search this blog

Tuesday, February 20, 2018

Migration of the Bell Beakers—but not from Iberia (Olalde et al. 2018)


At last, after many months of waiting, the paper that I've been calling the Bell Beaker Behemoth will finally appear at Nature today or tomorrow, depending on your time zone [Update: the paper is here]. The accompanying dataset is already online, and it's twice as big as what the paper's bioRxiv preprint promised, packing 400 new samples from Neolithic, Copper Age and Bronze Age Europe (freely available via the Reich Lab here).

I'll incorporate these samples into my collection of ancients very shortly, and then put them through their paces in the usual and new ways.

Nevertheless, despite the much larger and more varied new dataset, I know for a fact that the conclusions in the paper are the same as those in the preprint (which we discussed here). The authors tentatively accept the archaeologically-based academic consensus that the Bell Beaker phenomenon originated in Copper Age Iberia. But they admit that they can't find evidence in their data that its expansion across much of the rest of Europe was accompanied by significant gene flow from Iberia.

However, they do see in their data a large-scale migration of Central European Beakers to Western Europe around 2500 BC, bringing with them, amongst other things, steppe or Yamnaya-related admixture to the region for the first time. Many of the new samples are from the British Isles - where the impact of this migration was profound, resulting in roughly a 90% turnover of the population - and they appear to have been collected specifically to reaffirm this conclusion.

How exactly this massive population turnover came about isn't yet known. But early indications from other parts of Europe, where similar population shifts have been inferred from ancient DNA for the Late Neolithic/Early Bronze Age period, are that plague epidemics and deadly violence may have been important factors (see here and here).

I don't have a strong opinion about the place of origin of the Beaker cultural package, and I don't find the Iberian model entirely satisfying, mostly because it doesn't gel with the latest ancient DNA data. On the other hand, I've made up my mind as to who the Central European Beakers rich in steppe ancestry and also Y-haplogroup R1b-M269 were, and you can read about that here.

What are your thoughts after looking over the new samples? It's a big dataset alright, but does it do justice to the massive and complex Bell Beaker phenomenon? If not, then what's missing? Who's actually happy that the puzzle of the origin of the Beakers has now been solved? Feel free to let me know in the comments below.

Update 21/02/2018: I've updated my Global 25 datasheets with most of the ancient samples from Olalde et al. 2018 and Mathieson et al. 2018 (see list here).

Global25 datasheet ancient scaled

Global25 pop averages ancient scaled

Global25 datasheet ancient

Global25 pop averages ancient

See also...

Who's your (proto) daddy Western Europeans?

177 comments:

rozenblatt said...

TWIMC: preprint of Langobard DNA study: https://www.biorxiv.org/content/early/2018/02/20/268250

Arza said...

Mathieson data is also online. Ukraine_Eneolithic_outlier was moved to _Neolithic, but there are two new EN samples (male and female). Some quick and dirty modelling:

New one:
Ukraine_Eneolithic_I5884
Ukraine_Neolithic, Steppe_EMBA, Trypillia
best coefficients: 0.702 -0.019 0.317
std. errors: 0.051 0.069 0.038
chisq: 6.486
tail: 0.484230548

R1a-guy
Ukraine_Eneolithic_I6561
Ukraine_Neolithic, Steppe_EMBA, Trypillia
best coefficients: 0.038 0.527 0.435
std. errors: 0.049 0.070 0.043
chisq: 5.946
tail: 0.546069571

Chad said...

I've also got the Beakers, Matthieson SE Europe, and Vanuatu new samples. I'll do my first Admixture run soon.

Joe said...

Where can the samples for the Mathieson SE Europe paper be found? Link?

Samuel Andrews said...

@Rob,

Hopefully, they mention how the holy Balkan-Caucasus two way deserves all the credit for this migration of Steppe-rich R1b people. No DNA papers has given it its due credit. Weird.

Samuel Andrews said...

I'm interested to see if British Beaker folk really do have a close relation to modern Isles Celts.

Despite what I've said before, now I tend to think the BEaker folk weren't the proto-Celts. Iron age Celts from England as well as modern English have a strange French-affinity in David's north Europe PCA. Also, the Y DNA of Welsh is less understood than the Y DNA of Irish. Maybe they don't all have R1b L21 like the Irish do. Maybe, Celts arrived in the Isles from France, with people heavy in Beaker ancestry, in the Bronze age.

Samuel Andrews said...


Also, Celts from iron age England place between Irish and French in the North Europe PCA. British Beaker+proto Celt mix?

bellbeakerblogger said...

Since the Matthieson is out, I'd like to see some comparisons between Iron Gates and Blatterhohle. They're all easy to pick out, R1b and U5. Take your pick.

"What's missing"

Michelsberg elite genomes at the cemetery in Gougenheim.

Arza said...

Re: Mycenaeans

Can our Dereivkan be a proto-Greek?

Ukraine_Eneolithic_I6561, Minoan_Odigitria
best coefficients: 0.276 0.724
std. errors: 0.108 0.108
chisq: 9.293 tail: 0.75044264

Maybe. What about that guy from the Balkans?

Balkans_BronzeAge_I2163, Minoan_Odigitria
best coefficients: 0.217 0.783
std. errors: 0.066 0.066
chisq: 7.471 tail: 0.876294233

Much better.

Balkans_BronzeAge_I2163, Europe_LNBA_CWC, Minoan_Odigitria
best coefficients: 0.003 0.258 0.740
std. errors: 0.364 0.384 0.073
chisq: 6.486 tail: 0.889656925

But a specific subset of CWC seems to be taking over the lead.

But in Olalde we have a bunch of new samples. Let's check that angle.

Corded_Ware_Czech, Minoan_Odigitria
best coefficients: 0.348 0.652
std. errors: 0.115 0.115
chisq: 5.579 tail: 0.9601576

Oh boi. I mean "ahoy!".

Poland_BA? What's that?!?

Poland_BA, Minoan_Odigitria
best coefficients: 0.233 0.767
std. errors: 0.084 0.084
chisq: 5.194 tail: 0.970675418

Arza said...

^^^
qpAdm log: https://drive.google.com/open?id=1tcRxRGYnjeC3UoUTJd3uw0RAZeYYOwXX

Chad said...

Michelsberg isn't going to be any different than German_MN, England and Scitland_MN, and French MN. R1b L51 is not from Western Europe, if that's what you're insinuating.

AWood said...

@Samuel Andrews

Hopefully the British Bell Beaker have results for downstream mutations of P312+. If we see that 2000-1800 BC Britain is rich in L21+, the high rate of L21+ in Ireland (and M222+ by extension) is likely a founder effect from the Bronze migrations on the continent. Recall that Ireland is quite remote and Rome never really touched it. It would also be the least impacted by later Celtic tribes.

Who knows what language they spoke, but they certainly had steppe ancestry.

Samuel Andrews said...

@Awood,

The preprint already showed British Beakers were rich in R1b L21.

Nirjhar007 said...

Finally one of those days that we gene lovers wait for :) .

Nirjhar007 said...

around 500 BC

Nirjhar007 said...

Nope around 300 BC with Mtdna H2a1.

Chad said...

A little K5. It makes some type of interesting Steppe component. It appears to already include the Anatolian ancestry in Yamnaya, with Yamnaya being this group, plus Khavalynsk. It is ADMIXTURE of course, so until we see a population like this, it is probably nothing. It is interesting though to see Yamnaya, Corded, and Beaker get similar numbers of this population with later groups getting really watered down. I will look at formal statistics in a couple days. I have a paper due tomorrow night.

This has a bit of quality control and a lot of MN pops taken out to avoid that cluster. about 70% of the samples made it in.

https://docs.google.com/spreadsheets/d/1OMUvoV6epqDV-S5v291iPf3CI-PP3QZsJtjo8pB0-lA/edit?usp=sharing

Chad said...

Don't get too involved in that cluster for the Steppe. It could easily and likely be a result of the number of Beaker and BA samples overwhelming the number of Yamnaya samples.

Chad said...

Langobard study

https://www.biorxiv.org/content/early/2018/02/20/268250

Sofia Aurora said...

Excellent news! At last the first major surprise of 2018!
Davidski may we assume that the South Asia paper is next?

Arza said...

@ EastPole
You need to see this: https://s6.postimg.org/5qbm0ew1t/Ukraine_Baltic_Poland.png

@ Chad
Thanks!

Rob said...

Dave / Arza
One of the Lombards was R1a-Z283, in quick glance; as was the accidental BA Hungarian

Aram said...

Genetiker assigned the Spanish BB (those without Steppe) R1b-s as V88. And there is two Z2103 in BB now. One in Poland and the other in Hungary.
The Hungarian is even positive for L584 (found among Armenians). But I don't think it is true L584 :)

Rob said...

Late Paleo iboussieres (south France) also R1b-L754. Surprising all the earliest R1b expanded along south-Central Europe (south of alps)

Michalis Moriopoulos said...

@ Chad & Dave

Not sure if you're aware, but some Mehri samples have also been released in a new paper on Southern Arabia:
https://datadryad.org//resource/doi:10.5061/dryad.1pm3r

The Mehri are a very interesting population phenotypically and genetically (I ran one of them through Eurogenes K15 and he had both the highest "East Med" and the highest "Red Sea" amounts I've ever seen). Well worth including in any dataset, I'd say.

Davidski said...

@Mike the Jedi

Not sure if you're aware, but some Mehri samples have also been released in a new paper on Southern Arabia: https://datadryad.org//resource/doi:10.5061/dryad.1pm3r

Thanks, I'll run them in the Global 25 as soon as I can.

Archaelog said...

@Davidski Are there Hungarian Yamna genomes in the new dataset? Haven't been able to open and see for myself


@Salden Which is the article published by the lab in the Youtube video? The one with the R1b Greek IA sample like you said? If there is a paper, can you link it here?






