHG-related ancestry than Stuttgart, or none at all, and then acquired some WHG-related ancestry along with ANE from the steppe.

The fact that this WHG-related ancestry was not from Western Europe, and might be better termed UHG, doesn't matter.

ZeGrammarNazi,

Yes, I saw that. Also, check out what they say about the K=15 run. I suspect that confirms my claim that the Kalash have a shit load of ANE.

Tesmos said...

How numerous was the population in Far Eastern Europe/Western Siberia? Was there a overpopulation in that region?

Matt said...

Another thing I would say which occurs to me is that while this comment in the paper is fine, but when it does say

"e.g., for a European population with ~15% ANE ancestry, almost half of its ancestry must come from a Lezgin-like population with ~29% ANE ancestry; this would account for the ANE ancestry but would greatly dilute its WHG-related ancestry, and yet present-day Europeans have increased affinity to WHG in Extended Data Fig. 4 relative to Stuttgart"

This does seem to be assuming that a North Caucasian (or just beyond) population with 71% Near East, 29% ANE is mixing with Loschbour or Stuttgart, who differ yet both have 0% ANE.

What about if it mixes with Motala12, who have 81% WHG and 19% ANE? And Motala like populations seem plasuibly the typical thing for ancient Eastern Europe.

A 50:50 Motala12-North Caucasus mix would be 35.5% Near East, 40.5% WHG and 24% ANE. A 75:25 Motala12-North Caucasus mix would be 17.75% Near East, 60.75% WHG and 21.5% ANE.

Once we've got this Motala12-North Caucasus mix, add some back and forth with Bell Beakers who were a little more like straight ANE free EEF (perhaps with a little WHG and ANE spice?) and maybe that's modern Europe....

It doesn't feel like the simplest model (although it may have the advantage not needing any ghosts), but perhaps the archaeology and more ancient genetics will come to stand with it.

barakobama said...

"It doesn't feel like the simplest model (although it may have the advantage not needing any ghosts), but perhaps the archaeology and more ancient genetics will come to stand with it."

Reich would probably same something if copper age Samara people were just like modern people in the north Caucasus. If PIE was Caucasus-like why would Laz authors deny that theory while probably having Samara and Corded ware genomes? We'll probably learn next week if Corded ware people had lezgin-like ancestry.

It's an interesting idea that Indo Europeans were Caucasus-like, but there are alot of holes in that theory. A Motala12+Lezgin model only roughly works for northeast Europeans, plus it's ignoring Neolithic farmers.

Both west Asians and Europeans mostly have ancient near eastern mtDNA, but are mostly apart of very separate clades. Typical-west Asian clades are very rare in Europe, and Europeans are mostly in the same clades found in Neolithic Europeans. This goes along with the idea with little gene-flow between west Asia and Europe after the Neolithic. Also, Yamna mtDNA did not show affinity to west Asia, but we can't make many strong conclusions about their mtDNA because of low-quality results.

Matt said...

It's an interesting idea that Indo Europeans were Caucasus-like, but there are alot of holes in that theory. A Motala12+Lezgin model only roughly works for northeast Europeans, plus it's ignoring Neolithic farmers.

Yeah, to be clear if this worked at all, it would be -

Corded = Motala + North Caucausus

then Europe = Corded + EEF (more or less)

with later phenomena like Beaker, etc. being West->East and East->West hybrids of Corded+EEF that mainly served to homogenise North Europe
So a radiation of a Motala-like people across Europe has to happen in some form, just they may have ENF from a North Caucasus-like source first (this wouldn't really lower their ANE, only their WHG). The paper says it can't exclude the presence of EEF like elements in whatever culture brought ANE to Europe, and a North Caucasus+Motala like fusion population seems like hybrid that would be plausible.

I'm definitely not an Indo-European expert, but in additional to the fairly established pontic steppe origin and radiation Davidski talks about, I think there are some pieces of evidence like vocabulary sharing between Semitic and early Indo-European and the earliest splits in the Indo-European family (pre Corded Ware) being in Anatolia that will need to be reconciled with any model.

Intuitively, feels more right that Corded / Indo-European (tribal grain cultivating farmers with herding elements) should have some (diluted over generations, "introgressed") link to ENF in some way, more than some ultra Motala people who are just simple hunter gatherers leapfrogging up to a more "advanced" level of cultural development.

On what Reich has said, what we know so far about Samara so far is here -
http://eurogenes.blogspot.co.uk/2014/06/coming-soon-genome-wide-data-from-more.html

"Samples from the Samara region possess Ancient North Eurasian (ANE) admixture related to a recently published 24,000 year old Upper Paleolithic Siberian genome. This contrasts with both European agriculturalists and with European hunter-gatherers from Luxembourg and Iberia who had little such ancestry (Lazaridis et al. arXiv.org 2013). "

"Samara experienced major population turnovers over time: early samples (>6000 years) belong primarily to mtDNA haplogroups U4 and U5, typical of European hunter-gatherers but later ones include haplogroups W, H, T, I, K, J." These look like West Asian mtdna haplogroups, but the actual levels are not described here, nor is how much later this transition takes place.

Grey said...

"The finding of high ANE ancestry in the North Caucasus might suggest that the Caucasus is a potential source of this type of ancestry in Europe."

One random thought on that:

If ANE developed in particularly hostile environment in the north then they might have a few adaptations specifically suited to places that were harder to survive.

(As latitude and altitude have some of the same pressures.)

Then if even small numbers drifted into mountainous regions further south which were also particularly hostile then maybe those genes would be selected in place?

The sort of adaptations I'm thinking of would be for example an adaptation that allowed someone to digest an extra 1% calories out of their food or something that helped them keep 1% warmer - something that could turn 8% into 16% over time.

Mike Thomas said...

"This of course leaves pre-Turkic far Eastern Europe, Western Siberia and/or Central Asia as the source(s) of the ANE-rich population movements that apparently had such a profound impact on most of the European gene pool after the final Neolithic"

But wasn't this ANE component already present in some Mesolithic samples (eg Sweden)

ryukendo kendow said...

There is always going to be a level of uncertainty that not even ancient genomes can resolve, because there are too many variables.

There are two questions: 1) PIE contributed to both C+S.Asian and Europeans; who did they contribute more to? If they contributed significantly to both C+S.Asia and Europe, then that precludes a very high level of WHG in the source pop. If they contributed strongly to Europeans and little to S+C.Asians, then this precludes low WHG.

2) How greatly did they admix with the aboriginal population before they moved to further legs of their journey? If they admixed with WHG in E+N.Europe before continuing SW, i.e. Matt's scenario, this makes possible low WHG in PIE. If they admixed strongly with C.Asian cultures, e.g. BMAC, this makes possible high WHG in PIE. If both then everything is possible.

The samara may not answer this question with 100% clarity, because one can always assert that Samara is already admixed with later influences from whichever direction one wants, and we have to go back even earlier. The only thing one can say is that some X is the most likely hypothesis.

My own reasoning is that Basal Eurasian/Agri influences in PIE is a foregone conclusion. Pastoralism has always grown out from the edges of agricultural/semi-agri areas, instead of arising suddenly in HG ones. Then there is spread of metallurgy as well, also from ME. Also c.f. mythology.

We know for sure that there is ANE in PIE.

This leaves WHG. Because R1a in C+S Asia shows signs of recent expansion, the demographic impact of PIE in C+S Asia cannot be that low. This means that whatever WHG PIE carried must be on the low side.

My two cents.

Grey said...

Isn't there the possibility that PIE displaced neighboring populations to the west creating a wave effect with the actual IE wave as the *last* wave?

Mike Thomas said...

Ryukendo "PIE contributed to both C+S.Asian and Europeans; who did they contribute more to? ...etc"

Sure. but that's only if you buy into the idea that there actually existed "proto-Indo-Europeans" as a culturally and genetically (more or less) homogeneous block of people from which later speakers of IE desced from , at least in part.

I, for one, do not because of the fundamental flaws of the models of sweeping expansions from a single focal point.

Davidski said...

ANE was indeed present in Scandinavia, and probably east of Scandinavia in increasing frequencies, before the Copper Age. But the ANE present in most of Europe today didn't come from Scandinavian hunter-gatherers.

That's because if it did, then Europeans would now be almost indistinguishable from Scandinavian hunter-gatherers, who carried a ratio of ANE to WHG of up to 19%. This is obviously not the case, because present-day Europeans carry much higher ratios of ANE to WHG, particularly those speaking Indo-European languages. The French, for example, have a ratio of ANE to WHG of almost 50%.

Using the currently available pre-Bronze Age ancient European genomes, there's really no way we can get to a ratio of ANE of almost 50% with models that don't include a significant migration deep into Europe from somewhere east of Scandinavia.

Keep in mind also that late Neolithic Pitted Ware Scandinavian hunter-gatherers, who carried around 15% of ANE based on Ajvide58, lived by foraging and hunting seals, so they couldn't have been very numerous, and it's very unlikely that they were able to spread this subsistence package with them to Central Europe and be successful enough with it to dominate the region. Not only that, but it's clear that they were mostly replaced in Scandinavia, because their Y-chromosome lineages are now extinct there.

Also, there seems to have been something different about the ANE that arrived in much of Europe after the Neolithic. I refer you to page 67 of the supp info linked to above, and the paragraph that deals with the K=15 ADMIXTURE run.

In regards to the Proto-Indo-Europeans not coming from a relatively compact area, from what I've read, linguistic data supports the case that they did. They probably lived between the proto-Uralics and proto-Kartvelians, which is the Eastern European steppe/forest steppe, the only area which lacks non-Indo-European substrates. More info here...

http://eurogenes.blogspot.com.au/2013/10/tracing-indo-europeans-conference-videos.html

As for the shifting structure of the mtDNA gene pool in the Samara region, for now, I'd say that it can be explained by increasingly close contacts from the late Neolithic onwards between the Neolithic farmers of what is now Ukraine and the descendants of far Eastern European hunter-gatherers, who at some point became mobile pastoralists. You can read about these contacts in the second part of this paper. Scroll down to the part titled "A Case Study".

https://www.academia.edu/3535031/Migration_in_archaeology_the_baby_and_the_bathwater

The shift from foraging to mobile pastoralism in far Eastern Europe also gives us a plausible model how a population with a high ratio of ANE might be able to move fairly quickly deep into Europe and become successful there. That's because they'd have a novel subsistence package and, very likely, new social structure to help them spread out across wide areas and dominate these newly acquired territories, unlike the seal-hunting Scandinavian hunter-gatherers, who basically moved wherever their seals decided to go.

