search this blog

Monday, June 22, 2015

First look at an ancient genome from Neolithic Anatolia


Felix at GGT is in the process of uploading the genomes from the recent Pinhasi et al. paper. The file for the early Neolithic sample from Barcin, Turkey, is basically ready. I analyzed it with my K8 model and got these results (click on the image to enlarge).


I was only able to use a couple hundred SNPs for the test, so the outcome can't be taken too seriously. But it does make sense. The lack of Ancient North Eurasian (ANE) ancestry isn't surprising, because it mirrors the results of early European farmers we've seen to date.

Moreover, the relatively high level of Western European Hunter-Gatherer (WHG) ancestry, or at least something very similar, is also in line with expectations, considering that the sample was dug up in far western Anatolia, almost on the European border.

I also ran an Identical-by-State (IBS) affinity test using the Human Origins dataset and around 1800 SNPs. The results broadly back up the K8 analysis, with southern Europeans topping the list.


Citation...

Pinhasi R, Fernandes D, Sirak K, Novak M, Connell S, Alpaslan-Roodenberg S, et al. (2015) Optimal Ancient DNA Yields from the Inner Ear Part of the Human Petrous Bone. PLoS ONE 10(6): e0129102. doi:10.1371/journal.pone.0129102

See also...

The Near East ain't what it used to be

115 comments:

Mike Thomas said...

Great breakdown Dave

I'm surprised somewhat by the 20% WHG. But then again, perhaps not !

I wonder how specimens from Syria- Jordan-Palestine would look. I'd imagine; minimal if any "WHG".

I guess 'WHG' needs to be changed to EuroHG, given they appear to be a truly basal group which existed through to upper near East, whilst EHG are an admixed group (?)

Krefter said...

Awesome. ADMIXTURE results for BAR100 have been posted elsewhere. I put those results and ANE K8 results in a spreadsheet along with Stuttgart's results in the same tests.

https://docs.google.com/spreadsheets/d/1rTFReasuRyyWY169OhCjoW9bvVByu-Lrb50Xqu76Pak/edit#gid=0

How could BAR100's results make so much sense if she's of such low coverage? Anyways, I guess this means Early Farmers in Germany, Hungary, and Spain around 7,000 years ago were probably something like 85-90% Neolithic Anatolian. Even 5,000 years ago most farmers in Europe 60-70% Neolithic Anatolian.

People who know about Archaeology: So, is Turkey/Anatolia where the farmers who colonized Europe came from? Can we x-out the Balkans and SouthWest Asia?

Mike Thomas said...

Don't think so

I suspect Many of the WHG in EEF came from the northern balkans (west Croatia, Serbia, Hungary) but not South (ie greece) .

I think the Anatolian pre-Neolithic people were basically "WHG" (perhaps a misnomer now).

But there was massive movement from the levant to anatolia then europe. In fact this mihjt be why modern Arabs etc are so different to ENF: they almost all packed up and left for europe. !

Krefter said...

@Mike,

"I suspect Many of the WHG in EEF came from the northern balkans (west Croatia, Serbia, Hungary) but not South (ie greece) ."

That makes sense.

"But there was massive movement from the levant to anatolia then europe. In fact this mihjt be why modern Arabs etc are so different to ENF: they almost all packed up and left for europe. !"

That's what we see in mtDNA and Y DNA. EEF related lineages are very rare in West Asia today. It seems like EEF was a brother of West Asians' ancestors who decided to immigrate to Europe.

André de Vasconcelos said...

Awesome work, thanks.

Mike Thomas said...

Dave / Krefter
Was this sample male or female?
Any prospect for haploid
Markers ?

Chad Rohlfsen said...

I don't understand some of the thinking here. This could very well be about identical to the first farmers. Just because the first farmers came from the Near East, doesn't mean that those farmers had to be more Basal Eurasian than anyone. How do we know that they didn't drop Basal Eurasian in the Near East? Don't expect other Anatolians to be any different. They will likely form a cluster.

Chad Rohlfsen said...

Take that 8% SSA out of the Bedouins and they'll sit just kind of Caucasus shifted of this Anatolian.

Davidski said...

I think what we're seeing here is the result of a more Near Eastern population than this Barcin sample mixing with European-like foragers around the Sea of Marmara, and/or maybe migrations/backflow from Europe to Anatolia, to create these types of Anatolian farmers.

I don't think the farmers in Jordan or Syria will show any WHG. But I could be wrong, and they might show a bit.

Chad Rohlfsen said...

I don't think it'll get much more extreme than this. Even surrounded by pure WHG, EEF didn't increase more than 35% max in Gok2, after over 4k years. Most were only 15-20% after all that time. The Jordanian may be a only a bit more basal.

PF said...

I too am curious where exactly are the southern borders for WHG, and also when/how it interacted with EEF.

Exciting times.

André de Vasconcelos said...

A bit unrelated, but are there any (early) Neolithic samples from North Africa?
I'd really like to see how much WHG and SSA they had at the time.

Davidski said...

We might have to wait a year or two for usable Neolithic genomes from North Africa.

Anyway, I just updated the post with an IBS spreadsheet. It too makes sense.

Roy King said...

Davidski is right. There was a dense occupation of Mesolithic hunter-foragers around the Sea of Marmara. Aegean archaeologists don't see much of a connection between Pottery Neolithic (Barcin area) NW Anatolia and the Neolithic of Greek Theassaly and the Peloponnese. That said, the results might suggest that PN NW Anatolia contributed to the Starcevo/Impressed Ware of the Balkans and/or Cardial Ware of Italy. IT is amazing that the PN Anatolian sample has such a large WHG component to her admixture. It is often thought that NW Anatolian PN derived from an admixture of some farmers from the Catal Hoyuk area of Central Anatolia with local hunter-forager-fishers from the Sea of Marmara.

Arch Hades said...

Wow, just like the European farmers it's pulling very far 'west' in PCA. So relatively recent ANE ancestry is what's pulling everyone modern east?

Krefter said...

@Roy,

What do you think of R1b1c-V88 in Neolithic Spain. It's been confirmed.

https://docs.google.com/spreadsheets/d/12G2cfjG0wHWarsl5bB99ridFmvUWzqlZfZ6_e_R6oIA/edit?usp=drive_web

Krefter said...

