search this blog

Friday, September 30, 2016

Male-dominated conquest of Europe by Bronze Age steppe pastoralists


Just in at bioRxiv:

Abstract: Dramatic events in human prehistory, such as the spread of agriculture to Europe from Anatolia and the Late Neolithic/Bronze Age (LNBA) migration from the Pontic-Caspian steppe, can be investigated using patterns of genetic variation among the people that lived in those times. In particular, studies of differing female and male demographic histories on the basis of ancient genomes can provide information about complexities of social structures and cultural interactions in prehistoric populations. We use a mechanistic admixture model to compare the sex-specifically-inherited X chromosome to the autosomes in 20 early Neolithic and 16 LNBA human remains. Contrary to previous hypotheses suggested by the patrilocality of many agricultural populations, we find no evidence of sex-biased admixture during the migration that spread farming across Europe during the early Neolithic. For later migrations from the Pontic steppe during the LNBA, however, we estimate a dramatic male bias, with ~5-14 migrating males for every migrating female. We find evidence of ongoing, primarily male, migration from the steppe to central Europe over a period of multiple generations, with a level of sex bias that excludes a pulse migration during a single generation. The contrasting patterns of sex-specific migration during these two migrations suggest a view of differing cultural histories in which the Neolithic transition was driven by mass migration of both males and females in roughly equal numbers, perhaps whole families, whereas the later Bronze Age migration and cultural shift were instead driven by male migration, potentially connected to new technology and conquest.



Goldberg et al., Familial migration of the Neolithic contrasts massive male migration during Bronze Age in Europe inferred from ancient X chromosomes, bioRxiv, Posted September 30, 2016, doi: http://dx.doi.org/10.1101/078360

Update 15/03/2017: "Failure to Replicate a Genetic Signal for Sex Bias in the Steppe Migration into Central Europe"

84 comments:

MaxT said...

I guess Gimbutas and David Anthony were right. Indo-Europeanization was basically driven by male migration and elite recruitment.

Grey said...

neolithic more like north america
I-E more like south america

Colin Welling said...

sounds iffy given the huge influx of yamnaya like genes into central europe.

Davidski said...

Looks like there were multiple waves of population movements from the steppe into Central Europe with three main Y-HGs involved: R1a-M417, R1b-L51 and also I2 (ancestors of Unetice), probably in that order.

Women also came, but usually not very many in relative terms.

Matt said...

Cool stuff. Although the mtdna and Y-dna have said similar things already, the autosome vs X should be less vulnerable to founder effects (e.g. after a non-sex biased admixture, drift could take a Y or mtdna from one population up to high frequency).

Few thoughts:

1) So for their process, is there more going on than:
a) Use ADMIXTURE to estimate ancestry (see page 6)
b) Separate out X chromosome from autosome
c) Compare ratios and compute likely relative population sizes that could give rise to those ratios.
?

If so perhaps if a similar thing could be done instead with qpAdm (this method estimates ancestry proportions from candidate populations via their f4 statistic correlations with outgroups).

Run qpAdm proportions for Europeans for a CHG + EHG + Anatolia_EN + WHG model or a Levant N + WHG + Iran N + EHG model, over the separate X and autosome, then compare.

2) Also interesting if this is doable for CHG+EHG into Yamnaya.

On a tangent, I'm not sure I wonder if that would perhaps introduce a potential issue into their Fst analysis. If we're comparing the Fst on X (FstX) and Fst on autosome (FstA) between Bronze Age and Early Neolithic Europeans, because if the gene flow into Yamnaya was sex biased towards male EHG, female CHG, then there would already be some excess of relatedness between Yamnaya and Early Neolithic Europeans on the X compared to autosome (via relatedness between CHG and Anatolian farmers).

Another tangent, with Table 1, it doesn't look like BA-HG FstX is not in excess of BA-HG FstA in the same way as the ratio differs strongly from 1 for BA-SP. That would be a bit surprising to me if I'm reading it correctly, because I would think that BA should be more differentiated against HG on the X chromosome than autosome, if their HG related ancestry ultimately comes more from males (via sex biased admixture in Yamnaya, then more sex biased admixture in BA).

Ariel said...