Mike Thomas said...

Several points:

* Why do you need to assume that it was only *Scandinavian* HGs that had pre-Neolithic ANE. I only used it as an example to illustrate that ANE already existed, probably,throughout Europe (at least central & north). Given the mid-to-late Neolithic bust, the rise in ANE in Europeans could be due to some "mesolithic revival" (to put it simplistically/ anarchonistically). And I wasnt at all insinuating that the ANE in present day Europeans is accounted for by some Scandinavian expansion.


*
In regards to the Proto-Indo-Europeans not coming from a relatively compact area, from what I've read, linguistic data supports the case that they did. They probably lived between the proto-Uralics and proto-Kartvelians, which is the Eastern European steppe/forest steppe, the only area which lacks non-Indo-European substrates. More info here..."

Yes, PIE *language* once came from a confined area, sure. But this does not mean, as i said above, there was a culturally and genetically distinct PIE *people*. Moreover, the lack of non IE interference in the EE steppe/ forest -steppe is illusory. Ie that is only so because we have no records of the region until historic times, (4000 years after its supposed origin). The apparent lack of substrate, and supposed purity or baltic and Slavic languages, is itself dependent on the Kurgan model which it aims to prove, and lack of any idea of pre-slavic lanagues of the region. Anyhow, some scholars (eg Kaufmann) supposed a Uralic substratum in *all* Slavic languages, and not just the northern ones, and Hening Andersen argues that there is proof of pre-balto-Slavic lanagues preserved in balto-Slavic. (ie in so-called discrepant dorsals)

**As for the shifting structure of the mtDNA gene pool in the Samara region, for now, I'd say that it can be explained by increasingly close contacts from the late Neolithic onwards between the Neolithic farmers of what is now Ukraine and the descendants of far Eastern European hunter-gatherers, who at some point became mobile pastoralists. You can read about these contacts in the second part of this paper. Scroll down to the part titled "A Case Study"."

- Ive not yet read that study, but you'll find that pastoralism was very much conditioned by contacts with Balkano-Tripoljan agricultural sites, which already figured nascent pastroalism in their economy. In fact, the very reason for the growth of the Yamnaya culture was colonization from the Tropolye region (!) yes, the very opposite of Anthony's deeply flawed Kurgan mode.

Mike Thomas said...

'Not only that, but it's clear that they were mostly replaced in Scandinavia, because their Y-chromosome lineages are now extinct there."

"Replaced" is a relative term. You cannot rely on Y chromosomes alone. A populations entire Y chropmosome profile can become 'replaced' without them being so in reality, even in the absence of male-mediated invasion, competition & extinction. All it takes is time, say 1000 years and some bad harvests. ...

Davidski said...

I'm assuming that only Scandinavian (and Eastern European) hunter-gatherers carried ANE, because Loschbour from Mesolithic Luxembourg lacks ANE, and so does Gokhem2, the late Neolithic Swedish TRB farmer, who shows considerable hunter-gatherer admixture, and whose recent ancestors most likely arrived in Scandinavia from mainland North-Central Europe.

So we have several datapoints from across space and time in Europe, including from the very late Neolithic and even Chalcolithic (Oetzi the Iceman from the Tyrolean Alps), which show a ratio of ANE that is practically zero.

Like I said on my other blog, the ANE that suddenly appeared across Europe during the late stone age/early metal ages wasn't hiding in caves or up in trees, or just across the river. It had to have arrived there from some distance away, but it certainly didn't come from Scandinavia.

Balaji said...

Davidski,

You are right that Laz et al. only say that Lezgins and other Near Easterners have less WHG than Stuttgart. But you yourself have estimated the WHG/UHG in Lezgins as 15% with ANE of 27%.

http://bga101.blogspot.com/2014/09/eurogenes-ane-k7.html

Let us suppose that the Lezgins acquired all their ANE and all their WHG/UHG from the same source and that this source only had ANE and WHG/UHG. This will give us an upper limit for the WHG in this source population. This number is 36%. Now suppose we assume that the source population had BEA and ANE in equal amounts as well as WHG/UHG, then the percentage of WHG/UHG will be 22%. As Rykendow Kendow has pointed out the source population could only have expanded into the territory of others if it had advanced technology such as metallurgy implying it had considerable BEA. Therefore 22% as the upper limit for WHG/UHG into the ANE source population of Lezgins is more reasonable.

Davidski said...

Well, the steppe pastoralists began expanding without metallurgy. For instance, the horse bits and boar tusk ornaments found among the late Neolithic remains from around the middle Volga aren't made of metal, but later they show up elsewhere, like in Central Asia and the Balkans, looking almost identical, but they're either made entirely of bronze or iron, or like the board tusks, studded with copper beads.

Davidski said...

Actually, let me correct that. Cheek pieces (designed to hold horse bits) made of bone are found at early Bronze Age sites on the European steppe and near the Urals, and then later very similar cheek pieces made of bronze appear in Central Asia.

Mike Thomas said...

Blogger Mike Thomas said...
"Actually, let me correct that. Cheek pieces (designed to hold horse bits) made of bone are found at early Bronze Age sites on the European steppe and near the Urals, and then later very similar cheek pieces made of bronze appear in Central Asia"

The entire horse arguement is highly disputed by all apart from those with a vested interest in Kurganist arguements, and the latter people aren't even specialists in Bronze Age archaeology but 'pop' historians who write for the half-informed masses.

Equestrianism only began much much later, on the outskirts of the Parthian Empire c. 800s BC.

So how did your non-existent horse people invade Europe ? They were poor, scant, and resorted to **eating** rather than riding their horses into battle. Hmm. maybe they had nuclear weapons ?

Davidski said...

There's no dispute that the earliest horse bits and cheek pieces made of bone are found on the steppe near the Urals.

Their use then spreads both west and east, to the Carpathian Basin and Central Asia, respectively, and later metal copies are found in Central and West Asia.

So the horse people did exist. If we assume that they didn't know how to ride horses, which I really don't believe, then obviously they used the cheek pieces and bits for something else, like pulling chariots and wagons.

I actually linked to a recent lecture on the horse issue above, but you seemed to have missed it, so here's the direct link...

https://www.youtube.com/watch?v=HliaR2Ep24s&list=PLAXoDomeFLX90fTHi0W8lYBtEoZHSBH2i&index=6

Whatever you might think of David Anthony, calling him a pop writer is too much, and on top of that simply ignoring what he has to say is also pretty stupid, especially as his long standing arguments are now being backed up by ancient DNA.

Mike Thomas said...

Sure, horses were beginning to be domesticated for traction and secondary products. They were not ridden into conquering vast land masses. Nor does his theory that the act of domestication spread from the Pontic to Botai hold, as this has been recently proven to have been a local evolution. Finally, there area major flawes of his chronology. Eg Anafiesvo culture *predates* the Yamnaya culture.

Even worse, his treatment of the supposed invasions into the Danube valley are laughable. Every Balkan Bronze Age specialist has rejected Anthony's theories.

Matt said...

@Davidski Using the currently available pre-Bronze Age ancient European genomes, there's really no way we can get to a ratio of ANE of almost 50% with models that don't include a significant migration deep into Europe from somewhere east of Scandinavia.

Sure, I wouldn't claim Scandanavian / Baltic hunter gatherers are plausible on their own. They'd have to mix with EEF to get Neolithic levels up, and this would reduce ANE to insufficient levels (as in the paper).

Also North Caucasians wouldn't be sufficient, as they would lack enough WHG mixing with EEF like Stuttgart and not have enough ANE mixing with WHG like Loschbour.

What I'm proposing is a hybrid between North Caucasians and Scandanavian / Baltic hunter gatherers (approximately) could more or less have the right level of WHG to ANE (mix the 0:30 ratio of North Caucasians with the 1:5 ratio of Scandinavian / Baltic hunter gatherers and you could get the right level), and crucially, as it already has ENF, you wouldn't need as much mixture from EEF (who we don't see existing in some areas of Europe).

Just an idea anyway. Anthony's description of the early PIE as “tribal farmers who cultivated grain, herded sheep and cattle, collected honey from honeybees, drove wagons, made wool or felt textiles, plowed fields at least occasionally or knew people who did, sacrificed sheep, cattle and horses to a troublesome array of sky gods, and fully expected the gods to reciprocate the favor” sounds a little like it could be for a stone age North Caucasus population who picked up Scandanavian HG type ancestry along the way. And Laziridis is agnostic on whether the North Caucasus / West Asians picked up ANE during the initial radiation of ENF farmers, or later (seems to me that with the high level of language diversity there, they probably weren't affected by later movements from the steppe, but it's possible they were).

Davidski said...

My model currently involves an almost purely ANE population with high frequencies of Y-haplogroup R mixing with farm girls of Balkan origin in Ukraine during the Neolithic, and then after the Neolithic pushing west and mixing with populations more or less similar to Gokhem2, with varying levels of WHG and even some ANE.

I don't see direct gene flow from the Caucasus being very important in this, although the hypothesized ANE/EEF mixed groups in Ukraine and southern Russia would probably be very similar to my Lezgin/MA-1/AG-2 composite.

Chad Rohlfsen said...

I don't think this really changes what I was saying. I think that R1b was an ANE lineage in Northern West Asia and the Southern Steppes. It expanded with the arrival of agriculture and pastoralism, eventually moving into Europe, via the Black Sea coast, and expanding from there. We can't compare current Caucasus populations with what was probably there during the Neolithic or Early Metal Ages. I still think R1b will be dominant in Khavalynsk.

ryukendo kendow said...

'... ANE population with high frequencies of Y-haplogroup R mixing with farm girls of Balkan origin in Ukraine...'

'I don't see direct gene flow from the Caucasus being very important in this'

Why?

Davidski said...