Davidski said...

@Chetan

There aren't any Hungarian Yamnaya samples in either the Olalde or Mathieson datasets.

There are four new Hungary_BA and four new Hungary_LCA samples in the Olalde dataset. But I have no idea how these samples are related to Yamnaya, if at all.

And the presentation that Salden linked to is based on as yet unpublished data, so there's no paper yet.

Archaelog said...

@Davidski Ah that's such a pity. Was finally expecting to find the R1b Hungarian Yamna samples. Wasn't the Hungarian tumuli part of the LCA - EBA though? I don't have a clear idea of the archaeological dating in those parts. Do you know if any such samples are being sequenced anywhere for future publishing?

Davidski said...

@Chetan

Wasn't the Hungarian tumuli part of the LCA - EBA though?

Yes, and the above mentioned Hungarian samples might be closely related to Yamnaya. We'll find out soon.

Do you know if any such samples are being sequenced anywhere for future publishing?

No, and I remember Iain M. telling me a few months ago that he wasn't aware of such samples being tested anywhere. But maybe that's why those Hungarian LCA/BA samples are now in this dataset?

Archaelog said...

Revised Late IE migrations proposed routes https://imgur.com/a/8wUOt

Anthro Survey said...

Pretty wired about that Langobard study(as are Simon_W and Rob probably).

The results aren't that unexpected for Colegno, I guess, given what's known from history.

-Would have been nice(r) if they included HGDP Bergamasques, though, because we can't be super sure if those TSI-like Colegno locals were more Tuscan-shifted than modern Piedmontese or if just a projection issue. Could shed light on Germanic admixture in that part of North Italy.

-Assuming three were really so TSI-shifted compared to the two more Iberian-shifted individuals, there's an explanation for the structure.
The latter originally hailed from the hills around Colegno(Avigliana or so, per Strontium data), while the TSI-shifted people were Padanic lowlanders more influenced by East Aegean DNA at some point.
Modern Padanic Piedmontese(from what I've seen) cluster around Bergamo, indicating a homogenization of these sub-populations over time---growth of medieval and early modern cities, etc.

-As for the TSI-shifted Hungarian locals---
Strontium data doesn't rule them out being originally from Pannonia, albeit from a different locale. Now, this was a bit unexpected becasese I didn't think Circum-Aegean influence made THAT much of an impact north of the Danube.

-We can surmise that "proper" south Illyrians were at least similar to Tuscans, if not Abruzzese, from this. So, the modern Balkan cline is almost entirely due to Slavic admixture. My past usage of Bergamasques and Tuscans for modeling Macedonians, Serbs, etc. likely wasn't off the mark.

-Modern Hungarians convincingly look like a four way mixture of North "Illyrians", Slavs, Germans and Magyars. We'll see when Dave puts them through G25.

Nirjhar007 said...

If anybody cares :

Happy International Mother Language Day

Anthro Survey said...

@Dave

When you update the G25 sheet, can you also add Guanche please?

Davidski said...

@Anthro Survey

I don't have the Guanche genotypes yet.

Davidski said...

@All

The Global 25 datasheets have been updated with the following ancient samples from Mathieson 2018 and Olalde 2018.

https://drive.google.com/file/d/1NkmxB9zmV_tMJfM_SCI7_awdmyMLIKrk/view?usp=sharing

I've also added a good number of new modern-day individuals, and removed some outliers, so the pop averages for many of the modern-day populations are now different.

Anthro Survey said...

@Dave
Ah, ok. I saw the Genetiker guy included them in his ADMIXTURE, so I assumed they were readily available to the general public.

Grey said...

if BB is early in Iberia (according to archaeology) and yet there is no trace so far of migration between there and the central European BB then maybe the Iberian ones arrived via a coastal route - hence faster/earlier than those taking the land route?

EastPole said...

@Chetan

“Revised Late IE migrations proposed routes https://imgur.com/a/8wUOt”


There is an article in “Scientific American” March 2018 Issue on IE languages:

https://www.scientificamerican.com/article/new-evidence-fuels-debate-over-the-origin-of-modern-languages/

It is refers to old DNA studies. It is interesting because they say that “there were pendulum migrations back and forth,”. It looks like steppe IE migrated to central Europe and then back to to the steppe, to Sintashta and India. This is the best explanation of why Slavic and Indo-Iranian languages show many similarities.

Rob said...

@ Chetan
Your map has some major issues - R1a going to Greece and Armenia, when it should obviously be Z2013-R1b

Rob said...

@ Dave

For the Balkan _BA, did you remove the R1a-Z93 outlier ?

Davidski said...

@Rob

Balkans_BA:I2163 is the R1a-Z93 Bulgaria_MLBA sample.

Anthro Survey said...

Oh, and looks like the North Africa study is updated, but not yet published.
https://www.biorxiv.org/content/early/2018/02/20/191569

No conclusive f4s yet, but it incorporates Guanches from the other paper and they resemble KEB, unsurprisingly.

Another new paper, but this one seems a bit amateurish and over-relies on ADMIXTURE output.
https://www.biorxiv.org/content/early/2018/02/20/268524






Archaelog said...

@Rob Yes I know. That was and still is one of my strongest doubts. While Greek is Centum it also shares a lot of features with I-Ir. So it is certainly possible that proto Greek was carried by a second wave of R1b from the steppes or the Balkans (later than the R1b that carried Italic-Celtic)

@East Pole Yes even if the Corded Ware group that gave rise to Sintashta didn't come all the way from Central Europe, it seems they did have a lot of contacts with that region and the EEF admixture probably entered that way. From the maternal side I think.

Rob said...

@ Chetan

Yes I agree that's possible. First Greeks were Z93, and Z2013 came later with LBA/ Iron Age highland Greek'ized groups. Will need a lot of samples to show that.

Rob said...

@ Anthro

Of course a paper from the centre of Europe will always be good :)
Moreover, it's really getting into the knitty-gritty of social structure and how it intertwines with population genetics.
I'll wait for now when /if Dave opens a thread for it for comments, because the paper needs propper digesting.

Davidski said...

@All

That Shriner preprint is a bunch of crap.

Rob said...

Dave thanks for processing the data. It's quite interesting for just a rough quick look ofthe G25, using Chalcolithic sources.

For example, as Alberto suggested, there does seem to be a significant genetic shift in Ukraine between Eneolithic & Yamnaya period, one emanating from the Don-Kuban-Volga interfluvial.

Yamnaya_Ukraine
"Samara_Eneolithic:I0122" 63.75 + "CHG" 21.25/ "Armenia_ChL" 8.95
"Ukraine_Eneolithic" 6.05
"Balkans_ChL" 0


Hungary Mako - EBA, quite distinctive. The high Ukr Eneol. is as if they, in turn, moved West of the Carpathians as Yamnaya came in. High WHG admixture.

Hungary_BA:I1502
"Ukraine_Eneolithic" 37.1 + "Narva_Lithuania" 20.4
"Balkans_ChL" 36.55 + "Tiszapolgar_ECA" 5.15


Ukraine Eneolithic also prominent in Balkan BA:
-Balkans_BA
"Ukraine_Eneolithic" 39
"Greece_Peloponnese_N" 38.25 / "Balkans_ChL" 11.75 /"Greece_N" 5.05
"Armenia_ChL" 4.2


The non-steppe Admixed Vucedol

Vucedol:I2792
"Tiszapolgar_ECA" 47.45 / "Greece_Peloponnese_N" 29.3/ "Greece_N" 17.75 / "Balkans_ChL" 5.5 (near complete continuity)

The Steppe admixed Vucedol
Vucedol:I3499
"Tiszapolgar_ECA" 41.1 + "Balkans_ChL" 12.2
"Samara_Eneolithic:I0122" 14.4 + "Ukraine_Eneolithic" 7.5 + CHG" 12.5
"Anatolia_ChL" 8.35

For, later: eg
Beaker_Central_Europe
"Globular_Amphora" 44.2
"Yamnaya_Samara:" 20.45/ "Yamnaya_Ukraine" 16.65/ "Ukraine_Eneolithic" 10.1
"Vucedol:I3499" 6.5

Eren said...

@David:
Would it be possible to include your Basal-rich K7 synthetic pops in the Global25? I'm mostly interested in the Basal-rich one.

jv said...

If there was 90% population
turnover during the Beaker period in the British Isles, why do some modern British Isles folks retain so much Western Hunter Gatherer admixture? Was it reintroduced during the Viking Age? Or did resurface at the end of the Neolithic when some of the Neo Farmer groups were struggling? (Some of my Irish Family were originally Norse. Toner Family is originally Tomar[and defeated by the Irish in Dublin and had the Tomar neck ring stolen!] and Norse & the Scottish/Irish McCabe’s were originally Norse)

Anonymous said...

Just a small note for other commenters: while the Bell Beaker migrants to the British Isles did most likely (though not certainly), given the genetics, speak an Indo-European language, they were NOT Proto-Celtic speakers.

If you look at the significant changes that must have occured between Proto-Indo-European and Proto-Celtic, and if you then look at the relatively minor changes that then must have occured between Proto-Celtic and Old Irish, it is just not linguistically feasible to have a Proto-Celtic so early - the Celtic languages cannot have split by that time-period. Indeed, Proto-Celtic is linguistically hypothesised to have been spoken around 900BC, give or take a couple of centuries, and the archaeology seems to back that up.
(and when you consider that proto-Celtic itself either descends from Proto-Italo-Celtic, or else developed alongside proto-Celtic with sprachbund effects, that further hammers down the coffin lid of the proto-Celtic Beaker theory.

That does leave the possibility that Bell Beakers in the Isles spoke pre-proto-Celtic, long before the "late proto-Celtic" that broke up later on. But that would mean there would have to be a mass Celtic invasion FROM Britain TO the continent, which goes against all the other evidence.