"Looks like there were multiple waves of population movements from the steppe into Central Europe with three main Y-HGs involved: R1a-M417, R1b-L51 and also I2 (ancestors of Unetice), probably in that order."

Just a remainder that you dissed me and others many times by saying that no IE or EHG-rich bronze age groups had I2 as their main hapologroup. I still find the possibility of eastern european groups with high amount of ANE and I2 very compelling, at the end of the day we have already SHG. The situations of eastern Europe before the bronze age might be similar to anatolia in the neolithic, some farmers being more WHG shifted, whereas others have more natufians. And btw there are still the same problem with the R1b-51 phylogenesis.

Davidski said...

@Ariel

No idea what you're blathering about there.

Obviously, Unetice is both rich in EHG and I2, and I knew that since the genomes were published.

But I2 in Eastern Europe ultimately comes from WHG, because there was no EHG west of Russia until the Corded Ware migration brought it there.

Ariel said...

"But I2 in Eastern Europe ultimately comes from WHG, because there was no EHG west of Russia until the Corded Ware migration brought it there."

SHG exist. It's like a freudian delusion. SHG exist. They were ANE and WHG. There is already a prove that you are wrong. There was no EHG west of Russia!? Total delusion. A lot of the EHG/ANE in eastern europe came from non IE groups that were scattered around the area before the bronze age.

Rob said...

Good to further confirm
But what did all the steppe women do ? There must have been just thousands roaming the steppe

Davidski said...

SHG exist. It's like a freudian delusion. SHG exist. They were ANE and WHG. There is already a prove that you are wrong.

Sweden isn't in Eastern Europe, and not even Swedes have much SHG.

There was no EHG west of Russia!? Total delusion. A lot of the EHG/ANE in eastern europe came from non IE groups that were scattered around the area before the bronze age.

Very funny. Your face will drop when you see what Southeast Baltic foragers are like.

Matt said...

Some other things, thinking about this paper could suggest to look at these questions:


1)
- Take separate autosome and X chromosome datasets for various populations
- Run ADMIXTURE clustering for both
- Compare ADMIXTURE clustering for both sets

See where they differ and see which populations form clusters in X vs autosome. You might see populations form tight clusters in X where they don't in the autosome, indicating a bottleneck of women, or populations form clusters earlier in autosome, indicating bottleneck of men.

2) Compute FST matrices for autosome (FstA) and X chromosome (FstX) datasets. Run a ratio between the two matrices FstA/FstX. Then you could run an MDS on that it would place the population together which were dispropotionately connected by male / female geneflow, or by neither/both in equal amounts.

(Provided you run them properly, fst matrices in MDS visually give the approximation of the same genetic distances captured by PCA, etc - http://i.imgur.com/SMOeUo9.png / http://imgur.com/1cQ8NkA / http://i.imgur.com/yZrHboq.png).

Karl_K said...

@Rob

"But what did all the steppe women do ? There must have been just thousands roaming the steppe"

This result doesn't really mean that mostly men migrated to Europe.

It means that there was a dramatic sex bias in mating between the ancestry groups within Europe.

So if each man from the steppe took a few additional European Neolithic Farmer wives in addition to the wife they brought with them, it would still result in the same sex bias.

There is no need for lonely steppe women anywhere. There is the question of the Eurooean Farmer men, though, who basically died out eventually.

postneo said...

Actually they many not have not accounted for one scenario.

European mtdna and female mediated adna had already diffused in to the steppe by EBA while the steppe was comparatively depopulated.

So there was not much difference of steppe incursion on the female side.

MaxT said...

@Karl_K

"There is the question of the Eurooean Farmer men, though, who basically died out eventually."

Laughable, that's not true. They din't "die" out, EEF is still dominant ancestry of Europe. See (Lazaridis et al 2014)



Shaikorth said...

@Matt

We could go into deeper detail if someone had the time: run Chromopainter on the EGDP or SGDP set, separately for autosomes and X, and then Globetrotter on the results, or Broushaki et al. style fits. If there are differences we'd see their magnitude and direction clearly from that.

Grey said...

Rob

"But what did all the steppe women do ? There must have been just thousands roaming the steppe"

pastoralist polygyny maybe (both on the steppe and in the expansion)

Karl_K said...