What evidence is there of gene flow from the Caucasus to the East European plain?

Both uniparental and genome-wide markers show links between Europe and the Near East via the Mediterranean and the Balkans, but there's a major gap between Eastern Europe and the Caucasus, and the only reason it's not even bigger is because North Caucasians are shifted northeast towards the Volga-Ural region.

Mike Thomas said...

YEs, several studies have shown the suprising gap between Caucasus and adjacent Russians and Ukrainians, etc. But one could argue its a 'chicken or egg' arguement. Is it because there has been turnover in the Caucasus itself (often thought to be a 'relict area') or the plains of southern Russia or Ukraine (a common thoroughfare).

Whilst it could be the latter, Id be weary of automoatically thinking of the Caucasus as a 'relict' region. Even remote, mountainous regions are liable to migrations, population turnover, etc.

Davidski said...

The Caucasus isn't any sort of a genetic refuge.

The point is that there is no direct genetic relationship between Europe and the Caucasus. The relationship that does exist is indirect and was mediated via two different groups:

1) Near Eastern Neolithic farmers

2) ANE steppe nomads

That's why there's now an almighty gap between Eastern Europe and the Caucasus on West Eurasian PCA, but populations from both regions form parallel clines that stretch from west to east.

Mike Thomas said...

True ; but given the well known cultural and historical toes between the Caucasus and Black Sea steppe, the large gap between current populations must be explained by more recent events . Eg recent slavic expansions in Ukraine and Southern Russia. Also poor sampling of non-Slavic Russians eg tartars etc

Mike Thomas said...

*ties*

Davidski said...

Tatars and Finns from the middle Volga forest steppe are part of the European cline, although Tatars are shifted southeast, along with Chuvashs and Maris, probably due to inflated East Eurasian admixture. You can see that here...

https://drive.google.com/file/d/0B9o3EYTdM8lQV1VNT2dzSUl0RDQ/view?usp=sharing

Chad Rohlfsen said...

Population turnover in the Southern Steppes is the reason for the gap. Slavic expansion... Russians relocating people from the region.. plus Turkic tribes coming in. The turnover was on the North side of the Caucasus.

Davidski said...

Typically Caucasian groups may have lived further north than they do today before the Turkic and Slavic expansions, but that's not the point.

The point is that there's no evidence of a genetically intermediate population between Europeans and Caucasians inhabiting the western steppe. Whatever data we might be looking at, it's always the same outcome.

The closest proxies we have for people who lived in the North Caucasus and along the middle Volga during prehistoric times are Abkhasians and Erzya, respectively. They're actually placed parallel to each other on the PCA of West Eurasia linked to above, which is a result that fits geography, but they're also as different from each other as all other typically North Caucasian and Eastern European populations.

Mike Thomas said...

"The point is that there's no evidence of a genetically intermediate population between Europeans and Caucasians inhabiting the western steppe. Whatever data we might be looking at, it's always the same outcome."

I'd bet ancient DNA from anywhere between the Dnieper & Kuban from .c 200BC - 200 AD would be such an intermediate population

ryukendo kendow said...

@ Davidski
'The point is that there's no evidence of a genetically intermediate population between Europeans and Caucasians inhabiting the western steppe.'

If indeed it is true that PIE carried low WHG, then such a population would not exist.

There is, for example, no population that is intermediate between native americans and Europeans, though it is possible to construct a 'detour' through ANE. I would think that the situation is exactly the same for Caucasus and E.Euro.

There might be other ways to get a handle on the situation: 1) we can look for substructure within WHG, and whether the french differ from the basque in having East-Europe-related WHG vs, say, more La-brana or Loschour related WHG in Basque. We can also trace the affinities of UHG in Central Asians and compare. Maybe make Loschour, La-brana, Ajvide and Malta each their own separate component, and run ADMIXTURE to see what happens to Europeans+Central Asians, esp the latter if they score in the aWHG components.

2) We can strip Central asians of East+South Eurasian contribution, and see what we get. Autosomal is difficult, but I think a pca of mtDna with such stripping would also be informative.
De sarkissian in 2011 published a pca including rostov scythian and bronze age kazakhstan. The position of Central Asians in that pca makes me think that if influence from E.Asian and 'other' can be stripped from C.Asian, they would occupy precisely the necessary position.

http://digital.library.adelaide.edu.au/dspace/handle/2440/74221

barakobama said...

If Mesolithic people in Russia were ~100% ANE and copper age ones were ~45% EEF(like the zombie), why do modern east Europeans have the highest amount of WHG in the world? Take away east Asian ancestry of Finno-Urgic and Turkic Russians and they would be typical northeast Europeans. Their higher amount of ANE can at least partly be attributed to their Siberian ancestry.

Does this suggest a mass migration of WHG-heavy west Europeans into eastern Europeans? Or were there Neolithic farmers with big chunks of WHG as far east as Lithuania or Russia who mixed with Indo Europeans?

Shouldn't it be possible to slice an iron age Indo European from central Asia with modern Indo European+Turkic mutts? I would assume Sycthians or whoever were basically the same as copper age Russians despite some east Asian admixture. What about with Indians is their a good pre-Indo European proxy population of India?

ryukendo kendow, I like your ideas. WHG ancestry in west Asia, northwest Africa, and Europe can't be indistinguishable from each other. Same goes for ENF(SW Asian vs Mediterranean). Loschbour's WHG is probably a mix of Moatal12's and La Brana-1's WHG. Motala12's ANE should be a good proxy for PIE's ANE. Comparing it with ANE in modern populations might be able to distinguish regional ANE and native ANE from IE ANE. Making splits between differnt HGs and modern WHG will probably be able to show whether IEs had a lot of WHG or not(eg. if Irish WHG looks like mix of Loschbour and east European WHG).

Matt said...

I think the model where ANE comes in from a relatively high ANE:WHG population with low ENF mixture is totally a plausible model from what we know at the moment. My point is really only, considering only Europeans and autosomal genetics, there are two possibilities for how we can get the WHG:ANE ratios right, and PCA to fit while retaining the levels we need–

- a population which is a theoretical (but reasonable) exaggeration of the Losch-> Motala trend and Middle East->North Caucasus trends (origin population further East), then admixture between these and more and less WHG admixed ENF (more in Europe, less in West Asia).

- ENF picks up ANE during its expansion into West Asia / Caucasus, then direct admixture between Motala+North Caucasus (origin population further South), which also brings in ENF establishes ancient post-Neolithic far East Europeans, who spread into Europe.

Graphically I can show the difference between the two on the following badly modified annotated graph screencapture from the supplement's figure S10.8 -

http://imgur.com/tYQmfWC

The black cline is a theoretical North Caucasus-Motala cline, as could've been created by an ancient North Caucasus like population mixing with increasingly WHG yet equally ANE populations to their North (e.g. if they admixed on the East European plain). It is a model where the two discontinuous clines once were continuous.

One reason (perhaps the only reason) to favor this model is that, if we're postulating a cline which was rubbed out, this is the shortest such cline possible (and it fits geography).

The red clines are a continuation of the two diverging clines. Their convergence point would end up quite close to MA-1.

In either case, either clines would've been "rubbed out" by later Turkic population movements...

(the blue clines are North Caucasus to EEF and WHG, so you can see why the space between them wouldn't fit, and why the space between black and dark blue would).

(Eurogenes graph version - http://imgur.com/VCgAbJ2)

Autosomally, Europeans mostly all can be fit fine as a fit between North Caucasus, Motala12 and Stuttgart. The all fit within the triangle created by North Caucasus, Stuttgart and Motala, and where they don't, there's can't be accounted for by the hunter+EEF model either (due to East Eurasian or Middle East admixture).

It doesn't necessarily mean it had to happen this way, but I'd have thought that the authors of the paper would mention the possibility and it's odd that they haven't. In fact its hard to understand why they constrained the models they discussed to only use Western European hunter gatherers, EEF and North Caucasus and not test Scandinavian hunter gatherers, EEF and North Caucasus.

adna should soon tell us which is the more optimal.

Of course, there are reasons why Davidski's model makes sense – the existence of higher ANE:WHG than the European maximum is present in populations in the present day with East Eurasian mixture, and also the spread of R1a, which isn't present in the Caucasus in this way and which we have no current sign of in the Scandanavian HGs.

The approach as suggested by RK to extract all East Eurasian contributions from a really dense set of Central Asian references and see which cline area becomes more filled in, might be pretty useful.

Davidski said...

Bazza,

Believe it or not, present-day far Eastern Europeans, like the Russians and Finns near the Volga, have a lot of Baltic and even Western European ancestry. They acquired this from various people who expanded east from around what is now Poland, like the prehistoric Balts, Slavs, Germanics, and then Lithuanians, Poles, Germans, even some Dutch and French, etc.

There's been a lot of WHG around the Baltic since the end of the Ice Age, and I reckon the Bell Beakers and then Celts brought more Western European admixture during the Bronze Age. You can see this in the uniparental markers too, with some Atlantic-specific Y-DNA and mtDNA all over Poland and deep into Russia.

Tesmos said...

Davidski

Wow that's really interesting. Would you say that most Northern Europeans have admixture from every single Indo-European group like the Germanics/prehistoric Balts/Slavs and Celts aswell or is that an another story?

Alberto said...

When talking about ANE the keyword is ANCIENT. Therefore it cannot be directly related to modern populations.

Modern Europeans are based on 3 different populations (using the k7b denominations):

- Atlantic_Baltic: Basically the Paleo/Mesolithic population that inhabited all of Europe, including Kazakhstan. It could be divided in 2 subgroups: Eastern (from East Ukraine to Kazakhstan) and Western (all the rest of Europe.
- Southern: From Northwest Africa to the Middle East, probably to somewhere in modern day Iran. Could also be divided in 2 subgroups: North African and Near/Eastern.
- West_Asian: That were originally in Central Asia, just south of the Eastern European group in Kazakhstan (modern day Uzbekistan, Turkmenistan, Kyrgyzstan, Tajikistan,...).

Then there was a 4th Caucasoid group, quite different from the other 3, the South Asian, in modern day India, Pakistan, parts of Iran, maybe Afghanistan...