No, clearly the celtic languages must have entered the British Isles from the mainland more than a thousand years after the bell beaker culture.

Now, we could suggest that the BBs spoke a sort of "macro-Celtic" - an IE language that wasn't definitionally celtic, but that was more closely related to celtic than to other IE families. However, there's absolutely no reason to suppose this. Given that celtic didn't arrive for another thousand or even two thousand years, and probably arrived from central Europe, there's just no reason other than national myth to associate the early IE settlers of the Isles (or Iberia) with Celtic languages specifically. We might, I suppose, guess at a broadly "Western" (non-Graeco-Aryan) language, but even that is just guessing.

Besides, at these dates, no IE languages except Anatolian and possibly Tocharian would really have been that strongly differentiated.

Ric Hern said...

Are there any genetic connection between Vucedol and Bell Beakers in Portugal ? Maybe I2a ? Sea Route...?

mike said...

I think people need to get their heads around the fact that Celtic celtic in a genetic sense, Celtic in a societal structural cultural sense and Celtic as in the sound shifts that distinguishes it from italo-celtic were not identical or contemporary things. As long as the beaker people were west indo-europeans then the obvious participation in intense interaction across western and west central Europe c2400-700BC (which almost certainly meant wives moving to seal diplomacy) means the shifts that define Celtic could have been spread arealy among the elites across western Europe then spread down the social scale. These elites could have kept a constant convergence process that meant sound shifts spread in synchronicity over vast areas and divergence did not happen as long as the network existed

Slumbery said...

@jv

The newcomers also had some Western HG ancestry at the time they reached Britain.

Ric Hern said...

@ vacuouswastrel

What is interesting is the Irish which seems to have clearly separated from the majority of British DNA at a very early stage. So for a long time the Irish and British basically kept to themselves. I really do not see a major Continental Celtic Cultural or genetic influence that moved into Ireland that could have changed their Language untill the Norse Migration...Unless Saint Patrick introduced Celtic with the help of Religion ? Which I doubt...

But yes maybe the so called Celtic Languages was never One but rather were Late-Proto-Indo-European dialects who leveled out in some places creating something similar in the process....

MomOfZoha said...

@David:

Do you have the most updated moderns-only and ancients-only versions (separate sheets) of the scaled PCA 25 coordinates (and averages) that you could possibly share? It is just much easier to merge two files than it is to pick out each of the ancients from one. Thank you...

szahel said...

Hi everyone,

I usually just read the posts, but this article it seems important.

https://link.springer.com/article/10.1007%2Fs12520-018-0609-7?utm_source=foter_ro&utm_medium=link&utm_campaign=foter_ro_201802

Nirjhar007 said...

SEE paper :
https://www.nature.com/articles/nature25778

BBC paper :
https://www.nature.com/articles/nature25738

Anonymous said...

I appear to have missed some updated. Sion one third steppe..

Davidski said...

@szahel

I usually just read the posts, but this article it seems important.

https://link.springer.com/article/10.1007%2Fs12520-018-0609-7?utm_source=foter_ro&utm_medium=link&utm_campaign=foter_ro_201802


Interesting paper, but the data they're using is low resolution. I'll have to wait for the next generation sequencing version.

@MomOfZoha

Do you have the most updated moderns-only and ancients-only versions (separate sheets) of the scaled PCA 25 coordinates (and averages) that you could possibly share? It is just much easier to merge two files than it is to pick out each of the ancients from one.

Sorry, I don't.

Anonymous said...

@vacuouswastrel

The North-West Block theory proposes that NW Europe knew a unknown language which showed affiliation [1] to Italic languages. So the theory could be that the BB's filled Europe with a language family of which most languages died out but from which two survived, Celtic and Italic.

[1] The Netherlands knows a lot of names related to water which show an -apa (reconstructed) suffix.

@Katriina

Germanic is sometimes proposed to be a mixture of Kentum and Satem languages, I keep reading online. But I can't seem to find any serious article on it. Is there something valid to this theory?

Joukowski Transform said...

@vacuouswastrel Interesting, but since there doesn't seem to be much evidence of Celtic culture in the south of Ireland, especially the far south that kept distinct and apart from the rest. How exactly would Celtic have spread there?
Also, I think the earliest recorded written Irish is mostly from that area.

Archaelog said...

@vacuouswastrel I would say the Bell Beaker-spread language was still close to the late PIE of Yamna rather than proto-Celtic. So it could be described as pre-Celtic but not proto-Celtic. Whatever late PIE derived language arrived first in Britain could have taken a completely different trajectory than proto-celtic which evolved on the continent and later spread to Britain in the Iron Age overlaying and completely wiping out the earlier IE language that was present there.

Grey said...

" But that would mean there would have to be a mass Celtic invasion FROM Britain TO the continent, which goes against all the other evidence."

if other evidence contradicts it then it can't be but...

if a bunch of dairy farmers arrived in a region which was a) relatively bad for neolithic crop farming and b) perfect for cattle then you might expect a dramatic and rapid population expansion with maybe some boomerang effect

and

copper + tin = bronze weapons

and

if this R1b distribution map was a blood spatter diagram you'd be looking for the source at a point south of ireland and west of brittany.

http://dgmweb.net/DNA/Graphics/392x635-R1b-DNA-Distribution.gif

(not saying that's evidence of anything - just odd)

Rob said...

Czech EBA heavy I2c admixture like German Unetice & Hungary LCA

Archaelog said...

@Rob So 2 out of three Hungary_BA samples (non Beaker) are R1b and one is I2a. And the LCA samples are all G2a. Not a single R1a. More evidence against Anthony's theory and against those who say the Corded Ware was not IE speaking.

Rob said...

@ Chetan
Your tally is totally incorrect because you’ve missed a whole lot of previous samples
There is one Hungry BA R1a funnily enough in the Lombards paper
But overall, I (still) think that CWC spread via forest zone and BB via Romania / east Hungary

Archaelog said...

@Rob David said that particular R1a guy could have been a CW migrant. But glad to see that you agree the CWC spread via the forest zone.

Rob said...

@ Epoch

“Germanic is sometimes proposed to be a mixture of Kentum and Satem languages, I keep reading online. But I can't seem to find any serious article on it. “

Then it’s not much a theory is it ?
Have you come across anyone credible suggestig that Germanic has Satem features ?
In fact even Baltic is only partly satemized

It seems Germanics (the real Germanics; not Nordic BA) were heavily I2a2 and R1b; and are steppe + GAC
GAC is missing in early Baltic CWC ; it’s basically straight Yamnaya . Even lacks Narva

Ryan said...

@Chad, David, all - what do you think of Bell Beaker Blogger's third neolithic theory as an explanation for the archaeology? IE that the Beaker culture was able to initially spread back through Europe from Iberia by using a common cultural network from a recent Neolithic expansion? This post:

http://bellbeakerblogger.blogspot.ca/search?updated-max=2017-11-24T10:36:00-06:00&max-results=5&start=5&by-date=false

I think it would explain well the La Brana -> KO-1 shift we see in Iberia between the MN and CA in extended data figured 4.

Why are folks assuming Bell Beakers spoke an IE language at all? I think it's a huge leap to think that whenever steppe cultures mixed with non-steppe cultures that the steppe language always won out, and I would suggest that the archaeology and historical evidence still favours a predominantly non-IE speaking Atlantic fringe. I don't think there's any question that many of those Bronze Age inhabitants of the Altantic coast had a great deal of genetic material from IE speakers, but I don't think that means they were speaking an IE language.

I'd note that that has no bearing on the source of R1b-M269 among Beaker peoples. I realize my views on R1b are fringe among this group here - I'm not resting my case here on that. Y-chromosomes can jump cultures. Wessex is a good example of that - with such a classic Anglo-Saxon kingdom being founded by someone with a Celtic name, so even if you view Yamnaya->Bell Beaker for R1b-M269 as case closed I don't think that should rule out non-IE languages in the Atlantic fringe. Especially since one of those languages persists today.

I guess what I'm suggesting is that we have this interaction between CWC and non-steppe Beakers and other BA cultures, and out of that interaction comes a mixed culture. The eastern part of that mixed culture ends up speaking an IE language which becomes the Urnfield culture and ultimately Italo-Celtic and Lusitanian.

On the western side the non-steppe language wins out and becomes the Atlantic Bronze Age.

Does that sound plausible?

I think the best way to test this would be to see if the admixture the CWC ancestry in late Beakers and Atlantic Bronze Age samples is sex biased. If it has a male bias then I'd think the Northwest Bloc theory is probably correct and the Atlantic Bronze Age was a (mostly) IE culture. If the admixture has a female bias or is relatively even then I'd suggest that a Vasconic or other pre-steppe Bronze Age language would more likely.

capra internetensis said...

@ Rob

it's been suggested that Germanic has a form of the RUKI rule, though that falls rather short of being mixed with satem.

https://www.researchgate.net/publication/260258597_Germanic_and_the_RUKI_dialects

Rob said...

@ Capra
Ah yes RUKI seems shared, ? via Z282

AWood said...

@Ryan

The latest data from the BB paper in Iberia shows M269+ entering around 2500 BC with significant "Steppe" ancestry. After this initial wave, migration seems to be limited as by 1600 BC the Steppe component is quite diluted in the P312+ Portugese samples, yet still present. This is due to the fact that it was male mediated migration and breeding with local EEF females in Iberia. If language is often transferred via women, it shouldn't be surprising that Iberian and Basque as adopted languages in some of the communities. It wouldn't necessarily be the same in Britain for example, since the "Steppe" level remained rather constant as the local EEF population of the earlier inhabitants was largely diminished. Thus IE speaking remained IE speaking.

Samuel Andrews said...