WHG-related ancestry exists in all West Eurasians, even Bedouins. NorthWest Africans score almost 20% in ANE K8, and they don't score in ANE. High Basal Eurasian and ANE applies to some Middle Easterns, but not all.

Davidski said...

BedouinB and some Saudis don't show any WHG-UHG. They're basically just ENF/SSA.

And yes, everyone apart from BAR100, the EEF and WHG samples, and Sardinians are pulling east because of ANE and, in some cases, East and South Asian ancestry.

Mike Thomas said...

"BedouinB and some Saudis don't show any WHG-UHG"

I think because they're such a subset, endogammous, and drifted population groups.

Davidski said...

No, a lot of samples from southeast of Turkey have very little or no WHG-UHG. Even most Armenians only show a couple per cent.

Nirjhar007 said...

When we will know about the hgs?:).

Mike Thomas said...

Dave
Yes ok but my point stands- Saudis are not useful groups for comparisons. They're a mere subset of Semites. So they're not even representative of Semitic peoples, let alone anything else

Alberto said...

Great to see this sample up already, thanks!

I hoped it would be a high coverage one after the Pinhasi paper stating they got a high percentage of endogenous DNA from it, but it seems that quantity != quality.

I think, like Chad, that this might be quite representative of ancient Near Easterners, at least norther parts of it (Eastern Anatolia, Syria, North Iraq, Lebanon,...) 15-20% WHG might have been there from the beginning. It was probably migrations from more souther parts that diluted the WHG to the 0-3% levels we see today.

But I wonder how it happened. NW Africans do keep good levels of WHG. In Moroccans it could be attributed to gene flow from Iberia (there is a good amount of R1b there), but Mozabites? So I don't think the dilution came from NE Africa (or from SSA admixture). Probably from Arabia. Or maybe Iran? We'll need to wait for other samples to really know.

Matt said...

Cool. I think a position between the Middle East and Early European Farmers is really pretty consistent with what we'd expect, really surprised / impressed that this level of coverage can distinguish it.

As well as the SNP issue, I think there's probably some degree of projection compression. Even leaving aside the SNP count which you note, there's the way when D stats are run on Sardinians vs MN Europeans, the Sardinians come closer to at least Yamnaya e.g. D( Chimp Yamnaya Sardinian Spain_MN) D= 0.0122, Z= 4.587, as also shown on this PCA on M. Myllylä 's blog http://2.bp.blogspot.com/-XVEQYj8R9Oc/VXMAfdPqdOI/AAAAAAAAAmk/BHm9LOhTdiQ/s1600/pca12.gif which doesn't seem to show up for them on K8.

Aram Palyan said...

I think Levant will be different. Near East has a bioregional substructures.
Taurus Mountains divide Anatolia from Levant. Neolithic Farmers in Europe are most probably related to Anatolia.
In Levant and Mesopotamia we should find another type (related but different ) of farmers. IMHO.
You can see here Taurus Mountains at the South of Turkey.
https://ru.wikipedia.org/wiki/%D0%A2%D0%B0%D0%B2%D1%80_%28%D1%85%D1%80%D0%B5%D0%B1%D0%B5%D1%82%29#/media/File:Taurus-gebirge.jpg

Davidski said...

Matt,

The K8 focuses on broad ancestral components, in particular ANE in dimension two. So the farmers will never sprawl out like that past the Sardinians, because if they did, it would indicate that they had different levels of ANE, which they don't.

We don't know why Sardinians are slightly closer to Yamnaya relative to the Spanish_MN sample. But that Z score isn't as high as Mauri's plot would suggest, and it might not have anything to do with ANE.

I'd say the difference should be something as subtle as this...

https://drive.google.com/file/d/0B9o3EYTdM8lQb1NROWh4aWsxMkE/view?usp=sharing

Helgenes50 said...

Via the Roman Empire, they probably have a slight amount of Yamnaya

Alberto said...

@David

I was cross checking the results from your K8 with your K9 "Teal" runs. For example, this Jordanian's results:

Jordan445 K8:
ANE: 5.5%
NE: 79%
WHG: 5.4%
SS: 7%

Jordan445 K9:
Teal: 35%
Blue: 2.7%
Orange: 54%
Light_blue: 7%

The teal component is around 38% ANE in K8, so that alone should put this Jordanian at 13%+ ANE. And there's no other reason to take all that teal if not ANE, because there would be room for taking more orange (and no blue) to match all the components better.

So I think that Bedouins/Yemenite_Jews as ENF references are doing funny things to some populations. I don't know if there's a quick and easy way to check if that's the case (like using EEF as ENF reference and check if Jordanians or similar get the same ANE or not).

Shaikorth said...


Neolithics can plot beyond Sardinians if there's enough of them in the PCA to form a cluster. This happens in Mauri's plot and in Haak figure 2. If there's only one, it will plot with or right next to Sardinians, so this is a case where PCA's and f-stats perhaps should be treated separately.

Grey said...

Roy King

"There was a dense occupation of Mesolithic hunter-foragers around the Sea of Marmara. Aegean archaeologists don't see much of a connection between Pottery Neolithic (Barcin area) NW Anatolia and the Neolithic of Greek Theassaly and the Peloponnese."

One might follow from the other if the dense occupation acted as a barrier to migration.

The farmers - from wherever - may have had to sail round the Marmara HGs to get to Thessaly.

Davidski said...

Alberto,

ADMIXTURE components can have somewhat different structure in different individuals and populations. For instance, when the K9 teal component appears in most West Eurasians it's in large part Near Eastern, but not when it appears in the EHG samples. For them it's 100% ANE.

Alberto said...

@David

Yes, I understand that, but there is usually a reason for it. For example, the blue component is 18% ANE, so there's not way that EHG can get enough ANE from it. So they have to take teal that has 40% ANE, even if it's not a good match because of its 60% ENF.

But the case above shows that this Jordanian is choosing teal freely, without constraints from other components not providing what it needs. So I think in this case it's meaningful, or at least there is a chance it is. That's why I asked if there's an easy way to check it.