Early European Farmers migrated from the Near/Middle East into Europe, and had no West Asian admixture. However, later during the Neolithic, the West Asians moved west and south, Mixing with all the Southern population 'till Egypt and all the South Asian population 'till North India. Nowadays, all Near/Middle Eastern populations have a strong West Asian admixture, unlike EEF (or modern Sardinians), who supposedly came from there.

These West Asians must have been relatively north originally, since they had fair skin and a good percentage of fair eyes (though no blond hair). And quite broad skulls. Nowadays, they are best represented by Caucasus populations, but also by Turkmen (except for their East Asian/Turkic admixture). Around Pakistan they mixed with the very different South Asians, so even if they have a high percentage of West Asian, they don't look so much as the original ones.

Eastern Europeans (R1, or at least R1a) probably had some of the West Asian admixture originally, since both populations were bordering each other somewhere in south Kazakhstan.

Alberto said...

The question of course is when exactly did West Asians move from Central Asia to West ans South Asia.

Probably it was a continuous migration for a couple millennia (around the mid-to-late Neolithic), not an abrupt invasion. Maybe they were the reason why some Near/Middle Easterners migrated to Europe in the first place (pushed by West Asians). Difficult to know with so scarce data about these West Asians.

Davidski said...

Alberto,

The West Asian cluster is not a population. It's a modern composite of allele frequencies from ancient populations like Neolithic farmers from the Near East and ANE.

So you can't track ancient population movements with modern Admixture clusters, which are artifacts of sampling, recent isolation, drift and so on.


Tesmos,

I'd say everyone in Europe has admixture from everyone else, just in different quantities.

Alberto said...

Davidski,

And what is your explanation for EEF (or modern Sardinians, even Basques) having a lot of "Southern" admixture but ~0% West Asian admixture?

Again, as in k7b spreadsheet, Assyrians have 40% Sounthern and 50% West Asian. Lebanese have 44% and 39% respectively. If EEF were a mixture of these populations and European WHG, why did the West Asian admixture disappear in the mix???

Unless EEF came from North West Africa, which is unlikely, it's obvious that during the early to mid Neolithic the Near Eastern populations had no West Asian admixture (also see this study: http://arxiv.org/abs/1312.6639v2 - page 67 of the PDF, when commenting on K=15 results, where they mention this same thing).

The problem is ANE, not West Asian. ANE is a red herring because it's so ancient that it affects all populations from Amerindians to South Asians and Europeans. West Asian is much more modern and traceable. Where did West Asian come from? I say Central Asia. I could be wrong, but there are not many other choices...

Davidski said...

The red herring is the West Asian cluster, because it's the result of recent drift in the Caucasus.

It doesn't represent a pure population that lived somewhere in the past and then expanded for some reason. All it represents is a set of allele frequencies that can appear in populations that have no direct relationship to each other.

It's like the Ashkenazi cluster in some Admixture runs, which shows up all the way from Iran to Morocco, simply because Jews have ancestry from West Asia and North Africa, and not because Ashkenazi Jews expanded across this area.

Alberto said...

West Asian could be a red herring, yes. But then we'd need a better theory, not a worse one (the ANE is definitely worse, it explains nothing).

People around Pakistan are a mix of South Indians and at least another population (that was closer to Europeans). If this other population was indeed European (the Eastern branch), then Pakistanis would have Atlantic_Baltic admixture to a high degree. Instead they have mostly West Asian.

A recent drift in the Caucasus? Ok, and according to this the Caucasus population was mostly "Southern", that's why most "Southern" populations show strong affinity to it? But then... why EEF (or Sardinians) don't show this affinity? Why in Pakistan they show such strong affinity, but have no "Southern" admixture (and they obviously are a mix of South Asians and another population, but if this other population was European why not being Atlantic_Baltic instead of West Asian?

We need a better theory. ANE and "recent drift in the Caucasus" are too vague. A 3rd population (and Central Asia seems to be in the right place) sounds more likely to me.

As usual, time (and ancient DNA) will tell.

Davidski said...

Admixture doesn't provide a formal mixture test. If you over-interpret the output without cross checking it with data from other analyses, you risk coming up with ridiculous theories, like the one about the Bronze Age Indo-European invasion of Europe from West Asia, which was popular online a couple of years ago.

The West Asian cluster is linked in some way to the ANE that entered most of Europe after the Neolithic, just like the Kalash cluster in the K=15 run in the Lazaridis et al. study (see page 67 of the supp info PDF).

But without more ancient genomes from across space and time, it's impossible to say what we're really seeing, because groups like the Abkhazians, Dagestanis, Brahui and Kalash, who are all very good at triggering Admixture clusters, are essentially stuck high up in the mountains and genetically highly drifted. They're certainly not the sources of any invasions or last surviving enclaves of pure ancient populations.

It sounds like the Corded Ware and Samara Valley genomes might provide some answers. Let's hope they do.

Alberto said...

Yes, I completely agree with all that you say above.

But that's exactly why I dislike this whole ANE thing. It's vague, based on 24.000 y.o. genome (mostly), ancestor to completely separate populations (that genome itself appears as a mixture -ancestor- of East European, South Asian and Amerindian, so no need to say how useful it is for establishing the migrations during the Neolithic).

That comment on page 67 about k=15 is the quite revealing, but linking it to this vague ANE is a pity (it actually is related, but that relation is far and just a broad hint).

Samara genomes should not be very surprising, I think. Older ones should give us the East European type of the Atlantic_Baltic group (as Loschbour or La BraƱa gave the Western type). Newer ones might have some small admixture of other more southern populations (or South Eastern).

If we could have a similar studies with populations from Mesopotamia, Indus Valley, even Central Asia... THAT would be much more useful (but yes, difficult to do).

Grey said...

@Mike Thomas

"In fact, the very reason for the growth of the Yamnaya culture was colonization from the Tropolye region (!) yes, the very opposite of Anthony's deeply flawed Kurgan mode."

I think one of the things people are missing here is time spent on horseback.

Medieval peasants had horses - they walked behind then while plowing or drove them to pull carts.

Medieval knights had horses - they practiced fighting on horseback by riding to hunt every day.

On the steppe both peasants and aristos spent all day on horse back.

#

Expansion generally requires some kind of military advantage - with numbers being the decider if there is nothing specific.

A population from an environment where the food-getting requires spending all day on horse back will have an advantage over a population where the only people who can take the time to spend all day on horse back are a small warrior elite who have their food provided for them.

#

"Equestrianism only began much much later, on the outskirts of the Parthian Empire c. 800s BC."

There are videos all over youtube of people riding horses without a bridle at all or just using a rope around the horse's neck like this little girl.

https://www.youtube.com/watch?v=j2sYlGVB1_w

It seems possible to me that after domestication the less developed equestrianism was at any particular point in time - in terms of riding equipment - then the *bigger* the potential advantage there may have been to time spent in the saddle.

#

"Sure, horses were beginning to be domesticated for traction and secondary products. They were not ridden into conquering vast land masses."

There's a third option between nothing and mass conquering cavalry armies - raiding.

Take the depopulation of the Med. coast as a result of Moorish raiders for example (or imo the possible depopulation of the eastern coast of England as a result of Saxon raiders).

Early PIE raiding of Cucuteni type populations (for example) might be where they got some of that WHG.

So instead of a big invasion an early copper age expansion may have been

stage 1: raiding
stage 2: raided population move away
stage 3: expansion into now depopulated lands

with the chariot army type invasions only coming in the Bronze Age after the development of more cavalry / chariot technology.

This would imply the earlier copper age expansions into western Europe were mostly those displaced Cucuteni type populations perhaps with an IE warrior elite.

Mike Thomas said...

Grey thanks for your thoughtful response. But its seriously flawed:

"On the steppe both peasants and aristos spent all day on horse back."

->Speculation, and highly unlikely. At least, not in 3000 BC, or 2000BC, or even 1500 BC.

"It seems possible to me that after domestication the less developed equestrianism was at any particular point in time - in terms of riding equipment - then the *bigger* the potential advantage there may have been to time spent in the saddle."


-> Unlikely that unsaddled, undeveloped horse-riding was any advantage at all. Any early horseriding was for amusement and bravery. No military advantage conferred. As above, the first true equestrian groups were c. the Median empire- the 'true' Scythians of central Asia.

"There's a third option between nothing and mass conquering cavalry armies - raiding'.

->As above, for 'raiding' to be effective, youd need to actually be an effective equestrian. This happened much much later. Your examples of Moors are simply absurd, given that the Moors existed in known history, ie after equestrianism came about. So it is a useless, anchronistic analogy.

"So instead of a big invasion an early copper age expansion may have been
stage 1: raiding
stage 2: raided population move away
stage 3: expansion into now depopulated lands

with the chariot army type invasions only coming in the Bronze Age after the development of more cavalry / chariot technology.'
This would imply the earlier copper age expansions into western Europe were mostly those displaced Cucuteni type populations perhaps with an IE warrior elite."

-> DEmonstrably false. Firstly, the direction of Balkan collapse is wrong. The first centres to collapse were in northern Serbia, then Bulgaria, then Romania, then Cucuteni. This makes no sense if there was ""raiding" from the Yamnaya community, which didnt even ride horses, and was far less developed than that of the Balkan world. Secondly, it has actually been demonstrated, that they Yamnaya culture developed as the Cucuteni culture began to transform, becoming more extensive (c.f. intesive) in economy, with greater emphasis on agropastoralism, and greater mobility. So the 'Cucutenians' invaded (ie colonized) the steppe (!), complementing/ supplanting the pre-existing foragers/ hunters, not the other (wrong) way round.

The only cited proof of Yamnaya invasions into ROmania, Hungary etc is the presence of kurgans. However, this was a marker of male social elite status, and not of some IE invasion. Just like the spread of inhumation in later history was due to the adoption of shared Christian ideology and beliefs, and not due to some invasion of people from Jerusalem ! To rest your hat on this one piece of evidence is hopelessly reductionist. For example, in the kurgan graves in Hungary, there are items of Aegean origin. So according to your reasoning, some Aegean people also invaded Hungary. Lastly, the isotopic evidence from a recent study of an early Kurgan burial in the Carpathian plain showed the interred was born in nearby Transylvania, and not the Caspian steppe.