Oldest R1b-M269 samples in Iberia comes from two different Bell Beaker sites in Madrid. All the samples from these sites are placed in the "BK_Spain_MAD2."

They had 34% Steppe ancestry. One had 50% Steppe ancestry.

Done Deal. R1b P312 is from the Steppe and most of its expansions were done by bearers of the Bell Beaker culture.

Wish Maju could see this.

Samuel Andrews said...

The R1b P312 Beaker folk in Spain and Scotland and Poland in my opinion definitely spoke a common language and same ethno-linguistic identity. So, that if you asked them, they would explain that they belong to the same large ethnic group.

A single people group, for a definite but unknown reason, decided to spread across their entire world within a dozen generations. That's what the ancient DNA shows in my opinon.

I really wouldn't be surprised if the Beaker folk were the proto-Celts. We've learned time an dtime agian from ancient DNA that language shift doesn't happen without a genetic shift. Without evidence of major genetic shift in western Europe in the Iron age, the BEaker=Celts thing, I think makes the most sense.

Chad said...

@Ryan,

There is no movement from Iberia. That LaBrana>KO1 shift is mediated by different Megalith groups from France moving into Iberia, as the most plausible. R1b L51 is from Eastern Europe, just like Z2103. This dead horse has been beaten too much. People need to move on and accept what the data shows. How many hundreds more samples do we need for some people to get it through their head regarding Iberia and France? BB and Yamnaya likely from from the same population around 4000-4500 BCE and it wasn't anywhere with a lot of Anatolian ancestry.

Those R1bs in Iberia are V88, as I said they would be. Let's all agree that this is pretty well over. BB is IE. It would be ridiculous to start that argument it is not. BB, Yamnaya, and CWC are IE speakers. Whether or not there is something leading to Celtic in BB is the question. Likely, there is something pre-Proto-Celtic, with one region growing to dominate around 1000BCE and that splitting to P and Q Celtic.

Chad said...

I will say one thing, unrelated to Beaker. This Lombard study makes it look like Germanic migration did potentially cause some good replacement in Northern Italy. It may be on the magnitude of the Saxon deal and sure adds some questions about the amount of Visigoth ancestry in Iberians. I'm also really looking forward to something showing the Frankish contribution to Gaul as well. Comparing Gothic Poles to later Slavic ones will also be interesting.

Rob said...

@ Chad
What makes you think Lombards made big impact in north Italy?

Grey said...

AWood

"This is due to the fact that it was male mediated migration and breeding with local EEF females in Iberia. If language is often transferred via women, it shouldn't be surprising that Iberian and Basque as adopted languages in some of the communities."

If there was already a well established maritime trading network along the Atlantic coast rooted in southern Portugal then if the first BB to reach Iberia were a minority maybe they just adopted the language of the Atlantic Megalith culture - except maybe keeping their personal names (my understanding is part of the Celtic from the west idea comes from inscriptions of Celtic seeming personal names?)

Southern Portugal is/was in the Mediterranean climate zone so was suitable for the westernmost outpost of the neolithic farming package but Atlantic Megalith culture settlements further along the Atlantic coast weren't so may not have had the same population density so the balance of numbers between them and BB might have been different (especially if the AM culture settlements needed to be on the coast to supplement their grain diet with fish whereas BB cattle herders could maybe move inland as well?)

Bernd said...

@chad rohlfsen

The Frank contribution in Gaul should be quite high north of the Loire. If you look at the POBI study, the germanic component that makes up about 30-40% of SE England is shared with North France and is almost certainly related to the Franks.

Ryan said...

@Chad - "There is no movement from Iberia. That LaBrana>KO1 shift is mediated by different Megalith groups from France moving into Iberia, as the most plausible."

Agreed - that was actually my point.

"R1b L51 is from Eastern Europe, just like Z2103. This dead horse has been beaten too much. People need to move on and accept what the data shows. How many hundreds more samples do we need for some people to get it through their head regarding Iberia and France?"

My point wasn't about L51 or R1b at all actually. Not sure why you are bringing those up.

" BB and Yamnaya likely from from the same population around 4000-4500 BCE and it wasn't anywhere with a lot of Anatolian ancestry. "

Olalde excludes this possibility actually, since there are numerous Beaker even in central Europe samples that completely lack steppe ancestry.

" Let's all agree that this is pretty well over. BB is IE."

Surely you only mean that *some* BB were IE. Unless you have some explanation for the Beakers with 0 Yamnaya ancestry.

Ryan said...

Not really sure where the whole BB didn't come from anywhere with a lot of Anatolian ancestry thing comes from either. BBs are up to 90% Anatolian Neolithic according to this paper, and many of them have 0 steppe ancestry.

Chad said...

Just looking at the changes from Tuscan to Southern Italian-like natives to what these and later Northern Italians (Bergamo etc.) look like. We would need Rarecoal to be used on locals, but I think an input of 20% wouldn't be out of line. I will check some things once I do my new dataset here.

Chad said...

Ryan,

The Beakers I refer to are the full-package Beakers with the R1b and steppe ancestry. Obviously. None of the R1b males are missing steppe ancestry. Some acculturation with a rare non-steppe person among them will not cause a language shift among them. They are almost certainly IE.

I say that because you have been one that has pushed a non-IE and non-eastern origin here. And no, R1b CE Beakers do not come from a population with high Anatolian ancestry. They acquired that in the move west. It is GAC ancestry, and not Iberian. The original CA population they come from likely had very minimal Anatolian ancestry.

Samuel Andrews said...

The questions now are

-How did they pick up 40-50% non-Steppe ancestry and who did they get it from.
-Who were their Steppe ancestors and where did they live.

It's entirely possible R1b L151 originated anywhere from Crimea to Samara. But actually, I suspect they lived in the western Steppe near EEF people. They were probably already mixed before they headed west.

Samuel Andrews said...

Majority of men in Puerto Rico, Ireland, Warsaw, and some tiny villages in Pakistan carry Y DNA from Chalcolithic PIE Steppe clans.

Rob said...

@ Ryan
BB didn’t come from Iberia
It’s components can be traced to earlier periods to remedello, Balkans, Yamnaya , Majkop etc
What this study detected is the young (steppe) phase of Beaker expansion.

Anthro Survey said...

@Chad

Again, don't forget that Northern Italy likely had some mad structure in the mid 6th century AD, unlike today.

In particular, the people of the lowlands or those living close to the Po river should have been more susceptible to East Aegean(and possibly even Syrian/Early Christian during Roman times) influences. That East Aegean admixture may well have been a result of heavy Sabine/south Umbrian colonists in some regions as Simon_W suggested before and not necessarily a result of the 2200-1800BC wave or Etruscans. There may have been a caste-system of sorts in Roman Northern Italy with servile/lower castes tending towards Anatolia_BA or Samaritan-like admixture.

Then, there are those three Basque-like Colegno locals.
If you average all of the locals out, you'll get a population resembling that of the modern Bergamasque in 2D. Something also tells me that those Basque-like Piedmontese were probably more typical for the region, overall, than the Cypriot-shifted ones.

I'm not saying Northern Italy was UNaffected, but the degree is comparable to that of Slavic admixture in Albanian, not Saxon impact in Eastern Britain. Overall, Veneto should take the cake, not Piedmont. In fact there is reason to think that pre-Germanic Venetians were actually more Cypriot-shifted than an average Insubrian/Ligurian. Veneto get a lot of R-U106 and I1, too.

A bit unhappy that the PCA included no West Asian and Levantine populations. We'll confirm their exact behavior when Dave puts them through G25.

John Thomas said...

Well, the so-called 'Pictish' language, spoken in Scotland in Roman era times has always puzzled philologists. It is said to have been a Celtic tongue, a non Celtic IE tongue, or even a non IE tongue altogether.
All the more confusing since it is known by only a few place names scattered here and there in eastern Scotland.

Alberto said...

@Chad

BB and Yamnaya likely from from the same population around 4000-4500 BCE

Yes, I have to agree that this has been quite clear for a couple of years, and that R1b-L23+ came to Western Europe from the east. The question is, where would you place that populations at that time?

The latest Ukraine Neolithic HG sample we have is from around 4400 BCE. Then the earliest Yamnaya-like sample appears in the eastern edge of Ukraine at 4000 BCE, with R1a-M417 and strangely a large amount of AN ancestry (an outlier? And from where such ancestry, west or south?).

In Samara we have samples from that 4500-4000 period. Especially we have the most important sample from that Khvalynsk culture, the big chief - kind of a rock-star of the time who should have had hundreds of sons, according to some. But he's not even R1b-m269, nor really as Yamnaya-like as to be ancestral to Bell Beakers. And his mate is R1a-old (and not Yamnaya like).

So where would you place this Yamnaya-like, R1b-L23 population at 4500-4000 BCE? Which culture?

Rob said...

@ Dave/ Chad

Had a look at Myceneans with Peloponessian Neolithic

Mycenaean
"Greece_Peloponnese_N" 49.6
"Tiszapolgar_ECA" 16.65 + "Yamnaya_Samara:I0370" 9.1
"Tepecik_Ciftlik_N" 8.45/"CHG" 7.3/ "Levant_N" 3.35

Even with Greek Pelo N. it requires some additional Near Eastern
And I think the steppe component is Yanaya / Hungary Chalc., an expected route IMO.

Davidski said...

@Rob

Try the Greece_Peloponnese_N from ~5,000 BP.

Greece_Peloponnese_N I3709 3990-3804 calBCE (5130±30 BP)
Greece_Peloponnese_N I3920 3933-3706 calBCE (5000±25 BP)

Ryan said...

@Chad -

Let me put it this way - how much cultural continuity (in any direction) do you think there is between the steppe-Beakers and the non-steppe-Beakers as identified in this paper?