Davidski said...

There's no easy to way to check it that I can think of. Jordanians have some east African admixture, but there's no east African cluster in the K9. So that might be the problem? Or not.

Alberto said...

The light_blue is Sub-Saharan and it matches the 7% of K8, so I don't think that's the problem. I think it's related to Bedouins showing 0% ANE when they have some 25-30% teal, which affects populations with similar admixture.

Maybe we have to wait to get good ancient samples from the near east to be able to measure it accurately.

Davidski said...

Well, as we now know from multiple lines of evidence, Stuttgart is basically a western Anatolian farmer.

And it's only the BedouinA who show the teal at high levels, and they do have ANE. BedouinB weren't in the K9 because they were screwing it up.

Roy King said...

@Krefter
"What do you think of R1b1c-V88 in Neolithic Spain. It's been confirmed."
Yes, that's huge! Either V88 in Neolithic Spain is a relic hunter-gatherer who adopted farming, in which case V88 could have been associated with a Late Mesolithic migration from the Levant/Near East to North Africa and the Western Mediterranean as a trapezoidal/Castelnovian Mesolithic culture, Otherwise, V88 could have brought the Cardial Neolithic along both shores of the Mediterranean to Spain. In either case, the finding demonstrates a Mediterranean (Northern and North African) interconnection.

Mike Thomas said...

Roy

If by Mediterranean interconnections you mean direct from nth africa and iberia, I'd doubt that seafaring had come along that far by the mesolithic or even neolithic. The parasiminous explanation is spread from near East westward concurrently to africa and Southern Europe, most likely with Agro-pastoralism

Matt said...

David: We don't know why Sardinians are slightly closer to Yamnaya relative to the Spanish_MN sample. But that Z score isn't as high as Mauri's plot would suggest, and it might not have anything to do with ANE.

Shaikorth: Neolithics can plot beyond Sardinians if there's enough of them in the PCA to form a cluster. This happens in Mauri's plot and in Haak figure 2. If there's only one, it will plot with or right next to Sardinians, so this is a case where PCA's and f-stats perhaps should be treated separately.

Fair points both, yes the D stat variation is small, it does look like the EN do space out when there are enough to form a cluster and yes, D-stats vs Yamnaya don't seem to match up to the PCA (or geography) very much.

To be balanced, looking at the D(Test,CW;Yamnaya,Chimp) stats Tobus ran off for me, there *is* variation there in the EN : Starcevo_EN Z= -0.0554 D= -6.928 and Spain_EN Z= -0.0439 D= -12.013.

While modern for the same expression is like Sardinian Z=-0.0496, D=-17.613, Basque D= -0.035, Z=-12.904, Hungarian D=-0.0321, Z=-11.807, Bulgarian D = -0.0379, Z= -13.794, French South D=-0.0355, Z=-11.997.

Helgenes50 said...

The Barcin's results with K36 are interesting.
I like this calculator


0.00% Amerindian
0.00% Arabian
0.00% Armenian
32.73% Basque
0.00% Central_African
0.00% Central_Euro
0.00% East_African
0.00% East_Asian
0.00% East_Balkan
0.00% East_Central_Asian
0.00% East_Central_Euro
6.73% East_Med
0.00% Eastern_Euro
0.00% Fennoscandian
0.00% French
0.00% Iberian
0.00% Indo-Chinese
0.10% Italian
0.00% Malayan
0.00% Near_Eastern
13.88% North_African
0.00% North_Atlantic
0.00% North_Caucasian
0.00% North_Sea
0.00% Northeast_African
0.00% Oceanian
0.00% Omotic
0.00% Pygmy
0.00% Siberian
0.00% South_Asian
0.00% South_Central_Asian
0.00% South_Chinese
0.00% Volga-Ural
0.00% West_African
0.00% West_Caucasian
46.55% West_Med

Grey said...

Looks like what you might expect someone from the Atlantic Megalith culture to look like.

Sure he wasn't a merchant?

Grey said...

or her

Chad Rohlfsen said...

We already have Megalith samples, the MN groups. They're not like Barcin.

Grey said...

Sure, it's just odd looking - if you asked someone where ancient dna like this came from

33% Basque
7% east_med
14% north african
46% west_med

I'd imagine a lot of people's first thought would be Atlantic Megalith especially with its known connection to NW Africa.

No doubt it's just an artifact of the labeling.

Krefter said...

@Grey,

BAR100's cousins were (prob.)the main ancestors(84000YBP, they were already in Europe) of all Neolithic Europeans, not just the ones who made Megaliths.

postneo said...

Also needs to be seen how far east WHG/ENF propogated. IN the indian subcintinent Jats have some anomalous WHG not seen in Pashtuns farther to the west.

PF said...

@Helgenes50

Interesting. 0.10% Italian, 0% Near Eastern, all going to West_Med. I don't think the k36 is making sense here -- or is it telling us something?

Krefter said...

I added BAR100 to my 4mix spreadsheet. Look at some results I get with her.

French Basque: 27% WHG + 25% Corded_B + 48% Neo_Turkey + 0% Mbuti_Pygmy @ D = 0.0069

Tuscan: 3% WHG + 37% Corded_B + 60% Neo_Turkey + 0% Mbuti_Pygmy @ D = 0.0074

South Irish: 18% WHG + 61% Corded_B + 21% Neo_Turkey + 0% Mbuti_Pygmy @ D = 0.0069

Helgenes50 said...

@ PF

Yes indeed, I agree, BAR100 in This calculator is very West_Med.
and although the coverage is very low, the results, IMO, are not too noisy

Gioiello said...

@ Roy King
Have you realized that Iberian R-V88 (xV35, V69) is the oldest sample of this hg found so far and that it is ancestral to all the African and Middle Easterner R-V88 which are Y7771 and V69 derived? And that the oldest line is that which brings to the Sardinian M18? And why Sam Vass of the Jewish cluster (I am saying from ten years that it is Iberian derived) doesn't publish his Big Y (I have his data but not his .BAM file) above all when he thought that PF6289 was a link with Auwalu Musa he believed a descendant of a "lost tribe" but from their STRs I said that they were separated from at least 8000 years? And why smal (Sergey Malyshev) doesn't publish the tree of R-V88 and also of R1b1-L389* so everything will be clear?