But this is beyond the point. I feel uninclined to debate intricacies of archaeology to people who hold a religious compulsion to believe a myth just because it is appealing to them on some personal, ideological or whatever basis. I am not saying that is the case with you, but you'd need to read some specialist works by Bronze Age archaeologists to believ what I am saying, and not the Encyclopedic, introductory -level stuff by Mallory and Anthony, for all their respectful contributions to the field.

ryukendo kendow said...

@ Davidski @ Matt

I thought a bit more about a potential ADMIXTURE run involving forcing ancient genomes into their own components.

One thing we could try is leaving other aDNA unforced, e.g. stuttgart and Oetzi. This would hopefully reveal something about their WHG.

Another thing to try is forcing both Motala and AG2, since both contain ANE, in addition to either Loschour or La Brana, and perhaps a lezgin as well. This might help to shed some light on the two scenarios Matt describes, especially if we see how e.g. Basque turn out vs French.

This can also illuminate whether or not WHG in Eastern Europe stretching to Siberia has western affinities.

ryukendo kendow said...

@ Davidski

On another note, why did you delete the article on AME?

Davidski said...

I was getting too many messages from people asking where they can find the calc files and why the test isn't at GEDmatch, even after I published the ANE 7 test.

Grey said...

@Mike Thomas

"Speculation, and highly unlikely. At least, not in 3000 BC, or 2000BC, or even 1500 BC."

There is evidence all over youtube of people riding horses without any equestrian equipment at all so a default position that a population who learned to hunt/herd horses didn't learn to ride them as soon as it was physically possible is just silly.

Band of herders/raiders != organized body of cavalry/chariotry.

.

"-> Unlikely that unsaddled, undeveloped horse-riding was any advantage at all. Any early horseriding was for amusement and bravery. No military advantage conferred."

This is nonsense.

You don't need to fight on horse back for riding to be an advantage. The advantage is *mobility* which is what would confer the raiding advantage.

mounted infantry != cavalry

.

->As above, for 'raiding' to be effective, youd need to actually be an effective equestrian."

No you don't, you'd just need better mobility. You ride to the attack, attack on foot and then ride away into the night.

.

"Your examples of Moors are simply absurd, given that the Moors existed in known history, ie after equestrianism came about. So it is a useless, anchronistic analogy."

You seem to have missed the reason why the Moors are potentially the perfect analogy i.e. their *sea* raids had a mobility advantage but the raid itself was fought on foot.

So instead of ride to raid, attack on foot, ride away it was sail to raid, attack on foot, sail away i.e. exactly the same advantage as I am suggesting for the steppe-sea.

.

"Firstly, the direction of Balkan collapse is wrong"

That might be a valid argument. I'll leave that to others to argue

Grey said...

I'll use a more modern analogy to see if it makes the point clearer.

You have two armies fighting. One side has small arms (including machine guns) but no trucks while the other side has the same type of small arms but also has trucks but the trucks can't be used in actual combat because they are too vulnerable to machine gun fire.

Does the side with the trucks still have a distinct military advantage despite that simply because of greater mobility outside combat?

The answer is duh.

barakobama said...

"I feel uninclined to debate intricacies of archaeology to people who hold a religious compulsion to believe a myth just because it is appealing to them on some personal, ideological or whatever basis."

Personally I'd like it if you and other knowledgeable people argue and simply comment about archaeology.

You're right people are more attracted to theories that are more appealing and simplistic, goes along with their ideology(Hitler would try to prove the ancient Greeks were German, lol), or are exciting. Because the modern world gives stero typical characteristics to ethnic groups(middle easterns are fanatics, Polacks are retards who are unable to change lightbulbs, etc) people try to put defining characteristics on ancient people(Greeks were wise philosophers, Romans were discipline, ambitious, and militaristic, etc.) because it makes history more interesting, ancestral-pride, etc.

While academics try to avoid this it seems they end up making the same mistake. While they try to get away from labeling ancient people with certain personalities, they create weak neutered ancient people.

I have no knowledge beyond a few reads here and there. And so far I am defiantly not convinced that most ancient societies had a weak and passive lower class and a small minority of snobby-ass nobility which fought their wars, etc.

I've read some pre-medival writings here and there and it seems in most Chiefdom societies(Portugal-Mongolia in Iron age, right?) there wasn't a 1% knight nobility or at least the societies were not as structured as Hindi India. When boys became men oftenly they were given weapons, and men carried weapons in everyday life. The impression I get is that they simply gathered the best men available for their armies(at least for the most part). They seem to not have had super-structured and legalistic governments like we do, had no currency, were very eglitarian and communist(like a family), and were a family of hundreds to thousands of people living on their own.

More on the neutering of ancient people. We have to imagine that they had personalities and there were stero-typical characteristics for certain ethnic groups. Their reasoning for raiding wasn't only; "I need new farm land and to feed my people", it was also because they had martial pride and wanted to be bad-asses. Same with modern criminals.

I don't want to come off as abrasive. I'm not standing by these points either, it's just what I know currently.

Davidski said...

Let's not get too excited and stray too much off topic, shall we?

That ASHG talk on the Corded Ware and other late Neolithic/early Bronze Age genomes is in less than two weeks. Let's wait and see what Laz has to say.

I have a feeling we'll hear a few very useful insights, and we might even see a preprint very soon at bioRxiv on the topic, and maybe also on the ancient Samara Valley genomes.

Mike Thomas said...

Grey
My friend, YOU'RE missing the point. Yu're talking as if the Yamnaya were a coherent people, centralized empire that organized special ops, and who, whether they used their horses for attack or logistics, had a specialy organized tactics, logistics and *Blietskrieg* stratagem. Moreover, Whilst horses *CAN* be ridden without more specialized gear, they simply would not have been able to do their special 'feigned retreats', attack at pace with bows or jevellins, etc. Read Robert Derws's book - Early Riders- he is a bronze age and military archaeology specialist.

Barack
"as also because they had martial pride and wanted to be bad-asses. Same with modern criminals. "

Had had don't worry Barak, your a thoughtful chap. And as I stated, I am not some lefty, anti-masculine, anti-migrations hipppy. All I am saying is that their is a grave misconception about the Yamnaya culture that is the product of "popular impression" perpetuated by the more populist writers that is empirically wrong.

Firstly the Yamnaya were bunch of small-scale communities, nothng over and above individual clan groups. That much is obvious by the lack of heirarchicization of the grave invetnory, and any differences in goods was the product of 'personal choiace ' as to what was used to comnunicate the image they wanted to portray about the decease - a nomad, a warrior, a man, perhaps a hero.

Moreover, as van der Linden has argued, the Bronze Age in Europe, - ie the supposed time of Yamnaya invasions and rise of Corded Ware culture, - was one of peace (!) and strong inter-group communication, required becuase of the stress associated with the late Neolithic population demise. In fact, the traditional mage of peaceful farmers is wrong. Mass graves with shokcing deaths have been found in Neolithic Europe, and none during the Bronze Age. See how this is entirely opposite to what most people *think* they know. The whole idea about tumuli burials and sharp sex diferentiation in the CWC period was to communicate the status of being a male. The identity of CWC people was the centrality of males as 'heros' almost to the neglect of women.

I dont want to sound arrogant or abrasive either, but the entrenched beliefs that a lot of people have here *I also * had 10 years ago, before I went above and beyond the mainstream aricles. I simply researched more specialist and upto date books on the topic, Google Scholar not to mention the hundreds of article on academia.edu


"Firstly, the direction of Balkan collapse is wrong"

That might be a valid argument. I'll leave that to others to argue'

Grey: unless the Field-Masrhsall of the Yamnaya office corps decided to do a flanking manouvre to beat the "nation" of the Gulmenitsa cuture before turning to the "nation" of Cucuteni. Ha ha please

Mike Thomas said...

...continued...

What we're really seeing is the internal colpase of the Balkan Copper Age chiefdoms, possibly due to endemic war between themselves. IT is exactly at this time thath the Majkop chiefdoms arose in the North Caucasus, with their elaborate wealth due to becoming the new metallurgical power. They Yamnaya was a petty emulator of the Majkop kurgans- nothing speciali about it. And they certainly had few if any horses, and certainly did not ride them. As I had repeated repeatedly, the large number of horses in Botai, and nearby regions were beginning to be domesticated for their secondary products. So even if you insist on claiming that these horses were emplyed in some 3000 BC special ops hit and run missions , they first would have to have invaded the Yamnaya people first, who were culturally different to the get to Europe. And here the entire Kugrganist construct melts away.

I 'll leave with a quote from David Whittle, the Neolithic/ Bronze Age specialist. "The Kurgan hypothesis brings to the Arhcaeology of the Copper Age a vision of irresistible steppe forces, which have more to do with later historical cases like th Mongols than the very varied patterns of change outline above {ie the false chronology of Anthony, etc}. It relies on crude characterization ...(abd) generelizations which are flawed."..And arguements which "rapidly become circular"... (p 138 Neolithic Europe (Read that chapter, its a good preliminary start for the introductee...)

Mike Thomas said...

But sorry Davidski for straying...!

Mike Thomas said...

Sorry, Finally ; Grey -
War in the Copper - Bronze Age was localized conflicts of say 20 - 100 men fighting over local resource points and strategic loci. This has been shown from burials and weapon collections. They were not large scale, and not territory -driven. And certainly no logistic truck squadrons.

Grey said...

@Mike Thomas

Or in other words raiding - and what gives an advantage in raiding, mobility (whether ships or horses).

"and not territory -driven"

but as the Moorish maritime slave raids show this could still lead to the people on the receiving end moving away.

.

equestrianism from the bronze age only

vs

youtube

https://www.youtube.com/results?search_query=bridle+less+riding

chelli11 said...

I'm brand new to DNA but am well seasoned in genealogy. As far as I know, all the ancestors that can be verified via paper trail in this test are Latvian (Baltic) peasant farmers, so Baltic, North Sea, Eastern European roots were to be expected. But...