Also, how much cultural continuity do you think there is between the Atlantic Bronze Age, the Basques and the Vigesimal languages specific to the Atlantic fringe (proto-Celtic was decimal).

I think a lot. If you think not much that's not really something the genetics can resolve though I guess.

"The Beakers I refer to are the full-package Beakers with the R1b and steppe ancestry. Obviously."

The ones I refer to are all Beakers. :/ It's easy to say all Beakers have steppe ancestry when only Beakers with steppe ancestry are considered real, no? Question of interpretation I guess.

"None of the R1b males are missing steppe ancestry."

To be clear here - you mean none of the confirmed R1b-L51 males, right? I don't mean to be pedantic but the distinction is relevant here. (And no I'm not disputing that all the confirmed R1b-L51 males have steppe ancestry.)

"Some acculturation with a rare non-steppe person among them will not cause a language shift among them. They are almost certainly IE."

They're not rare though and that's a big part of what I'm getting at. It's half the BK_Hungary_HUN samples (4/8). It's 1/3 of the BK_Hungary_Sfu samples. It's 1/4 of the French samples and all 3 Italian samples.

"I say that because you have been one that has pushed a non-IE and non-eastern origin here."

Non IE, not non-eastern. R1b-L51's spread in Europe has always look centred around the Danube to me.

"And no, R1b CE Beakers do not come from a population with high Anatolian ancestry. They acquired that in the move west. It is GAC ancestry, and not Iberian. The original CA population they come from likely had very minimal Anatolian ancestry."

Where are these original Beakers then? It seems they only developed this cultural package after they encountered these low-steppe Beaker people you dismiss as outliers, no?

Aram said...

Chetan

Not all scholars agree that Armenian is fully Satem. There are suggestions it is partly satem. For me it is clear that Armenian was Satemaised by strong influence from I-I. This influence was so strong that initially Armenian was classified as a Iranian language.
Anyway I don't see any reason to think that Catacomb culture was 100% R1a.
There could be R1b also. And R1b is well structured in Armenia while R1a is not.

Archaelog said...

@David What do you think of the idea that it's the Bell Beakers in Poland who were the ancestors of the Balto-Slavic language and the Polish CW spoke a completely unrelated language? Highly unlikely in my opinion but it definitely has support in some circles. This question keeps nagging me nevertheless

Archaelog said...

@Aram Yes it could be R1b imo and that would place Greek and Armenian origins with a group of R1b speakers who migrated from steppes/Balkans under pressure from the incoming R1a from the north. Armenian has been called a very anomalous language in the past. It is clearly IE but its sound laws are bizarre compared to the other branches. Some have even called it an IE creole than a IE language proper. Whatever it is, we would need more historical and archaeological research to explain its origins.

Davidski said...

@Chetan

There's now a very strong archaeological and genetic trail leading from Corded Ware to Trzciniec to Balto-Slavs.

Bell Beakers in Poland, rich in R1b, can't explain Balts. The only way to explain Balts is via R1a-M417-rich Corded Ware/Trzciniec.

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

So I'm not sure which circles you're talking about? They sound like loony circles to me.

Archaelog said...

@Davidski Yes I agree David but the doubt still keeps nagging me. There are some bloggers who identify Unetice culture and the resulting Lusatian culture with Balto-Slavic and see the process as an R1b mediated shift to Indo-European among the late Corded Ware groups. (see https://indo-european.info/ie/Balto-Slavic for example)

I think and hope you are correct, because if the CW was not IE speaking, then all our models would need to be revived. In the case of Indo-Iranian, that would mean a Uralic/some other language speaking Abashevo completely took over an Indo-Iranian speaking East Yamna / Poltavka but somehow still miraculously started speaking the language of the people they just assimilated. Very contrived stuff. So I hope we never have to go there

Rob said...

@ Alberto

This will interest you, as it demonstrates what we've suspected.

Ukraine_N
Ukraine_Mesolithic 62.7 %
Iron_Gates_HG 27.7 %
EHG 6.8 %
CHG 2.6 %
LBK_EN 0.2 %

Ukraine_N_outlier
LBK_EN 96.7 %
Levant_N 3.3 %
Villabruna 0 %

As mentioned in the Paper, in Ukraine Neolithic there was a migration from the west, likely Balkans.

By the Chalcolithic, there is additional (or more complete) EEF admixture, but also CHG/EHG appearing.

Ukraine_Eneolithic
Ukraine_N 54.3 %
Balkans_N 25.2 %
EHG 10.8 %/ CHG 8.8 %

Then massive change in Yamnaya period
Yamnaya_Ukraine
EHG 55.7 %
CHG 36 %
LBK_EN 7.35 %

As for Y lineages, Mesolithic Ukraine has an I2a1, R1b-xP296, and R1a
Then significant arrival of I2a2 in Neolithic, alongside what is most likely R1b-V88 (showing longterm links with Iron Gatees).
The 2 male Eneolithics are R1b-Z2013 and R1a-M417
So is R1b-M269 from much further east ?

Davidski said...

@Chetan

The person that you're referring to appears to be very confused and perhaps even mentally unstable.

I think Mittnik et al. 2018 did a very decent job in demonstrating that Balto-Slavs originated from a migration of a Yamnaya-related people rich in R1a-M417 from the steppe to the Baltic and surrounds.

You should read this paper if you haven't already. The authors are very clear in how they view this issue.

https://www.nature.com/articles/s41467-018-02825-9

Archaelog said...

@David I have given it a read already but thanks for the clarification.

Anthro Survey said...

@Rob

Of course it will. Lefkandi I-Kastri....Anatolianization of pottery.

Alberto said...

@Rob

Yes, Ukraine Neolithic is not ancestral to Yamnaya. Yamnaya must have come from further east or south east. I was asking Chad if he has any theory about it, since the origin is neither Ukraine or Samara (and when it first appears in Eastern Ukraine, near the Don, it's R1a-M417 with a lot of AN ancestry, at 4000 BCE).

Rob said...

@ Alberto
Yes a head scratcher
I agree with what you’ve previously mentioned (& I think Ric too) that Majkop migjt have too much ANF to fit the bill
Perhaps we need to turn to mtDNA - some of that U5a and H2a :)

Davidski said...

How much Yamnaya-related or even CHG ancestry does Ukraine_Eneolithic I5884 have? You know, the new Dereivka sample belonging to R1b-M269(Z2103+).

If this guy is native to Ukraine, then Yamnaya formed via female gene flow from the southeast into local populations dominated by R1a-M417 and R1b-M269.

André de Vasconcelos said...

@Anthro

There are no Basque-like samples from Collegno. The northernmost southern samples don't even exactly cluster with modern Iberians https://image.ibb.co/dkkdfx/Screen_Hunter_2223_Feb_21_09_18.jpg
The paper seems adamant when it states:

"in Collegno it was notable that the five individuals with major southern ancestry are primarily assigned to Italy using PAA, exhibited local strontium signatures (...) migrating and settling in Collegno amongst a set of local individuals of primarily Italian origin" (page 11)


So there seems to be no indication that the southern European samples from Collegno are non-locals, although they could be descended from Roman citizens from elsewhere in the crumbling empire (or even the Apennine peninsula itself) considering the chaotic times these people lived in, but really, I'd expect people to move AWAY from North Italy, not into it if they had the choice.
Also, most of the southern samples do not cluster up north, but with Cretan-like peoples. The most northernmost southern is CL94, a male with an elaborate grave who was buried together with peoples of northern/central european autossomal profile, despite not being kindred. We don't know his origins, but considering the paper stated that "the five individuals with major southern ancestry (...) exhibited local strontium signatures" he was probably not local. He also had CEU ancestry.
CL23 clusters with modern N_Italians and was also a male, but had a simple burial near a (very) southern European individual and a few undetermined people with unknown goods. Considering he's not one of the samples with highest southern ancestry (he scores CEU) I would also accept the possibility he is non-local.
This leaves us with the Cretan-like individuals.


So, to wrap things up, and unlike I previously believed to be true, it does seem that Chad is right and the migration period brought a significant genetic change in northern Italy.

Rob said...

Yep I mentioned it
Indeed I often debated that M269 isn’t from too far East

knedokesto said...

@André de Vasconcelos

No matter who really are those who plot more to the south, there is no evidence that these southern European samples from Collegno, and even more so the Cretan-like individuals, can be taken as proxy for all northern Italians. It is just a speculation. Just as it remains a speculation the migration period brought to significant genetic change in northern Italy.

@ Anthro Survey

Even today many modern inhabitants of Emilia and Liguria get "Tuscan" as the first result on gedmatch, plotting on average as intermediate between Bergamo and the Tuscans. So that there were Tuscan-like people in the Po valley is not a big surprise. A little different case for those that go further south. In any case there is no evidence, as I have already said, that these results of Collegno can be taken as proxy for all northern Italians who lived at that time. It was not even the goal of the study.

André de Vasconcelos said...

That argument can be used for any situation, I'm sorry but it's not worth much.
This is the information we currently have, and the authors seem relatively sure they were locals. Whether it fits our percieved vision of a certain historical people or not is not very relevant for the data.

I wasn't expecting it either, but the data is there. I find it hard to believe that people in the late roman period, in a place like Northern Italy, had a large variety of autossomal profiles, and we just happened to stumble upon the southernmost ones

Ariel said...

I was testing the new Global 25 I found this, balkans BA and modern Northern italians are really similar...

Balkans_BA:I3313
"LBK_EN:I0100" 51.45
"Yamnaya_Samara:I0357" 28.95
"Armenia_ChL:I1632" 17.15
"WHG:Rochedane" 2.45
"Levant_N:I0867" 0

Italian_Bergamo:HGDP01152
"LBK_EN:I0100" 50.7
"Yamnaya_Samara:I0357" 31.75
"Armenia_ChL:I1632" 11.65
"WHG:Rochedane" 3.25
"Levant_N:I0867" 2.65

But then I saw the pca in that new paper about Langobards and the Colegno locals were Sicilian/Cretan-like, so I guess it's just a coincidence.
Btw, what do we have to think? That romans replaced the northern italian populations or the first italic tribes were very mediterranean all along.

knedokesto said...