Krefter said...

@Gioiello,
"Have you realized that Iberian R-V88 (xV35, V69) is the oldest sample of this hg found so far and that it is ancestral to all the African and Middle Easterner R-V88 which are Y7771 and V69 derived? And that the oldest line is that which brings to the Sardinian M18?"

That's interesting, I didn't know that. The old age of the sample though makes it believable. The idea R1b1 originated in all the way in Italy seems a bit far fetched But, keep posting your ideas and information. You are considering other origins though right?

Nirjhar007 said...

Ah yes The Parkhai Genome, Any News Krefter,David?

Chad Rohlfsen said...

Unrelated note here, the Okunevo look about 40% Native American, with the rest like EHG, and one of the Afontova Gora samples looks essentially 100% Karitiana, on many more runs than not.

Krefter said...

@Chad,

So Amerindian-like pops were living in Siberia just 2,000-3,000 years ago?

andrew said...

"Can we x-out the Balkans and SouthWest Asia?"

There were original domestications that became part of the Fertile Crescent Neolithic package in the Levant, in Anatolia, in Mesopotamia and the neighboring highlands, and in the Balkans (just as in the New World Neolithic package, corn and beans were domesticated in different places and integrated, and just as the Sahel Neolithic package was created from domestications in both West and East Africa that were integrated through trade).

Given the lack of a written language at the time and barriers to travel at the time, any exchange of domesticated crops had to be accompanied by at least some population exchange, but it wouldn't take much to produce lots of technology exchange.

The final coming together of the Fertile Crescent Neolithic package probably happened within a couple of centuries in the entire region with whichever Neolithic populations had the right combination of geography and population pressures spreading further into Europe and assimilating locals most readily at the every moving frontier of advance (which probably traced river basins).

Mesopotamia and the Levant may have contributed relatively little to the first wave of farmers, due to the mountain and water barriers to their migration, but it is likely that the first wave included both Anatolians and Balkan peoples.

Tobus said...

@Chad: one of the Afontova Gora samples looks essentially 100% Karitiana

Do you know which one?

Chad Rohlfsen said...

I'll post the number tomorrow. I've got some tests running.

Gioiello said...

@ Krefter
My last deep analysis about R-V88 is on eng.molgen, fortunately come to life yesterday, and palamede spoke about that also on Anthrogenica, demonstrating much courage. His post is there yet, even though nobody replied.
eng.molgen>R1b haplogroup>R1b-General&Misc>R-V88 in Sardinia (from the Italian Refugium?)
I am not able to send you a link, but there you may find many things to read, beyond that I wrote on Rootsweb (2007), DNA-Forums (2008), Worldfamilies, "Dienekes'Anthropology blog", Maju's blog, Anthrogenica, many FB blogs (but two banned me and I cannot look at them: R1a haplogroup and Isogg group) and many groups are restricted to Italian speakers.

Mike Thomas said...

Gio

I have expressed my doubts on your theory before, above all, because the obvious way the R1b phylogeny sits, and the ahistoric reconstruction your theory demands (=an otherwise undocumented expansion from Italy during the Holocene).

But I'm interested to hear out your story; so please summarise in simple point Form - and do not redirect to other sites or forums which I have no inclination to bother visiting. If your theory is sound, then it can be sunmarized succinctly here..

Gioiello said...

@ Mike Thomas
That I wrote on 27 April 2015 on eng.molgen to a person as to everyone could easily understand. I don't know if it is enough to you.
"Stoeni/Vettor/Pretotto, to speak to you risks to be useless unfortunately. 1) This is the theory of Anatole Klyosov, a sympathic person, but I didn't agree with him in anything. For having come from India, there should have been R-L389+ there, but India and Central Asia have only R-L389- and no subclades of R1b1-L389+ if not a few R-L23 come from Indo-Europeans, clearly from Eastern Europe with massive R1a under Z93 in spite of all the Indian nationalits. The map of the R1b1 FTDNA Project done by the infamous Author is clearly a flop: also the Kurganists say that R1b came from Siberia and not from Middle East. There hasn't be found no R-V88 in India or Central Asia.2) Only Western Europe has R1b1-L389+ with YCAII = 18-22, 18-23, 23-23 (see the test of InvisibleSun from Sicily), whereas the Caucasus, but practically Armenians, Indo-Europeans with a 30% of European descent and come from the Balkans, have R1b1-L389+ with YCAII =23-23. 3) R-L51 expanded clearly from Italy. 4) Bell Beakers expanded from 2900 AC from Italy, South France and Iberia and not from Samara. 5) Don't waste your time with the Anthrogenicians: I defeated they all once for ever. You will see next that all these subclades will be found in Italy and Western Europe. If you prefer, wait that one who understands very little of genetics like Rich and others finds R-L51 in Eastern European aDNA. When he finds it, phone me".

Chris Davies said...

Sardinia has unique branches of R1b R-V88 and also a unique branch of haplogroup A. Based on the hypothesis that R-V88 entered Africa from Italy and not vice-versa, one could make the same claim about haplogroup A. But nobody has.

Gioiello said...

@ Chris Davis
Of course what you say isn't true. Sardinia has a line of R-V88 which brings to M18 and to V35. Read my posts on eng.molgen where I demonstrated everything SNPs by SNPs. These two lines separated 13000 years ago and the line which bring to Sardinian V35 is the ancestor line also of the African Y7771 and V69, which for that are recent. The sample found in Iberia 7100 years ago demonstrates all that, and ask you, if the "R1b1 FTDNA Project" were right, why they don't publish the .BAM files of the Jewish cluster or any other sample from Africa or Middle East.

Gioiello said...

As I wrote to Wim Penninx on an Italian FB blog, you are analogical and not logical. You use the argumentum satellite haplotipo: what may be right for hg A isn't necessarily right for R-V88. Argumentum satellite haplotypo is pretty always the Fallacia satellite haplotypo.

Chad Rohlfsen said...

Afontovo Goro, Rise 554. Karasuk 497 isn't looking too far off either.

Helgenes50 said...