My Eurogenes EUtest V2 K15 results are:
Population 
North_Sea 26.01% 
Atlantic 11.12% 
Baltic 38.28% 
Eastern_Euro 21.16% 
West_Med 2.08% 
West_Asian - 
East_Med - 
Red_Sea - 
South_Asian 0.48% 
Southeast_Asian - 
Siberian - 
Amerindian - 
Oceanian - 
Northeast_African - 
Sub-Saharan 0.84% 

And my Dodecad World9 result are:
Population 
Amerindian 0.30% 
East_Asian 0.10% 
African - 
Atlantic_Baltic 83.24% 
Australasian 0.54% 
Siberian 1.78% 
Caucasus_Gedrosia 11.30% 
Southern 1.48% 
South_Asian 1.26% 


I originally decided to take an autosomal test to see if I could verify a theory of mine that my mother had either Jewish or Turkic roots from somewhere - does the West_Med/Caucasus_Gedrosia point to this??

Mateus Smith said...

Davidski,

On what you said about how "everyone in Europe has admixture from everyone else, just in different quantities", could you tell me some more about the influence of ancient populations (like Germanic tribes, Celts, prehistoric Balts, Slavs, etc.) to each region of Europe from a genetic perspective, please?

For instance, do modern south Germans, French, Swiss and English inherit more from their Celtic or Germanic ancestors? what about the elements brought by Romans and other populations?

I guess it's more complicated than that, but I appreciate your opinion, thanks!

Davidski said...

If you had Jewish and/or Turkic roots, then you'd be showing West Asian, East Med and probably also Red Sea in the V2 K15.

The reason the World9 results are showing much higher levels of Caucasus_Gedrosia than expected by looking at the reference population averages, is because Dodecad tests aren't corrected for the "calculator effect". What this means is that users will see higher levels of the exotic components than the reference populations, and generally much noisier results.

But I have no idea what the Caucasus_Gedrosia component represents. If I had to guess I'd say it's a reflection of the heavy genetic drift among some of the ethnic groups in the Caucasus and Hindu Kush, which share the same ancient genetic components. If so, it probably spills out into Europe because Europeans also share some of those ancient components, and not because there were any recent migrations from the Caucasus/Gedrosia areas of Asia into Europe.

Davidski said...

Mateus,

It's difficult to say much about these things until we see more ancient genomes from Europe. We still haven't seen any from the migration period. These will be the first...

http://eurogenes.blogspot.com.au/2014/09/spanish-like-celts-and-finnish-like.html

barakobama said...

GEDmatch's "Are your parents related" test results for MA1: 11.4cm(5th cousin 25 cm), Stuttgart and La Brana-1: No, Anzick-1: 95.2cM(3rd cousin 100cm), Loschbour 18.8cm(5th cousin 25cm).

I expected more closely related parents for stone age samples than that, because I got the impression that people in small communities like the ones they lived in were all cousins and mixed with their 1st cousins and siblings. The Motalas will defiantly come out as cousins, because some of them are FMS mtDNA matchs and Y SNP matchs.

Balaji said...

Davidski,

Extended Data Table 1 of the Laz paper includes the following nuggets.

f3(Abkhasian;Georgian,LaBrana)=-0.0004, z=-0.5
f3(Chechen;Georgian,Loschbour)=-0.0002, z=-0.3

This would indicate that Abkhasians and Chechens have more UHG than Georgians.

Similarly Seinundzeit noted that from your data:

Pathan;MA-1,Lezgin -6.19285E-005 0.000303485 -0.204058

This he interpreted as showing that Pathans have more ANE than Lezgins.

Now that you have so many ancient genomes as well as the genomes of present-day people, could you compute f3(Test;Ancient,Pop). Test is any of 100 present-day populations and Ancient could be MA1, LaBrana, Loscchbour, Stuttgart etc. Pop would be the present-day populations plus the Ancients.

The readers of your blog can then try to glean information that the Reich Lab people have not yet published or which they have overlooked.

Nirjhar007 said...

@Every One.....
Italian Academic Scholar and Indologist Prof. Giacomo Benedetti has a very refreshing document on the Cradle of PIE people and Culture-
http://new-indology.blogspot.in/2014/10/can-we-finally-identify-real-cradle-of.html
In the article The Kurds And Iranians have a very crucial role! So it would be a great honor if you kindly visit the Scholars blog and give your valuable thoughts...
Have a Great Day.

barakobama said...

"Had had don't worry Barak, your a thoughtful chap. And as I stated, I am not some lefty, anti-masculine, anti-migrations hipppy."

I'm not an arrogant and angry-bickering old man. I just wanted to address a problem I see in the way academics view ancient and modern nations. You seem to be level-headed and someone who knows a lot about archaeology. Me thinking you kind of have this agenda may just be my miss-interpretation of you and lacking knowledge on what you were writing about.

I want to elaborate on what I meant to express. Yes, I do get the impression there is some anti-maleness in the academic world(maybe I'm wrong, just the fibe I get), by exaggerating ancient people's need for comfort, civilization, structure, dependability on external circumstances and having only weak internal responses. I doubt this is a huge problem, but it does exist and needs to be addressed.

Maybe it's because of our western society that some assume ancient societies were extremely structured, intellectual-like, religiously superstitious(always fearing and sacrificing to the scary Gods), prudish, and had high-class Victorian-era English accents(Hollywood needs to stop deceiving the public, I can't bare to watch 300).

There's an assumption that civilization, intellectuality, urbanization, and sometimes on average weaker citizens, is always seen as better by people than barbarism, primitiveness, small nations, and sometimes masculinity.

It's actually more like the opposite but people rarely admit this, and many ethnic groups(African Americans, Irish) who have pride in their barbaric reputation claim others are raciest for not believing they are civilized. And to give esteem to nations academics try to prove they followed our strict ideas of a developed ordered society and my school history books even resort to lying(Claiming, Mayans were comparable and equals to ancient Rome).

Overall I get very annoyed of academics worshipping stero-typical western civilization. I have a feeling the assumption that Europe is a completely separate and united entity compared to Asia and has been since Greco-Roman times(Greeks and Romans if anything associated themselves more with near easterns than chiefdoms deep in Europe), began when Europe became Christian(separate and unified compared to Muslim middle east) and in a large part Romanized(Christianity came through Rome).

Rome and Greece were pretty-much the only culturally ancestral ancients medieval Europeans knew of, and they probably admired how advanced they were, wanted to claim cultural decent(I understand it's largely true). Later intellectual-ages in European history I'm guessing especially admired the intellectual component of the ancient Greeks and Romans, wanted to be like them, and associated themselves with them. This might be part of the origin of this agenda I see in academics, and some people's assumption all Europeans trace their heritage to Greeks and Romans.

Mike Thomas said...

@ GREY
OK I accept that the horse can have been ridden without saddle. However, the "Yamnaya horse lord" take over theory has still problems

(1) Horses constituted a small minority of faunal assemblages in Yamnya (western steppe) region. So they were cattle hearders, not horse breeders -cum- equestrian conquerors.

(2) You havent addressed the issue of chronology/ direction: the copper mining centres of Serbia collasped first, then Bulgaria, then Romania, lastly Cucuteni-Tripolye shifted to a dispersed , more 'coarse' culture. This doesn't signify an invasion from th east (ie Yamnaya), but rather an "implosion" (ie internal crises of some sort).

(3) The 'hit and run' tactics you talk of require more or less sophisticated, central planning and infrasturcutre - ie at least a complex chiefdom. Simply, this is not what the empirical evidence tells of Yamnaya; which were egalitarian collectives of clans/ extended kin groups/ petty chiefs all more or less on an equal footing.

@ Davidski " difficult to say much about these things until we see more ancient genomes from Europe. We still haven't seen any from the migration period. These will be the first...'

I also believe that there is some work being done on the Lombards in central -eastern Europe, and the testing of their supposed Elbe-Baltic 'homeland'. I know they're testing isotope signatures, also hope they test for DNA. "http://www.les-treilles.com/?p=3462"

@ Barak
I for one rever not Greeks and ROmans, but Celts, Viking and Slavs, ha ha

Davidski said...

Balaji,

Good idea. I'll run those tomorrow and post the results here. I'll limit the analysis to the Human Origins dataset, to make sure I'm mostly testing the same SNPs.

Helgenes50 said...

The samples from Hinxton are available ?

http://www.y-str.org/2014/10/hinxton-dna.html#gpluscomments

Davidski said...

Holy shit, let me check this.

ryukendo kendow said...

@ Davidski

Understand if you are busy, but could you look into the fixing components in ancient genomes in ADMIXTURE runs idea?

From what Barak posted abt the parents relation test some tests somehow distinguish between Loschour and La-Brana, so WHG substructure is probably discernible.

Matt--or I think its him--pointed out how the ENF component reaches zero in NE.Euro, when we know this cannot be the case. I've long suspected that a component centering in Baltic is more reflective of recent pop movement, e.g. slavic exp, than of WHG per se. Same as what you said of Georgian and Ashkenazi behaviour in ADMIXTURE. Centering components in ancient DNA would help us know if this is indeed the case.

Thanks if you do.

@ Balaji @ Davidski

WHG forms a clade with ANE to the exclusion of Basal, and Georgian are basal-rich compared to N.Cauc and etc.

Not to minimise the very probable fact that WHG is higher in the N.Cauc than in S.Cauc, but this result you mentioned might be mostly tracking ANE in N.Cauc instead of WHG. Similar to how using Dai vs Sardinian makes all Europeans admixed with Dai, when it is actually because of higher WHG+ANE which are a clade with Dai.

This is in contrast to, say Lezgin vs Kalash, because they are both pulled by Basal away from ANE, and WHG does not play too big a role in both of them, so their results are comparable.

Davidski said...

I can't fix components on samples of my choice. Admixture doesn't let me do that.

I can try and point it in a specific direction in supervised runs with various reference samples designated as pure, as well as very carefully chosen datasets, and then Admixture either picks up on the hints or it doesn't.

I've already tried copying the components from Laz et al. using Admixture, with a wide variety of datasets, large and small, and the only component I can really tease out is ANE.

The components that take shape with La Brana and Loschbour as references always resemble much more my WHG-UHG component than Laz's WHG, no matter how I design the datasets. The components triggered with Oetzi and Stuttgart always end up looking more like my ENF than Laz's EEF.

ryukendo kendow said...