@André de Vasconcelos

The information we have is about a specific Longobard cemetery, not about the genome of all the northern Italians of that era. The study doesn't state these results can be taken as proxy for all northern Italians who lived in the same period. This is a fact. Of course you're free to believe that migration period brought to significant genetic change in northern Italy. But it's still an unproven speculation.

Richard Rocca said...

@Rob,

You are correct and I've been touting that for years now. Remedello in Northern Italy, Rinaldone in Central Italy and Gaudo in Southern Italy all have Balkan like cultural traits and are dated to around 3400 BC. Below is a typical Remedello grave, which has many of the early Bell Beaker package traits (flexed single grave, NS orientation, archery equipment, ochre) and some burials even has steppe influences (silver t-headed pins, copper axes).

http://www.r1b.org/imgs/Remedello_T78.jpg

Therefore, a sea route is not necessary to explain eastern influences in Iberian Bell Beaker traits.

Unknown said...

Germanic is sometimes proposed to be a mixture of Kentum and Satem languages, I keep reading online. But I can't seem to find any serious article on it. Is there something valid to this theory?

Is this what you are talking about?

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

Alberto said...

Yes, that Ukraine_Eneolithic:I5884 sample is very strange. He doesn't seems to have Yamnaya admixture (no EHG or CHG), and ye he is from 2800 BCE (contemporary with Yamnaya_Ukraine or Yamnaya_Bulgaria) and has R1b-Z2103 Y-DNA.

But here we have a dilemma. Or you choose a strange explanation for this single sample, or you choose a strange explanation for all the other Yamnaya_Samara, Yamnaya_Kalmykia, Poltavka and Afanasievo samples.

It would certainly be an interesting twist in the tale if R1b-L23 came from Ukraine_Neolithic (or even from Balkan's farmers) and they started to take a massive amount of Yamnaya-like wives, to the point of replacing roughly 100% of their genetic make up.

@Rob

I agree that R1b-M269 could be from the west (Balkans, probably), but L23 is a different story. It had to happen first in an EHG-rich population or CHG-rich population or a mix of both. And at 4500-4000 BCE EHG-rich populations were in the Volga area and almost pure CHG we don't know where, but probably not in the Caucasus anymore.

Anonymous said...

@Unknown

No, not that. That is the non-IE part in Germanic.

jv said...

@ Slumbery
Thanks! Just trying to understand these Neolithic migrations into the British Isles. Everyone keeps writing about population turnovers but...... for some of us, with nearly 60% Western Gatherer ancestry and British Isles ancestry......well, where were these Hunter Gatherers hiding?

André de Vasconcelos said...

@jv

No one today has 60% WHG ancestry, be it Brits or Balts

Anonymous said...

@David

Do you have the Iron Gates samples already?

Ryan said...

@Alberto - "But here we have a dilemma. Or you choose a strange explanation for this single sample, or you choose a strange explanation for all the other Yamnaya_Samara, Yamnaya_Kalmykia, Poltavka and Afanasievo samples."

If R1b-Z2103 came into the Ukraine from a source like Iron Gates then there's no need for a strange explanation for any of them. It's just Iron Gates -> Steppe for all of those populations. Which sample is this though and from which paper? That's quite the smoking gun (that R1b-L23 is intrusive to the steppe) if true.

@Andre - "No one today has 60% WHG ancestry, be it Brits or Balts"

She may be going by old David's K7 calculator on Gedmatch, which I believe combined CHG and WHG. I score 62% there. I think Family Tree's calculator inflates WHG ancestry somehow too - it has me as 40% WHG (vs 18% Yamnaya and 42% EEF) which I find implausible too, though I'm not what exactly the problem is there.

Ryan said...

@David - "If this guy is native to Ukraine, then Yamnaya formed via female gene flow from the southeast into local populations dominated by R1a-M417 and R1b-M269." Am I right in thinking 1 sample isn't enough to use the X chromosome to check for sex bias?

Anonymous said...

@namedguest "EHG, which in turn is WHG+ANE."

No, it is very poor model, practically impossible. There have already been given why.

Anonymous said...

@supernord
I know, I know. But this is still the "official" explanation from Lazaridis.
The time frame for such event to happen is all wrong, but still, we have to work with what we have. Although people say that the WHG comes from Siberia, I think there's evidence for it being an offshot of the Basal, or even a continuity from other groups. This question is not settled, at least for me.
Are you familiar with Genetiker's Admixture graphs? They're great for seeing that:

This one here is his most recent K=11. Here, EHG and WHG show as only one blue component:
https://genetiker.files.wordpress.com/2018/02/higher-res-11.png

This here is K=14. It distinguishes between WHG (dark blue) and EHG (blue):
https://genetiker.files.wordpress.com/2018/02/all-32-14.png

M said...

There are also Northern European-like from Collegno (Italy) who are in the local strontium range from Italy. Strontium Isotope Analysis doesn't imply that someone is fully native, it implies he was born and raised there.

Schizoanalyst said...

Can someone try some models with Northern Italians using the new samples?

Tesmos said...

I5884 does has some Yamnaya-related ancestry? Although it's not that much.

Anthro Survey said...

@Dopa

I'll get to it later when I have some more time. Dn't know if using the Beaker from Parma to model their ancestry is the best idea considering that the Polada wave probably had a bigger impact, but hey, we can make a couple of safe assumptions here and there about autosomal similarity of these groups.

Davidski said...

@Alberto

My formal models do show that Yamnaya has minor EEF and WHG ancestry. This might be a signal of admixture from the Ukrainian steppes that brought R1b-M269 to the Caspian region.

@Ryan

Not much data on the X for this sample, but overall he models as having very minor or no Yamnaya-related admixture.

@epoch

Yes, I have the Iron Gates samples now.

@Tesmos

Sounds like you're looking at the ADMIXTURE bar graph from Mathieson et al. Keep in mind that this is based on a supervised analysis, which means that the samples that carry alleles ancestral to the Yamnaya component will look like they have some, or even a lot of, Yamnaya admixture. In other words, it's not that easy to work out the direction of gene flow from such analyses.

I need to take a closer look at this issue. Expect a blog post today or tomorrow.

EastPole said...

Botai horses probably have nothing to do with IE expansion. There were other domesticated horses to the west of Ural mountains.

“massive genomic turnover underpins the expansion of the horse stock that gave rise to modern domesticates, which coincides with large-scale human population expansions during the Early Bronze Age.”

http://science.sciencemag.org/content/early/2018/02/21/science.aao3297

http://www.sciencemag.org/news/2018/02/ancient-dna-upends-horse-family-tree

Davidski said...

Interesting, so Botai people aren't relevant to any of this.

They either went extinct, and/or were absorbed by pastoralists moving east from the Eastern European steppes.

Anonymous said...

@David

When and if you can spare the time, could you do:

D(Mbuti, Natufian; WHG, Iron_Gates)

Iain Mathieson mentioned this was +7 for Anatolians instead of Natufians.

https://twitter.com/mathiesoniain/status/935883881777070080

EastPole said...

@Davidski
“Interesting, so Botai people aren't relevant to any of this.”

It looks like it and it also seems to be supported by some:

“ancient human DNA that details migration and population patterns from that time. Indeed, they already have some evidence from unpublished studies.”

http://www.sciencemag.org/news/2018/02/ancient-dna-upends-horse-family-tree

Archaelog said...

@East Pole "“massive genomic turnover underpins the expansion of the horse stock that gave rise to modern domesticates, which coincides with large-scale human population expansions during the Early Bronze Age.”

Massive migration of horses from the steppes lol..

BTW, there was an article describing Botai as an stopping point for the initial Khavlynsk/Repin expansion into the east that would lead to the Afanasevo culture. I wonder if there were other older IE cultures in the east that were later replaced by the Andronovo expansion.

Anthro Survey said...

@Andre

I have to agree with knedokesto here. Indeed, we should be cautious about making assumptions as to which 6th century AD sub-populations of Northern Italy actually ended up accounting for the vast non-Germanic portion of modern Padanians' ancestry. Which cities/communes ended up demographically important in the long run and which dead ends? What was their exact structure like and how significant was it? What was the impact of re-settlers from the Emilian/Ligurian Appenines and pre-Alps? The region was more complex than people realize, I think.

There are technically only 3 highly Cretan-shifted locals and I don't deny them being born there. Fig 4 is pretty clear about this. CL121 is more like a Northern Tuscan if you look in S7.7. CL23 is like a Catalan or a Swiss Lumbard. CL94 and 47 are rather North Iberian/Occitan-like. CL31 is contaminated, so can't be too sure about that one. Sure, CL23 and 94 are not locals to the locale in question, but they could easily be locals in the sense of hailing from the surrounding Ligurian hills and having more local ancestry in the long run than those 3. Their ancestors needn't have moved there during times of turmoil, so to speak, but maybe 3 or 6 centuries earlier.

If Germanic ancestry is so great, why is R-u152 so prominent(except for Veneto)? The Hg data leans in the direction it mostly not being carried by Langobards: other R1b subclades like U106 were more important. The dearth of I2a2 across North Italy is also telling.

Anthro Survey said...

@Ariel

Yeah, we don't know the breakdown of Anatolia_BA ancestry in N. Italy by origin, atm---whether it was from the early Circum-Aegean exchange network circa 2000-1500BC(Polada era), Etruscan period or Roman period.