David,

Of all the available IE samples, on your K8 model.
Which one is the most Eastern? ( RISEs + Haak)

Alexandros said...

The WHG in the Anatolian farmer is extremely interesting! I really like Mike's proposition of reconsidering the labeling of WHG (EuroHG sounds good to me!). Apparently WHG people were present in the Balkans. I was always wondering whether HG-like admixture was present in Neolithic Greece prior to the arrival of the Near Eastern farmers and based on this finding, the answer is most propably yes! As I posted a couple of days ago, if this was a male, I would expect him to belong to hg G2 or I2 and the autosomal data seem to support that. We cannot conclude however whether the WHG admixture was already in NW Anatolia prior to the difffusuion of agriculture from the Levant or whether Balkannic HGs entered the region during the mid/late Neolithic. What we can say for sure is that such WHG admixture got diluted in Anatolia from the Bronze age onwards..

Mike Thomas said...

Gio
Thanks for the explanation

Alexandros said...

As with most people in the blog, I also believe that it is very unlikely that such WHG-like admixture will be found in the Neolithic Jordanian or any other Neolithic Levantine/SW Asian sample. If we do find such admixture, that would be a total shock!

Roy King said...

Is anyone thinking what I'm thinking, namely, if the Pottery Neolithic NW Anatolian sample clusters with the European Neolithic (circa 6400 BCE) and the Late Neolithic sample from Kumptepe (also NW Anatolian) from the abstract for the Croatian meeting has a component like modern Near Eastern samples that Kumptepe will show some of the Teal component and reflect a Late Neolithic immigration from the East (Eastern Anatolia/NE Anatolia)?

Chris Davies said...

@ Gioiello -

I am open to more unconventional theories.

But in the case of Sardinia I think that there is sufficient evidence for ancient gene-flow northward from Africa, which to some extent also impacts Corsica and northern Italy.

There is a multitude of evidence from papers on y dna, mtdna, HLA, GM immunoglobulin allotypes.

Off the top of my head I can think of E-V257/L19(xM81); E-V65*(xM78); A-M13; numerous mtDNA 'L' sequences; A*30:02-B*18:01-C*05:01-DRB1*03:01-DQA1*05:01-DQB1*02:01 [HLA haplotype] [https://en.wikipedia.org/wiki/A30-Cw5-B18-DR3-DQ2_%28HLA_Haplotype%29]; GM*1,17 23' 5* [GM immunoglobulin allotype]; etc.

So I would be suspicious of the claim that R1b R-V88 originated in Italy or Sardinia, rather than Africa or West Asia.

However your hypothesis *might* be correct!

Gioiello said...

@ Chris Davis
I wrote a lot also about that, for instance that there were very old hgs., for instance J and E (and others), from many thousands of years in Italy. One of them was just E-V65, and now certainly E-V13 has a subclade very old in Italy, perhaps older than the Balkan one. The young Vincenzo Passariello (Vincent here and Passa on Anthrogenica) belongs to hg. E and is studying this haplogroup with high interest, even though he is autosomally more Italian than me (84 against only 67, but the first 23andMe gave me a 100%, after I don't know what happened, but also for that I believe a little to the autosome, if not cum grano salis).

Gioiello said...

@Chris Davis

I remember a paper (perhaps you know that it is a fun to me to break in pieces the peer reviewd papers) in which the scholars tested an haplotype from Sicily for M18 and V35 and found it negative, but it is just an R-V88... evidently they thought it were impossible in Italy out of Sardinia and Corsica. But Marchesi, tested M269 from FTDNA, has been recognizd belonging to R-V88 in base of its STRs values after some letters of mine.

Davidski said...

Helgenes50,

The most eastern is still this Yamnaya sample from Haak et al. It's even more eastern than the Afanasievo samples from Allentoft et al. They only score around 32% ANE.

Population Yamnaya
ID I0429
ANE 0.3698
South_Eurasian 0.0541
Near_Eastern 0.215
East_Eurasian 0
WHG 0.3583
Oceanian 0.0002
Pygmy 0.0025
Sub-Saharan 0

But the current K8 overlaps poorly with the Allentoft dataset. I'll have to do something about that.

Gioiello said...

@ Chris Davis


From eng.molgen
European E-V257/L19
by Gioiello » Wed Jul 18, 2012 1:10 pm
Last post
Last post by Gioiello View the latest post
Wed Sep 10, 2014 9:54 am
6
Replies
Replies
1997
Views

Helgenes50 said...

But the current K8 overlaps poorly with the Allentoft dataset. I'll have to do something about that.

Thanks

Gioiello said...

@ Chris Davis

E1b1b1b1 CTS1271/M5064 * CTS6304 * CTS5480/PF2360... 106 SNPs formed 23800 ybp, TMRCA 13900 ybp
E1b1b1b1*
E-PF2431 FGC18913/Y10538 * PF2466 * PF2431... 21 SNPs formed 13900 ybp, TMRCA 10900 ybp
E-PF2431*
id:HG02317 ACB
E-PF2438 FGC18915/Y10559 * PF2469 * FGC18918/Y10542... 12 SNPs formed 10900 ybp, TMRCA 9200 ybp
E-PF2438*
id:HG01699 IBS
E-Y10539 Y11300 * FGC18923/Y10561 * FGC18922/Y10562... 20 SNPs formed 9200 ybp, TMRCA 6900 ybp
id:HG02798 GWD
id:YF02677
Sardinians are negative for Y10539 and it isn't said that HG02317 ACB is of African origin rather that Iberian one. The Sardinian samples 46, 47 and 48 are upstream M81, which is very young in Africa and I should do other studies to say if Europeans derived from Africans or the other way around. Anyway the samples 46, 47 and 48 are very old in Sardinia and in Iberia. After the separation from the M81 lines the Sardinian samples had 101 SNPs plus 46 (8), 47 (9) and 48 (22) private ones. I. e. the Sardinian line may have been separated from the M81 line also from the 13900 years given from YFull. The Sardinian line belongs in fact to the PF2438 subclade and is linked to the Iberian sample on the YFull tree. Of course these old samples should be compared, what I cannot do because I haven't their data.

Chris Davies said...