"various reference samples designated as pure"

This is exactly what I mean. Apologies for unclear language.

Could we fix BOTH La-Brana and, Ajvide as pure samples but in separate components? We could even repeat this for La-Brana, Loschour, Ajvide, Malta, etc. etc. This would force ADMIXTURE to model modern pops as products of past pops and pull out substructure in WHG/ANE.

For your AME run Basque scored anomalously high WHG. My suspicion is that Basque retains many La-Brana like sequences in its WHG, while most WHG in other European populations are derived from a more eastern source vis-a-vis their current position, compatible with Matt's Corded Ware Eastern Europe WHG introgression theory.

If we find discontinuity between French being far more Loschour/Ajvide like vs Basque more La-brana like, then this would be another piece of information.


barakobama said...

The raw data's Y SNPs has positive results for R1a1 SNPs.

http://www.y-str.org/2014/10/hinxton-dna.html#gpluscomments

Davidski said...

Like I said above, I can't force Admixture to form specific clusters. I can only designate various samples as pure, and then see what happens.

I've already tried using each of the ancient samples as "pure" references, and various combinations of them. Admixture just doesn't want to create the WHG and EEF components from Laz et al. It refuses to do so no matter what I try.

So either I don't yet have enough WHG and EEF samples, or these components never really existed as ancient populations, probably because EEF is modeled on Stuttgart, which is a mixed sample.

By the way, the WHG-UHG component always looks very similar, even if I don't use any Balto-Slavic samples in the Admixture run.

ryukendo kendow said...

@ Davidski
Does this mean that if you designate Loschour and La-Brana as pure, ADMIXTURE automatically assigns them to the same component?

Davidski said...

No, if I say that Loschbour and La Brana are each from different pure populations, Admixture automatically gives them 100% membership in these populations.

And if I mark them both as pure samples from the same ancestral population, it gives them 100% membership in that shared population.

But this need not have any impact on the rest of the dataset if these designations don't reflect the divisions that exist in the dataset. Admixture will just go ahead and create the clusters that work best.

Helgenes50 said...

I was comparing on Gedmatch the old genomes by using K36.
The results are very interesting.
Stuttgart, Malta and the three WHGs samples don't share the same components
as is the case with K15 for example where the Atlantic component is shared by Stuttgart and the WHGs. In K36, it is not the case the North Atlantic is only WHG
It's why, that's easier to see what we share with each of them.

The French component doesn't exist. This one is certainly composed of recent SNP ?

Davidski said...

Basically, the French are a very heterogeneous population genetically. So they're not very good at making clusters in these sorts of tests.

Shaikorth said...

K36 North Atlantic AFAIK is not quite the same as K15 Atlantic. The former peaks in Irish and the latter in Basques, thus North Atlantic is much less neolithic and its absence in Stuttgart is not surprising.

But don't read too much into it, these are broad strokes and ADMIXTURE with modern samples is bound to cause some distortion with ancient genomes.

Davidski said...

Preliminary results for Iron Age ERS389795...

K15

North_Sea 43.19
Atlantic 28.88
Baltic 6.46
Eastern_Euro 11.98
West_Med 6.71
West_Asian 1.74
East_Med 0.01
Red_Sea 1.01
South_Asian 0
Southeast_Asian 0
Siberian 0
Amerindian 0
Oceanian 0
Northeast_African 0
Sub-Saharan 0

K13

North_Atlantic 54.62
Baltic 26.12
West_Med 11.3
West_Asian 5.22
East_Med 0.07
Red_Sea 2.29
South_Asian 0
East_Asian 0
Siberian 0
Amerindian 0
Oceanian 0.01
Northeast_African 0.38
Sub-Saharan 0

So is that the Celt or the Anglo-Saxon? I have no idea?

Anyway, I need a favor. Can someone download this full autosomal csv and have a look if there are any SNPs after rs9948582. If there are, please upload the genotype calls in a small text file (zipped up if it ends up not being small).

https://drive.google.com/folderview?id=0B7vzRsRM2aOQR1BLbHduVUUxcjQ&usp=sharing&tid=0B7vzRsRM2aOQTENJUlB4OVVWeUE#list

I need to check that, and I'm on wi-fi now, so I can't get that file, which is 132 MBs.

Richard Rocca said...

These five are all pre-Roman Iron Age samples, so they are Celtic.

Davidski said...

Hey, is his Y-HG R1a?

Richard Rocca said...

I didn't go into all of the SNPs, but he is R1 (M306+) and R1a1 (PF6234+) but also R1a1 (M459-).

chelli11 said...

Thanks Davidski!
So the 2% West Med on the eurogenes test is probably just "noise"?

Davidski said...

It's probably a signal of ancestry from early Neolithic farmers, who were very Sardinian-like.

The West Med component peaks in Sardinia, and is the most important component for the Stuttgart genome from Neolithic Germany and Oetzi from the Copper Age Alps.

Alberto said...

@Richard Rocca

The abstract says this:

"We present whole genome sequences generated from five individuals that were found in archaeological excavations at the Wellcome Trust Genome Campus near Cambridge (UK), two of which are dated to around 2,000 years before present (Iron Age), and three to around 1,300 years before present (Anglo-Saxon period)."

So if this individual tested is R1a I would really doubt he is one of the 2 from Iron Age, but rather from the Anglo-Saxon period.

Davidski said...

There aren't any details about this file, but it's a high coverage genome alright, and this is what the ASHG abstract says about these samples:

"Good preservation status allowed us to generate one high coverage sequence (12x) from an Iron Age individual, and four low coverage sequences (1x-4x) from the other samples."

So this is probably the Iron Age Celt. He belongs to R1a, and this is where he clusters based on the K15 and K13 results shown above...

http://imageshack.com/a/img674/9271/XQ7f70.jpg

http://imageshack.com/a/img673/779/oKV8wT.jpg

Richard Rocca said...

There seems to be many more R1b SNPs now...

http://www.anthrogenica.com/showthread.php?97-Genetic-Genealogy-and-Ancient-DNA-in-the-News&p=54616&viewfull=1#post54616

Davidski said...

Ahh...I see.

Chad Rohlfsen said...

The individual turns out to be
44%EEF, 40%WHG, and 16% ANE using the calculator.. Are you sure it's R1a? If so it's likely a Saxon sample or recent migrant or descendant of one. Remember the Scandinavians in Kent, during the Iron Age.

Chad Rohlfsen said...

Those numbers are pretty Scandinavian....

Matt said...

@ Ryukendo, re the ENF component and why it doesn't show up in Northeast Europe, what might be interesting I've had a look at Davidski's PCA of Neolithic & Mesolithic and Synthetic Neolithic and Mesolithic sample and tried to make a screenshot of them that uses the same scale on the intra-West Eurasian samples, to try and show why this might be:

http://i.imgur.com/3M6ce2c.png

The synthetic ones are from an early run where Davidski gave us a PCA of them - I'm assuming there is not substantial shift between this and the present (I could be wrong). The West Eurasian samples on the synthetic run should be mostly the same samples.

Hopefully I've got the relative positions almost exactly right.

As you can see, the North Eurasian Forager has almost exactly the right position - its a little north and a little east of the true sample, but close (surprisingly to me as it doesn't show that high as a component when MA-1 is tested with ANE K=7 but it is what it is).

The problem component seems like the West European forager component, which seems like it takes a position that is a compromise of an ultra-Mediterranean farmer position and true West European foragers. (This probably also explains why the WEF component is so enriched in some populations where real West European forager ancestry doesn't seem to be).

As the WEF component is so far south ( i.e. less associated with ultra North European gene frequencies) relative to the real West European foragers, populations like the Lithuanians who fit a straight cline between this WEF and the NEF component don't need any ENF component and so get fitted without it (and I guess the Eastern Non-African components get added to give a little East shift without really affecting position otherwise?).

How to fix this component and get a real WEF to trigger (like how Davidski explains) would be a bit of a mystery to me though. For now just accept that the test is not accurate to ENF or WEF I guess.

ryukendo kendow said...

@ Davidski
Thanks for the explanation.

@ Matt
Thanks for yours too. Makes a lot of sense.

I always thought that this 'shifting around' only took place for Basal-like components, since unlike for WHG there is no ancient DNA for Basal. It seems that merely providing ancient DNA, like what we did here, is insufficient to get ADMIXTURE to produce a WHG-like component closely, unlike ANE for some reason.

An observation: this prob means that the west european forager component eats up some of the Basal component and migrates southwest on the pca, causing the early neolithic farmer component to be pushed further south than pure Basal Eurasian would be.

This adds a bit to my hypothesis that pure Basal would score northwest of bedouin and southwest from Sardinian and Basque, and that what pulls ENF and bedouin south in pca is African admixture, not basal, and thus that ADMIXTURE components that turn out hyper-bedouin cannot be used as a good proxy for Basal Eurasian.

Grey said...

@barak

"I expected more closely related parents for stone age samples than that, because I got the impression that people in small communities like the ones they lived in were all cousins and mixed with their 1st cousins and siblings."

A lot of foragers have very strict rules about who can marry who.

I think the reason is when they live in very small bands people have to be much stricter to prevent extreme inbreeding.

After farming, firstly there is now property to inherit so there is more incentive to keep it in the extended family and secondly the total numbers of the "band" are now a bit higher (say 3 or 4 large extended families in a marriage alliance) so the inbreeding effects might be diluted a bit.

Personally i think the foragers from that era will turn out to be slightly less inbred than the farmers because of this.

Just a guess though.

Grey said...

@Mike Thomas

"(3) The 'hit and run' tactics you talk of require more or less sophisticated, central planning and infrasturcutre - ie at least a complex chiefdom."

Among still existing pastoral raiding tribes in Africa they do it because of polygamy.

The young men can't get wives without cattle so they raid the farmers for cattle and women.

The young men organize it themselves in the same way urban street gangs do - no infrastructure.

This would also explain the male/female dna conundrum.

#

"You havent addressed the issue of chronology/ direction: the copper mining centres of Serbia collasped first, then Bulgaria, then Romania, lastly Cucuteni-Tripolye shifted to a dispersed , more 'coarse' culture."