No doubt that Apennine Italic tribes(esp elites) carried "steppe"-related Y-Hgs and ancestry, but, yes, definitely plenty of Anatolia_BA ancestry there by the time Samnites, etc. roll around.
So, Osco-Umbrian colonists would prolly be anywhere from HGDP Tuscans to Abruzzese, I'm guessing, but not "proper" South Italians as those seem to have a different, additional stream of Near Eastern ancestry.

Davidski said...

@All

My PCA of West Eurasia with the new ancient samples. Please let me know if anything's a miss.

https://drive.google.com/file/d/1mF1X0nXe-bjz95Pojl9AGV7wjt6m-XV1/view?usp=sharing

https://drive.google.com/file/d/1k-UJeTmpiiV7dCAZmnshIgTN8ASK2BeA/view?usp=sharing

Ryan said...

@David - Could you do a version of that PCA where you highlight only the samples from the steppe and group them by time period? (IE Mesolithic, Neolithic, Eneolithic, etc). Though I guess that might not tell us much since CHG and WHG both seem to lie on PC1.

Simon_W said...

Haha, Amorim et al. believe that CEU is central Europe. Somewhat surprised they write they found it hard to distinguish that sample from GBR.

Grey said...

"BB didn’t come from Iberia"

right but *if* there is a cultural connection between maritime (Iberian) BB and riverine (central Europe) BB and no evidence of gene flow either way doesn't that imply the possibility they both originally came from the same source but diverged through taking two very divergent paths.

say for the sake of argument maritime was pre proto celtic A (which ended up sitting on a lot of copper+tin along the Atlantic coast) and riverine was pre proto celtic B which developed into proto Celtic (and ended up sitting on a lot of iron in Bohemia) so eventually B > A cos iron > bronze leaving behind some odd stone inscriptions in the previously A regions then the question that pops up in my mind is how conservative are personal names?

#

"So where would you place this Yamnaya-like, R1b-L23 population at 4500-4000 BCE?"

pure guess for fun: valley downriver of the kargaly copper field in the southern Urals

Davidski said...

@Ryan et al.

Could you do a version of that PCA where you highlight only the samples from the steppe and group them by time period? (IE Mesolithic, Neolithic, Eneolithic, etc).

The relevant PCA datasheet is located here...

https://drive.google.com/file/d/1k-UJeTmpiiV7dCAZmnshIgTN8ASK2BeA/view?usp=sharing

You can plot it any which way you like with PAST, which is freely available here...

https://folk.uio.no/ohammer/past/index.html

Chad said...

@Alberto,

I think m269 is likely Ahrensburg-related, which moved east via Swiderians, like the Kunda. L23 probably arose around the middle Volga area. Probably closer to Voldograd than Samara.

My preference has been Mikhailovka and Kemi Oba, just do to the warrior package having parallels.

However, there is another angle. L51 is supposed to arise at 3600BCE. The clock is not static though, and mutations happen much more frequently as men age. Men having kids later could cause multiple mutations to happen sooner and maybe we could see a couple mutations in a generation or two. We could then have a situation where an L23* from Yamnaya in Hungary at 3000BCE is the source for L51. The oldest known L51 is from Hungary at this point, with L151 in Poland, Germany, and England almost simultaneously. L51 was obviously never a popular group, so the first carriers are not likely to be found. We'll see what Hungarian Yamnaya looks like, and go from there.

Alberto said...

I guess that technically it is possible that R1b-L23 came from the AN-WHG part of Yamnaya. More than Ukraine Neolithic, who where hunter-gatherers, the Balkans Chalcolithic could be the source of it, and after the Copper Age collapse they moved onto the steppe and became Yamnaya-like by extreme exogamy with steppe women.

But how likely is that?

Why not just prefer the "normal" option of L23 being from a Yamnaya-like population (or at least one that is ancestral to Yamnaya in 50% or more of its genetic makeup)?

Chad said...

I wouldn't be shocked if Maykop had Z2103, and other odd R1bs, like the KA R1b.

Rob said...

@ Rocca
Yes good point, & nice Pic.

@ Chad
Ahrensurgian derives from FMG and ultimately Azilian, after “mutating” due to the Lecher see event/ YD, via the Bromnian cultural bottleneck.

Therefore I’d bet Ahrensburgian is I2a1b people, and this explains the I2a1b in east Baltic Post -Swidros

But now that there is R1b even in southern France I wouldn’t rule out your scenario

Davidski said...

@Alberto

The Dereivka Z2103 can be mostly modeled as Ukraine Neolithic and Mesolithic. The rest is basically EEF.

There was plenty of R1b in Mesolithic and Neolithic Ukraine. So both M269 and Z2103 might be from Ukraine.

Alberto said...

@Chad

My above comment was mostly addressed to Davidski and Ryan who suggested L23 being from a more western population (compared to Yamnaya). I agree with you that it should be closer to the Volga in a Yamnaya-like population.

I hope they release samples from those cultures that you mention soon (Majkop, Mikhailovka, Kemi-Oba...)

Alberto said...

@Davidski

Yes, L23 could be from Ukraine, as long as female exogamy changed > 90% of their genes. It could be from Balkans farmers too. Or from some astray Italian farmer.

But why? All the evidence suggests it's from an eastern population like Yamnaya (or one ancestral to it to a high degree).

Chad said...

Rob,

Ahrensburg is basically at the same time as Azilian, and probably more related to the previous Hamburg group. Lithics are different between the two.

Chad said...

Fishermen with harpoons vs reindeer hunters with bow and arrow.

Davidski said...

@Alberto

I honestly can't see any strong evidence that can place the origin of L23 in the Balkans.

At this point, considering the new data, it's almost certainly the Caspian steppe or the North Pontic steppe, where R1b has had a long presence in the native peoples there.

Chad said...

Rob,

Are you thinking of Federmesser? They're more similar to Azilian.

Ariel said...

Regarding northern Italy, I made this PCA with fewer samples: https://imgur.com/0MtTSit
As you can see the average of Bronze Age Balkans is very close to Bergamo, also if you had some southern ancestry and some extra steppe to the BB samples from Northern Italy (that already have a southern shift compared to the BB from the Iberian peninsula) you end up with something very close to modern Nothern Italians (that could have happened before the roman era).
Note also the that the only Balkan_IA sample that we got is already shifted south, next to Tuscans and Albanians but with less steppe.
Balkans_IA:I5769
"LBK_EN:I0100" 55.35
"Armenia_ChL:I1632" 32.1
"Yamnaya_Samara:I0357" 12.55

Rob said...

@ Chad
Yes I think Federmesser (FMG) are similar to azilian, and ancestral to Tanged points
On the other hand, I think Hamburgian would be related to Magdalenian, thus was replaced.

Folker said...

Completely agree. Celtic proper is an Iron Age language. Moreover, Bell Beakers were not present homogeneously in Continental Europe, so large parts of it were IEnized later.

Folker said...

Unlikely. Western Europe were not fully IEnized until the Roman Empire. BBs were present in numerous places, but other cultures subsisted also. There were no continuity of a general cultural network after the BBs. It is likely that several language emerged from the PIE used by BBs, most of them got extinct with diffusion of Proto Celtic on the Continent (Urnfield probably), and obviously with diffusion of Celtic with Hallstatt (on the Continent) and La Tene (Continent, the Isles...). By the elites only? Probably not. But since those populations were very similar, difficult to say at this point.

Samuel Andrews said...

@Alberto,
"Yes, L23 could be from Ukraine, as long as female exogamy changed > 90% of their genes. It could be from Balkans farmers too. Or from some astray Italian farmer."

Imo, we don't fully understand the genetic makeup of Mesolithic Ukraine/Russia. EHG may not be a legitimate subpopulation. EHG might just be on one end of the ANE-WHG spectrum in Mesolithic eastern Europe.

R1b L23 UkraineHG, plus something mostly ANE, plus lots of CHG stuff, can make Yamnaya.

Samuel Andrews said...

Ancient DNA keeps delivering surprises. Looks like, R1b1a is a Mesolithic European lineage tied to WHG. Of course, we don't know how WHG got it. All we know is the major modern R1b1a clades all descend from WHG R1b1a.

"African" R1b-V88, "Central Asian" R1b-M78, and "Middle Eastern" R1b-Z2103 all found in ancient Europeans of mostly of WHG-decent. No one would have predicted this just two years ago.

Originally we thought Yamnaya and EHG were a good representative of the source population of R1b1a1-P297. Now, it appears even Yamnaya traced a fraction of their ancestry to the original R1b1a1 carriers like Villabruna. Then consider how little of a fraction of the ancestry people in modern Spain goes back to the original R1b1a1 carriers.

Alogo said...

Ariel, that seems about right to me. You plausibly need some extra steppe-heavy (likely of northwestern type in Iberia and Italy and northeastern type in the Balkans - preference for things like Nordic_IA vs Baltic_BA in Global25 too) and extra near eastern (somewhere between Armenia_BA and Levant_BA, seems more towards the former for the Balkans Central-North Italy and towards the latter for South Italy-Sicily and Iberia) ancestry to get from Bronze-Iron Age Iberia-Italy-Balkans to their modern counterparts, in varying proportions. That's assuming we do have somewhat generally representative samples.

Northern Europe seems to have lost some of its post-CW WHGroove too since the Bronze Age. I really want to see what happened up there too and there are similar considerations about whatever possible structure might exist.

Chad said...

Rob,

Okay. I don't think that'll be the case. Hamburgian is shoulder-backed and triangles, like central Europe. Federmesser is with crescents and round-backed like Azilian. Magdalenian is similar to epiGravettian points, but also with harpoons and tanged stuff, like later in Azilian. Hopefully, the samples come soon to sort it out.

That's just how it looks from stuff I've got here.

Chad said...

David,

That m269 in Ukraine looks like someone who could be between Bell Beaker, Yamnaya, and SHG. Just looking at some numbers here. I am creating the ped files right now, then onto geno. I'll take a good look at this stuff tomorrow after work.