@ Gioiello -

The issues as I see it are that Africa is chronically under-sampled relative to Eurasia. Plus we have a lack of ancient DNA samples from Africa and West Asia. And Africa today is substantially different today from 10kya, taking into account Saharan dessication, numerous population expansions & contractions, migrations, etc. And there is a general lack of interest in African genetics from many Euro-descended gene bloggers and commenters. So we can debate this ad nauseum, and people can claim whatever they want to claim. There are plenty of people adamant that haplogroup E entered Africa from Eurasia. I can neither deny nor confirm that R1b R-V88 in Africans is derived from migrations from Iberia or Italy. But I remain generally sceptical of the claim at this stage given the other genetic evidence of ancient migrations from Africa into Europe via Sardinia.

Gioiello said...

@ Chris Davies

Then for your fan read also this last post of mine on Human Genetic Population of my friend Vincent/Passa:

And if I demonstrated that the ancestor of all the E-M183, whom YFull esteem 2100 years old, wasn't a North African but a Sardinian? Samples 49 and 50 of the Sardinian pool from Francalacci 2015 have 2 and 19 private mutations. How is it possible? The other Sardinian E-M183 belong to two subclades: 1) samples 51-58 and 2) samples 59-114 (starlike), who have a sum of 24 SNPs in common before the starlike separation, thus they are long older than 2100 years.

Alberto said...

Are the other samples already available? It seems so from the site linked in the post. I hope we can get some decent data from them. Jordan and Turkmenistan are unexplored areas that should be very interesting if at least we get a rough idea of how they look like.

@Gioiello

Thanks for your explanations about your hypothesis. I think it's clear now. We just need to wait for ancient DNA to prove it or disprove it. The same team that just released these samples has 50 samples from Italy listed in their site (from UP through the Neolithic), so hopefully by the end of the year we will know.

Gioiello said...

@ Alberto

I thank you for this news. We have only to wait therefore. I was surprised that anyone had tested these samples before, because I know from many years that the anthropologist of Pisa University Mallegni gave many bones to Dr Caramelli and also to a
French team.

PF said...

The WHG in the neolithic Anatolian is indeed interesting. Perhaps this genome fills out a chunk of the puzzle, especially with respect to southern/southeastern Europeans?

More specifically, I think the way Eastern_Med populations were previously modeled generally as “near eastern farmer” + “European HG”, is more accurately modeled as “local near eastern” + “neolithic Anatolian.” That is, pretty much all the WHG in these populations is coming from the neolithic Anatolians and later Yamnaya-related populations, and not Loschbour-related populations.

Check out these 4mix numbers I’m getting with the K8 averages. Adding BAR100 along with Loschbour, Yamnaya, and Yemenite_Jew is producing some pretty damn good fits.

Population Loschbour BAR100 Yamnaya Yemenite_Jew D statistic
East_Sicilian 4 51 22 23 0.0028
Bosnian 15 45 40 0 0.0041
West_Sicilian 4 55 23 18 0.0045
Sardinian 21 73 0 6 0.0046
Sephardic_Jew 0 44 20 36 0.0049
Serbian_Bosnia 17 45 38 0 0.0057
Greek_Thessaly 5 63 31 1 0.006
Kosovar 7 62 31 0 0.0062
Central_Sicilian 4 51 21 24 0.0063
South_Italian 0 63 23 14 0.0069
Ashkenazi 2 42 24 32 0.0073
Ukrainian_Lviv 22 30 45 3 0.0075
Bulgarian 10 51 36 3 0.0079
Macedonian 11 54 35 0 0.0081
Spanish_Castilla_Y_Leon 25 34 19 22 0.0081
Spanish_Cataluna 25 38 22 15 0.0082
Moldavian 18 29 46 7 0.0084
Spanish_Murcia 26 20 16 38 0.0086
Basque_French 30 52 17 1 0.0087
North_Italian 14 63 23 0 0.0089

**Now, without knowing the exact math/methodologies behind 4mix, or how much stock to put in these results, the following is all speculative.** I guess what this suggests is that East_Med was spread around the Mediterranean by descendants of peoples who were already mixed between Anatolian farmers and a Red_Sea-related population, but NOT a Loschbour population. Likewise, Balkan people are a mix of the Anatolian farmers and Loschbour, but NOT Red_Sea.

I’d say the Anatolians are either from Southern Anatolia / Northern Levant, or moved downwards there. They are probably very closely related to the people who first discovered agriculture. Along the southern edges they mixed with a local Red Sea like population. Along the northern/western expansion, they mixed with Loschbour.

I'm also curious why "Neolithic Turkey + Yemenite_Jews + Yamnaya" fits so well for the South Italian, Sicilian, and Ashkenazi/Sephardic Jew cluster? An accident, or a genuine relic of Sea Peoples / Phoenicians / proto-Greek and/or Canaanite colonizers / something else?? And, if this cluster represents recent admixture between a near eastern group and post-Bronze age Europeans, as usually assumed, wouldn't we expect at least a bit more Loschbour in this modeling?

Next questions: 1) exactly when and who spread Yamnaya to East_Med, and how much WHG/ANE did this population have? 2) When/where did neolithic Anatolians get their WHG? 3) Are Yemenite_Jews really a good proxy for ancient southern Levantines? (Looking forward to that Jordanian genome and hopefully others.)

I’d be keen to hear whether my analysis makes any sense. And thanks to @Davidski and @Krefter for posting the 4mix package. BTW, is there a list somewhere of all pop gen tools available for R?

a said...

@Gioiello

I'm curious if we can have K8 results for the R1b v88 and other samples from spain- Haak et al 2015?
https://docs.google.com/spreadsheets/d/1xS830gSmb1QvPdrZqcF0YOdygLCIyDNJP5IXO-QychI/edit?pli=1#gid=1351765170

I'm curious why new tree has no R1b-v88 placement; a bit odd?

I'm curious if Felix is going to get around to input the Haak et al samples from 2015 in his ancient dna table? The 8000 year old Anotolian sample was sure done quick- with quick blog coverage to boot.

http://www.y-str.org/p/ancient-dna.html

Gioiello said...