I don't have a position on that. It may well be correct.

My point is simply that a horse culture potentially has a raiding mobility advantage before it has the technology for organized cavalry/chariotry armies.

i haz youtube as evidence


#

"This doesn't signify an invasion from th east (ie Yamnaya), but rather an "implosion" (ie internal crises of some sort).

Maybe so - but an internal implosion among the people who originally lived in the area west of the Black Sea would also imply less need for PIE to have a big cavalry army to expand into their territory.

Balaji said...

Ryukendo Kendow,

You wrote: "Not to minimise the very probable fact that WHG is higher in the N.Cauc than in S.Cauc, but this result you mentioned might be mostly tracking ANE in N.Cauc instead of WHG." This was in reference to the following f3 statistic.

f3(Abkhasian;Georgian,LaBrana)=-0.0004, z=-0.5

According to Laz, it is the Lezgins who have the highest ANE at 29% with Abkhasians having 19%. Yet the f3 stat for Lezgins is positive. Therefore the negative f3 stat for Abkhasians more likely is related to their having more UHG than Georgians.

Mike Thomas said...

Davidski, you argue further ANE intrusion into Euroep during the BA 9although it already existed there earlier), whislt Dienkes sees a "West Asian highland" intrusion at the same time. How do yu reconcile these two findings ?

Davidski said...

ANE existed earlier where? In Western and Central Europe? La Brana, Loschbour, Stuttgart, Gokhem2 and Oetzi say otherwise. So do the Sardinians from the HGDP.

Scandinavian hunter-gatherers had as much, or even more, in common with Eastern Europe than Western and Central Europe, so it's not surprising they carried ANE.

And there's absolutely no evidence of any intrusion from the West Asian highlands into Europe after the Neolithic. This blog post is about this fact. Only some parts of southern Europe show post-Neolithic admixture from the Near East.

It's really not particularly important what Dieneks sees. You should read Laz et al. a couple of times, and make sure you read between the lines too.

Mike Thomas said...

'Scandinavian hunter-gatherers had as much, or even more, in common with Eastern Europe than Western and Central Europe, so it's not surprising they carried ANE"

Yes, i meant Scandinavia.

"It's really not particularly important what Dieneks sees"

Not taking his word as law, of course, but I thought he demonstrated thie empirically in his blog....


Davidski said...

The descendants of Scandinavian foragers and seal hunters didn't populate Scandinavia after the Neolithic. Other groups did, in large part anyway, and some of them had to have come from regions where ANE, Y-HG R and Indo-European speech were present, which was probably Eastern Europe in the case of all three.

And Dienekes didn't demonstrate anything. He just confused a lot of people, some of whom are still running around and shouting about the Gedrosia component and other nonsense.

There was no Indo-European invasion of Europe from West Asia during the Bronze Age, and R1a isn't native to India. So two of Dienekes favorite theories have bit the dust.

Matt said...

@ Ryukendo I always thought that this 'shifting around' only took place for Basal-like components, since unlike for WHG there is no ancient DNA for Basal. It seems that merely providing ancient DNA, like what we did here, is insufficient to get ADMIXTURE to produce a WHG-like component closely, unlike ANE for some reason.

To explain what I think Davidski did to get the ANE and ENF components from his previous explanation, he *didn't* use actually an ancient samples to get the components.

The ancient samples would have very little sample weight (because there's like two or three of them), so they wouldn't trigger clusters in any panel of populations with enough size sufficient to test whether they're good clusters for our purpose - clusters trigger to explaining the most % of differences of the sample populations (what works best), so even though the ancient samples are very different, as they account for, like 1/1000 or something low of the sample size, they are hard to trigger off (I think). Also, as there are so few of them, it's hard for admixture to tell their individual differences from what is systematically different about them due to their population.

Plus, the ancient samples have little SNP overlap (with one another and modern samples), which reduces the power of ADMIXTURE to detect clusters even further (ADMIXTURE is dependent on detecting small differences in frequency over lots of SNPs, and lots of subjects, so more SNPs and samples the better, and clusters are noisy if you have far too few).

So because of all these problems, to get ANE and ENF, Davidski used portions of Native Americans and Bedouin (I think) samples that had East Eurasian and African segments extracted, as East Eurasian and African segments are fairly robustly identifiable, then used these to "trigger" components. He then continued to make slight adjustments (to the panel of input populations and SNPs, I think) until the populations looked like they had about the right median percentages of ANE in North Europeans and Siberians and ENF in Middle Eastern folks. These components were probably triggered in a sample set which was relatively heavy on West Eurasians.

So the WHG component, I think, ended up being what is basically what is left over in Europeans once the ANE and ENF have been triggered about where they ought to in their main populations of interest.

I think the problem with this is that this does seem to have left the WHG component as a wastebin / filler component - ADMIXTURE works to minimise distance between components and real population so we get a totally artificial population that can be mixed with ENF to create Sardinians and with ANE to create Northeast Europeans, but falls really short of the real WHG. Fine if the test is just to look at the isolated ANE, not useful if you're looking at the test to try and get a real measure of WHG/Basal Eurasian.

It's just difficult to see any way around this.

On your observation that the WEF component may be pushing the ENF component, if the process works as I think above, I think the WEF component is sort of a passive player in this - it is pushed to where it is, it doesn't push back.

Davidski said...

Here are some of those f3 ratios from the Human Origins...

Lezgins

https://drive.google.com/file/d/0B9o3EYTdM8lQUDlZNUtVZHpWeU0/view?usp=sharing

Pathans

https://drive.google.com/file/d/0B9o3EYTdM8lQOHBrSjJRanV3dEU/view?usp=sharing

Tajik Pomiri

https://drive.google.com/file/d/0B9o3EYTdM8lQaE1ZUWpFZmhMbGM/view?usp=sharing

Punjabi Lahore

https://drive.google.com/file/d/0B9o3EYTdM8lQa3JtdTVaQ0djdFk/view?usp=sharing

Kalash

https://drive.google.com/file/d/0B9o3EYTdM8lQSFBTVmVuYUI0S1k/view?usp=sharing

The Kalash don't actually get any negative f3 ratios, so I just posted their lowest 100 scores. I might need to curate their sample set and leave only the most outbred individuals to get something.

Anyway, I'll run this analysis properly over the holidays, and put up a blog post with my observations.

Seinundzeit said...

David,

Thanks! Very interesting output.

Seinundzeit said...

Also, if you do curate the sample set, it's probably best to keep only these Pashtun samples:

HGDP00214

HGDP00243

HGDP00213

HGDP00244

HGDP00259

HGDP00241

HGDP00254

HGDP00224

HGDP00262

HGDP00218

HGDP00234

HGDP00226

Based on what Everest did, these samples cluster very tightly together (I also have a personal reason for this, as it would make the HGDP Pashtun population behave much more like myself on these sorts of analyses, giving me a general idea of what I should expect for myself). Although, it's possible that some of these samples are missing from this data-set.

Balaji said...

Davidski,

Thanks for working on these f3 statistics. Are African samples included in the analysis? In your earlier analysis the most negative f3 statistics for Pathans involving MA1 were as follows.

athan;MA-1,Yoruban -0.00462641 0.00082202 -5.6281
Pathan;Yemenite_Jewish,MA-1 -0.00448872 0.000379465 -11.8291
Pathan;MA-1,Ethiopian_Amhara -0.00429134 0.000489328 -8.76988
Pathan;MA-1,Maasai -0.00416974 0.000633063 -6.58662

Davidski said...

Yes, Africans were included in this analysis. But this analysis had multiple ancient genomes, and it seems like Stuttgart and Oetzi took the place of the Africans.

Matt said...

Davidski, one thing which I did yesterday with the previous f3 stats you posted up before, which is kind of trivial but still maybe interesting, was that I ran PCA on the f3 stats.

When you use a lot of the ancient samples (MA1, various WHG, Stuttgart, Dai), and just West Eurasians (which I selected as populations above a certain threshold of f3 with Stuttgart), not too surprisingly you get a recapitulation of a similar parallel clines PCA.

http://i.imgur.com/62hzU0d.jpg
http://i.imgur.com/ocNNSDn.jpg - corresponding loadings

http://i.imgur.com/F3lcXiE.jpg
http://i.imgur.com/rmuA5tp.jpg - corresponding loadings

Obviously the loadings show that the PCA load mainly on WHG sharing vs non-WHG sharing ancestry (0 correlation with Dai, positive correlation with WHG, EEF and ANE in descreasing order) on the 1st PCA and then on MA-1/Dai vs Stuttgart sharing on the 2nd PCA, with WHG ancestry more or less neutral.

(PC3 and PC4 are together about an 80th in size PC1 and PC2 together, but for completeness, PC3 refines splits between MA-1 and Dai sharing, with populations who have lower MA-1 sharing than expected from PC2 plotting to one side of the graph (Sardinian, Uyghur) and those with higher to the other (Kalash, Palestinian, Lithuanian), PC4 then contrasts populations with higher residual affinity to MA-1 and Stuttgart (Caucasians) against those with residual WHG (Near Easterners).)

Using f3 statistics for MA-1, Karitiana, Dai, La Brana, Stuttgart and Australian Aborigine recapitulates the obvious world PCA we would expect for these panel of populations, although seems like with more influence by intra West Eurasian differences and the South Asians point more towards Oceanian populations, while retaining their pattern of pointing more towards MA-1 than would be expected from how much they point to Karitiana. I guess because f3 statistics sort of scale for drift?

PC1 & 2 - http://imgur.com/uyKl1dl

(zoom on Eurasia - http://i.imgur.com/3p4DHbv.jpg, West Eurasia - http://imgur.com/ngHqUKi)

PC3 & 4 -http://i.imgur.com/uiifSB3.jpg (zoom on Eurasia - http://i.imgur.com/3QB1S4B.jpg)

PC5 & 6 - http://i.imgur.com/7gKumMd.jpg
Loadings - http://imgur.com/b1TRCu7

Not sure if this actually clarifies anything, just surprised me that the same patterns come out in this data as well (but of course they do I guess).