The oldest L51, from Hungary, is about 2/3 Yamnaya.

Chad said...

What's really tripping me out is that English Beakers look to average higher Yamnaya that CE Beakers. I'll definitely look that over and scan the paper for any stats regarding that.

Simon_W said...

The addition of the Mathieson et al. samples brought about improved fits for my maternal grandmother and my maternal half.

Maternal grandmother:
"French" 41.65
"Nordic_IA" 20.25
"Croatian_vLBA:I3313" 16.35
"Scythian_Samara" 6
"Hungary_BA:I1504" 5.65
"Anatolia_ChL" 3.85
"Halberstadt_LBA" 2.35
"Bulgarian_IA" 1.9
"Levant_BA" 1.3
"England_Roman_outlier" 0.7
"Hungary_IA" 0
"French_East" 0
"Italian_Tuscan" 0
"Mycenaean" 0
"Anatolia_BA" 0
"Balkans_MLBA:I2163" 0
"Bulgarian_EBA" 0
"Croatian_MBA" 0

distance%=0.8376

Makes sense! Now the Anatolia_ChL admixture has shrunk to a mere 3.85%, that's more realistic for Roman Age admixture in southern Germany & Switzerland. While the Balkan and Pannonian admixture amount to a considerable 23.9%, which is augmented by 6% Steppe people. This reflects the eastern origin of some of the Suebi I suppose: the Quadi who had come up the Danube rather than from the north.

And my half Italian maternal half:

"French" 31
"Croatian_MBA" 15.8
"Nordic_IA" 11.25
"Beaker_Northern_Italy" 11.25
"French_South" 9.65
"Croatian_vLBA:I3313" 8.65
"Samaritan" 8.55
"Mozabite" 3.85
"French_East" 0
"Minoan_Lasithi" 0
"Hungary_IA" 0
"Cypriot" 0
"Hungary_BA:I1504" 0
"Remedello_BA" 0
"Mycenaean" 0
"Anatolia_BA" 0
"Anatolia_ChL" 0
"Levant_BA" 0
"England_Roman_outlier" 0
"Druze" 0
"Vucedol" 0
"Bulgarian_IA" 0
"Bulgarian_MLBA:I2163" 0
"Bulgarian_EBA" 0

distance%=2.6137

So my Italian ancestry is composed of:

"French" 4.47
"Croatian_MBA" 30.74
"Beaker_Northern_Italy" 21.89
"French_South" 18.77
"Samaritan" 16.63
"Mozabite" 7.49

Surprisingly substantial Balkan admixture there! At last, because archaeologically it made sense before. So basically a mix of ancient North Italians and migrants from Dalmatia. With a strong addition of Levantine and North African admixture from the Roman era.

Rob said...

@ Chad

“Okay. I don't think that'll be the case. Hamburgian is shoulder-backed and triangles, like central Europe. Federmesser is with crescents and round-backed like Azilian. Magdalenian is similar to epiGravettian points, but also with harpoons and tanged stuff, like later in Azilian. Hopefully, the samples come soon to sort it out.

That's just how it looks from stuff I've got here.”

Well yes but Central Europe was ‘Magdalenian” when north European plain was Hamburgian

André de Vasconcelos said...

@Anthro

CL94 is actually closest to modern Portuguese (or some regional Spanish), not Occitans or Basques, as can be seen per this image taken from eupedia by Angela https://i.imgur.com/X4un04P.png

I have no good answer for your question, as I said these results also caught me by surprise as I was not expecting them at all. Maybe in the future we'll have additional samples that contradict these results, but this is the data we currently have and it's actually rather homogenous for Collegno, these locals are not all over the place.
For the time being this is what we have to work with

Rob said...

Looking at that "Eneolithic" Ukraine R1b male (from 2800 BC) at a broad level

Ukraine_Eneolithic:I5884
Ukraine_N 69.1 %
LBK_EN 20.4 %
Iron_Gates_HG 8.1 %
Balkans_N 2.3 %


The Varna outlier female (4300 BC):
Varna_outlier:ANI163
LBK_EN 47.5 %
Ukraine_N 24.5 % + EHG 14.3 % + CHG 13.5 %


Aram said...

Chetan

Archaeology, linguistics and detailed analysis of names and toponymes in cuneiforms was already suggesting the presence of pre- Armenians in Norrh Armenia since MBA. Starting from LBA (Lchashen Metsamor culture ) proto Armenian formed and during 1200 bce year event (crisis in Near East) they expanded a lot toward south and west. Both ancient DNA and the structure of Y dna confirms this scenario.

All this said I am not suggesting that Greeks are a R1b nation. We need to see their elite shaft graves to have an idea who they were.

Rob said...

@ Chad

"What's really tripping me out is that English Beakers look to average higher Yamnaya that CE Beakers. I'll definitely look that over and scan the paper for any stats regarding that."

Yep marginally but probably not historically significant. Brit Beakers just have a touch lower GAC admixture & no Balkan stuff c.f. Central Euro.

Beaker_Central_Europe
Globular_Amphora 45.75 %
Yamnaya_Samara:I0370 40.4 %
Vucedol:I3499 10.4%
..

Beaker_Britain
Yamnaya_Samara:I0370 53.75 %
Globular_Amphora 40.05 %
Narva_Lithuania 5.9 %

Rob said...

@ Aram
Do you think the Z2103 in near east could be from historic Armenians ?

Davidski said...

@epoch

Mbuti Barcin_N WHG Iron_Gates_HG 0.0092 5.171 1140577

Mbuti Natufian WHG Iron_Gates_HG 0.0056 2.222 502960

Anonymous said...

@David

Wow. That Anatolian affinity in Iron Gates HG's is responsible for that mysterious WHG Middle-Eastern affinity. There are *two* layers of WHG affinity in the Middle-East. One Anatolian and one Levant and Anatolian.

Thanks.

Anonymous said...

Is *not* responsible, that is

Anonymous said...

Ancient genomes revisit the ancestry of domestic and Przewalski’s horses

The Eneolithic Botai culture of the Central Asian steppes provides the earliest archaeological evidence for horse husbandry, ~5,500 ya, but the exact nature of early horse domestication remains controversial. We generated 42 ancient horse genomes, including 20 from Botai. Compared to 46 published ancient and modern horse genomes, our data indicate that Przewalski’s horses are the feral descendants of horses herded at Botai and not truly wild horses. All domestic horses dated from ~4,000 ya to present only show ~2.7% of Botai-related ancestry. This indicates that a massive genomic turnover underpins the expansion of the horse stock that gave rise to modern domesticates, which coincides with large-scale human population expansions during the Early Bronze Age.

http://science.sciencemag.org/content/early/2018/02/21/science.aao3297

Domesticated horses occurred from the West from the Kazakhstan.

Chad said...

Rob,
Some places, yes. Hopefully, someone is testing these groups

Aram said...

Rob

No I don't think so. Z2103 has 5 different subbranches. Only two of them probably expanded from Armenian Highlands. But even this two didn't necessarily spreaded Armenian language.
Manneans and Ancient Cappadocian languages also imho formed on the basis of Z2103.
The other one which for example has strong presence in some North Iranian groups is from Afansievo. Another one is from classic Iranians , Scytho Sarmatians. We have one from a Sarmatian burial.
And lets not forget Assyrians. At some point Z2103 probably became part of Aramean Semitic groups and they also spreaded some.

Simon_W said...

That's my Italian ancestry from Cesena:

"French" 4.47
"Croatian_MBA" 30.74
"Beaker_Northern_Italy" 21.89
"French_South" 18.77
"Samaritan" 16.63
"Mozabite" 7.49

This makes complete sense. The core is Croatian_MBA + Beaker_Northern_Italy.
If you check this map of late Chalcolithic Italy you'll see why this makes sense:

https://jpst.it/1bpio

The Tuscan Beakers were probably very similar to North Italian Beakers. And Croatian_MBA, although a little later, was probably similar to Cetina.

So I'll predict ancient central Italians like the Romans to have been like a mix of North Italian Beaker and Croatian_MBA. With stronger Hungary_BA related admixture (from the Protovillanovan) only in Tuscany and Umbria.

Swarthiness doesn't equal southern. See the map of light hair in Italy:
https://jpst.it/1bpk5

There are swarthy areas even in the north, and most of central Italy isn't very light either.

What strikes me on acient Roman busts is an often rather Dinarid character, very brachycephalic tower heads, big, beaked noses. Or they have rugged and angular faces, or with a strong positive chin or some with very broad faces like Pompey. And all these phenotypes can be found in modern central Italians as well. It's neither a particularly south Italian thing, nor Levantine looking. But I could imagine that a certain Balkan influence could bring about exactly this.

Compare Julius Caesar with Croatian actor Petar Dobric:

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

http://php88.free.fr/bdff/film/2003/0049/08/Petar%20Dobric.jpg

Of course the Nazi Nordicists were completely crazy. They speculated about Caesar having had blue eyes, because he allegedly had a "Nordic soul" even though the written sources unambiguously say he had very dark eyes.

Simon_W said...

From the Longobard paper: "Our study reveals for the first time the importance of biological relatedness in shaping social relationships." Fascinating. Back in these days, blood still mattered, at least for these Germanics.

Andrzejewski said...

Northern Italians are mostly descendants of the Cis-Alpine Gauls but mostly the Germanic Longobards. This is what distinguishes them from Central and Southern Italians, let alone island populations like Sicilians or Sardinians.

Andrzejewski said...

Further proof that Bell Beaker did *not* come from Spain:

4,500 years ago there was a Steppe-based invasion which wiped out practically every single male in the Iberian peninsula. Does it explain why Basque people are predominantly R1b-L23? Could be!

https://elpais.com/elpais/2018/10/03/inenglish/1538568010_930565.html