@ a
I think we already know the answer, because all the EEF examined so far were close to Sardinians, and look at what is coming out from the last paper of Trombetta et al. about hg. E: the oldest subclades of M35 are just V68 and V257 found already in Sardinia from the previous papers of Cruciani ad his team. And also about that I wrote tons of letters...

Nirjhar007 said...

Gioiello,
What y-dna are you expecting from the Barcin one ?

Gioiello said...

@ Nirjhar007
I did some previsions in the past and I failed all. I prefer to wait for the answer of the data, also because finding an haplogroup or another has no definite meaning without a whole glance upon all the rest. Think to all who thought having found the origin of hg. R1b with the first results from Yamnaya...

Nirjhar007 said...

All Right:)....

a said...

@Gioiello
R1b-v88 if correct/confirmed; is quite high in
K15- 87.12% combination West Med and Atlantic, higher than Sardinia and Oetzi and Basque all very old. How can this be if he migrated from Levant.
https://docs.google.com/spreadsheets/d/1xS830gSmb1QvPdrZqcF0YOdygLCIyDNJP5IXO-QychI/edit?pli=1#gid=1351765170

Levant is opposite- for example Yemenite have 4.45% Atlantic+West Med.
https://docs.google.com/spreadsheets/d/19c_bZjUV_RouKyGyLHmMDw57WwAVabXFJOaso_gcuRE/edit#gid=1872836177

Rather curious.

Richard Rocca said...

@Gioiello, you said..."Don't waste your time with the Anthrogenicians: I defeated they all once for ever. You will see next that all these subclades will be found in Italy and Western Europe. If you prefer, wait that one who understands very little of genetics like Rich and others finds R-L51 in Eastern European aDNA. When he finds it, phone me".

Since you are the self proclaimed defeater and destroyer of all other theories: Three more Copper Age Italians were found to add to Otzi, and all of them are typical Neolithics in both Y-DNA (I2a & G2a) and autosomal DNA, and not a single mention of the Remedello findings by you, the originator and only believer of your "Italian Refugium"??? Add to that the fact that the likely Italian Cardial-derived Early Neolithic Iberian is R-V88, and we all know that R-V88 has nothing to do with Western European R1b. Not one damned post about that either??? For all the denying that any immigration EVER occurred INTO Italy (no Yamnaya, no Celts, no Greeks, no Etruscans), you don't have a single post explaining why the Y-DNA and autosomal DNA of peninsular Italians has undergone such an important shift since those four Copper Age Italians? By not mentioning these findings, you are coming across as a fraud.

Also, you knocked STR dating for all of its flaws and now you are hanging all of your false hopes on SNP counting dates? Please, stick to the ancient DNA.

Mike Thomas said...

LOL. I think his blind -sited Italocentrism is pretty obvious to all.
Still talking about 'modal haplotypes' in 2015.

a said...

@Richard Rocca

Come on now Richard you were not even as hard on Connan the Barbarian>> now known as Connan the Barbarian in a chariot.

Grey said...

@Krefter

"BAR100's cousins were (prob.)the main ancestors(84000YBP, they were already in Europe) of all Neolithic Europeans, not just the ones who made Megaliths."

Well quite but would you imagine they'd have the north African before they moved instead of after?

On the other hand her being a mail order bride from the western med. at this time seems pretty unlikely.

Either way the north African bit seems to be the curious bit.

Grey said...

Speaking of R1b-V88

If for the sake of argument it did enter Africa from outside what would be the most likely route to where it exists now - through Egypt?

Taking those two ifs it makes me wonder about the claimed R1b of King Tut - what clade?

Gioiello said...

1) I never called you "Rich". Thus that Rich weren't you.
2) But your answer is meaningful. Onwards I'll think to you as a second "Rich", thus, as the "posthomous fidelity" which the great Viennese spoke about, I say to you all what I said to the other Rich.
3) Your argumentations are inconsistent, because about R-V88 from Sardinia-Italy or Iberia I wrote a lot from so long (I have letters of the great Cruciani and Scozzari who were, in spite of you, possibilists) and I think having now demonstrated my hypotheses (there are long argumentations on eng.molgen) and that is the first defeat for who did the R1b1 haplogroup tree before coming back to his plants, for who doesn't publish his Big Y, for all the theory that R1b and subclades came from Middle East.
4) About all the rest your indictions are deserving of you: we have explained here which is the truth about Remedello and about the R1b1 found in Eastern Europe. These fact don't tarnish my theory: you have to find R-L51 in Eastern Europe and about the fact that my R-Z2110 (tested with a Full Genome which has only 300 no-calls out of 72000 SNPs and not the 12000 of Big Y) is in Italy from at least 6100 years and is brother of the ancestor of the CTS7556 which generated all the Eastern CTS9219, thus the expansion happened from West to East. Anyway I have always submtted my ideas to the proof of the aDNA and for that I am waiting willing.
5) If your boss send you here to provoke me, hoping that I am banned once more, I hope that your wife's blood is more courageous and honest than yours. I have nothing more to say to a lackey who supports useless tests and wrong tree.

Richard Rocca said...

Once again you have skirted around the ANCIENT DNA data. I asky you the question again...Why is there no R-L51 in Remedello and Otzi, why are they so obviously shifted from modern day peninsular Italians, and why do modern day peninsular Italians have Yamnaya ancestry that Copper Age Italians and Cardial Ware Neolithic Iberians lacked, if accoring to you, no migration ever went into Italy and Iberian Cardial is derived from Italy?

Karl_K said...

Wow. This Gioiello guy is great. Almost as good as the one that cannot be named.

Chad Rohlfsen said...

Italic and Basque-centric bullshit is all over. No one can let go of 2003. L51 is not from Italy, Iberia, or West Asia. Get over it.

Krefter said...

@Grey
"Either way the north African bit seems to be the curious bit."

The sample is low coverage, so all ADMIXTURE results should be taken with a grain of salt. IMO, the only thing we can be confident of about BAR100 is that she was closely related to EEF. A relation to North Africa, maybe even with Sub Saharan ancestry, is of course possible though.

Alberto said...

Another interesting detail about this first sample from West Asia is that the top IBS list is completely dominated by European populations. Unless the Neolithic in West Anatolia was a local development, if this farmer came from around Syria or Lebanon, then it means that the population turnover since the Neolithic in the Near East has been quite more dramatic than the one in Europe. The only true Near Eastern populations in the list are:

Position 23: Druze
Position 48: Palestinian
Position 49: Iraqi Jew

By contrast, Europeans (including Ukrainians or Belarusians) top the list, with 18 out of the first 20, the other 2 being Jewish populations with likely European admixture.

Another interesting thing is that the Caucasus seems to have kept more of the ancient Near Eastern component than most Near Eastern populations, in spite of the high ANE.

Let's see if the sample from Jordan confirms this or maybe it's quite different and reveals that EEF came indeed from West Anatolia.. For me more unlikely, but how knows.

Mike Thomas said...

It really seems bewildering. I'm aware that the entire Natufian complex seems to have disintegrated perhaps with a climactic event, shifting northwestern migration. But to think that almost entire population moved on (hundreds of thousands ?), and some other subset from further elsewhere replaced it is amazing, but entirely plausible. We'll see

Krefter said...

@Alberto,
"Let's see if the sample from Jordan confirms this or maybe it's quite different and reveals that EEF came indeed from West Anatolia.. For me more unlikely, but how knows."

Felix is done with the Jordan sample. I read a post where someone said it's of lower coverage than BAR100. There's a sample from Neolithic Serbia of better coverage though. I hope someone analyses it.

y-str.org/p/ancient-dna.html

Krefter said...

@Alberto,
"Let's see if the sample from Jordan confirms this or maybe it's quite different and reveals that EEF came indeed from West Anatolia.. For me more unlikely, but how knows."

Felix is done with the Jordan sample. I read a post where someone said it's of lower coverage than BAR100. There's a sample from Neolithic Serbia of better coverage though. I hope someone analyses it.

y-str.org/p/ancient-dna.html

Davidski said...

I can't run the samples from Jordan or Turkmenistan, but the Vinca sample from Serbia looks like this...

K8 NG21-10
ANE 0
South_Eurasian 0
ENF 41.1
East_Eurasian 0
WHG 58.9
Oceanian 0
Pygmy 0
Sub-Saharan 0

If this is legit, then this guy was obviously of mixed origin, with lots of Balkan forager ancestry.

Mike Thomas said...

Wow, almost 60% WHG !
Any chance of haploid marker ?

Krefter said...

Davidski can you run Formal stats with the Vinca Serb? He might have had a pure-WHG grandparent.

Davidski said...

I can't run any formal stats, but the top 25 IBS list looks like this, indeed suggesting an EEF profile with inflated WHG.

Spanish_Aragon 0.830404
Scottish_Argyll 0.829988
Sardinian 0.829757
Icelandic 0.829569
English_Kent 0.829501
French_South 0.829052
Basque_French 0.828946
Bulgarian 0.828863
Spanish_Cantabria 0.828766
Spanish_Cataluna 0.828727
Spanish_Baleares 0.828701
Lithuanian 0.828313
Hungarian 0.827868
Croatian 0.827666
Norwegian 0.827628
Spanish_Castilla_y_Leon 0.827491
Italian_WestSicilian 0.827434
Orcadian 0.827429
Greek 0.827330
Finnish 0.827216
Turkish_Jew 0.827202
Spanish_Murcia 0.827101
Italian_Bergamo 0.827058
Italian_Tuscan 0.827024
Czech 0.827012

Chad Rohlfsen said...

Compared to the EEF model, that should be something like 57%EEF, 43%WHG. Gok 2 was something like 65/35. This is very interesting and unexpected. I'd figured they would be like Oetzi.

Chad Rohlfsen said...

Sure as hell isn't Halaf related, as Maju said...Vinca is certainly native European.

Nirjhar007 said...

Ans since When Maju takes Autosomes seriously?;)
Krefter, Whats the News on Parkhai one? i'm informed that it will be low coverage, so anything?.

Davidski said...

This Vinca sample basically looks half WHG and half BAR100.

But keep in mind that there are so few markers that the result might be biased in favor of WHG.

Helgenes50 said...

Serbia, Gomolava

9,23 % of Steppe is noisy ?

61.02% Near_Eastern
0.00% East_Asian
0.00% Siberian
0.00% Oceanian
26.07% WHG-UHG
3.68% Sub-Saharan
0.00% Hindu_Kush
9.23% Steppe
0.00% Amerindian
0.00% Southeast_Asian

Alberto said...

Yes, this sample from Serbia is very interesting. But being just one we cannot know if it really represents the whole population or it's some exception (after all, we already found a pure HG among farmers in Hungary, so this sample could be a son/daughter of such case of HG and Farmer couple). Though, on the other hand, the high incidence of I2a in the area could indeed mean that local HGs adopted farming early and the resulting population was like this single sample. And Vinca, like CT, could be an important source of LN population expansion.

Pity the low quality of the samples from Jordan and Turkmenistan. I hope they have better luck with other samples from those areas in the future.

Helgenes50 said...

Yes, this sample from Serbia is very interesting.

Probably due to a low coverage, I wonder where is his steppe from?
This one is supposed to be of IE-like origin

Aram Palyan said...

In Pinhasi study there should be some aDNA from Areni cave in Armenia.
It is Copper age
Is there any news about it?

Simon_W said...

I'm a little late with my 2 cents here, but yeah, the Vinca sample looks interesting. Vinca didn't belong to the earliest Neolithic cultures of this area, so in its chronological standing it might be rather compared to the MN cultures of central/northern Europe, and these too had more WHG admixture than their EN precursors. And evidence of early Neolithic cultures has not been found everywhere in the Balkans, so like in Blätterhöhle in Germany, some WHG may have roamed around for quite a while. But importantly, this one lacked ANE, so it cannot explain BA Hungary.

ANTON T said...

hello,
For k36 eurogenes can you please post a world map showing the regions of the 36 different regions defined? It will be very useful because there is ambiguity in the region names!

Simon_W said...

I doubt Davidski will make a map, but there is no need for it, just see the list of samples used for the K36 components here:
https://docs.google.com/file/d/0B9o3EYTdM8lQUlVTRmxQdGpZdW8/edit