G2a haplogroup Y chromosomes were at a frequency of about 80% in EEF men.

Now they are about 5% in Southern Europe, and very rare in the north.

I'd say that is a pretty dramatic difference.

MaxT said...

@Karl_K

Based on aDNA, EEF is dominant genetic ancestry of Europe. Y-DNA does not say much about admixtures.

Alberto said...

@Matt

Yes, interesting experiment, but also leaves us wanting to see some of those suggestions that you made. More tests would reveal how trustworthy this method is, because here it leaves a few questions.

From Table 1, indeed it looks strange that HG admixture into BA is slightly female biased, though it could be because of WHG and not EHG (WHG coming high via CE females)? But another question would be if BA-SP ratio (Q) is 0.237, deviating highly from the expected 0.75, and meaning a biased male mediated SP admixture in BA, then you would expect the opposite in BA-AF ratio (very female biased admixture of AF into BA), but the value of Q is 0.826, quite close to neutral. Including EHG and CHG for SP would have been very informative to see how it all works out.

From Table S1, using ADMIXTURE, also some strange results. Like samples having 100% SP admixture in the autosomal level but 0% on the X chromosome (the 100% SP samples seem to be Corded Ware and not Yamnaya, but even then a Yamnaya male and CE female would not produce a 100% Corded Ware son, since it would have an excess of CE admixture). Averages seem to work ok, though.

If I understand correctly, the whole thing is based on the X chromosome being female biased (being on average 75% maternally and 25% paternally inherited, unlike the non-sex chromosomes that are 50-50). But then is this not doable using the Y Chromosome -limited to male samples, obviously- that is 100% inherited from the paternal side and therefor it would have higher confidence? Too low number of SNPs? (but even then, low number of high confidence SNPs seem preferable to larger number of lower confidence ones? But maybe I'm missing something about this all).


@Karl_K

That's one of the reasons why some time ago I said that exogamy alone could not fully explain this phenomenon, but polygamy could. Still, scarce evidence for polygamy if it was so widespread.


@Colin Welling

Iffy, yes. I think that in spite of the interesting technical side of it, what the paper and its conclusions highlight is what a poor model this is and that we're missing important data to explain things in a more reasonable way than male dominance based on horses and chariots.

We'd need to explain too why all this sex biased migration between the steppe and Germany completely disappears between Germany and Ireland, since Irish BA samples are basically identical to German BB samples. A radical change in strategy (males migrating alone, or being "exogamously" polygamous to migrating in whole families and being monogamous/endogamous) but having the very same results over the local populations?

Chad Rohlfsen said...

EEF is above 50% for only SW Europeans. Rathlin1 is an early migrant without local MN admixture, which you need to make the Irish.

Samuel Andrews said...

@Chad,

Maxt is right though that out of all the ancestors EEF is the most important for all Europeans, except small exceptions like Saami and some by the Ural mountains and maybe Balts.

Matt said...

@ Alberto, re: Table 1, I think all the numbers in Table 1 show some form of greater relatedness on autosome than X chromosome, which would in theory tend to indicate flow all these populations has tended to be homogenised more by male movements, while the female side generally showing slightly more more persistence of ancient variation and drift. (Might make sense, if human history has generally tended to be more matrilocal at the broad scale; be interesting to see if that's a general rule).

(I'm not sure reading your post but in Table 1 these are basically Fst scores between the populations for autosome and X chromosome, and then the ratio between them, so when you have like, "CE-HG: FSTA 0.0526, FSTX 0.574, Q 0.911" that's not implying an ancestry proportion on the X or anything it's just like the ratio of FSTA to FSTX is 0.911. Apologies if you know all this.)

IIUC, the numbers would have to go above 1 to show a ratio of more similarity on X between two populations compared to the autosome (i.e. female geneflow is homogenising between populations, as shown by reduction of differentiation on X compared to autosome).

Though maybe the X chromosome could just be unusually structured between populations for natural selection reasons as a wrinkle to that.

It does seem unusual like you note that the BA-AF pattern is not just the inverse of the BA-SP, so where BA-SP shows more differentiation between Steppe and Bronze Age on X, less of autosome, there's not a pattern where BA-AF shows differentiation where the X chromosome is more similar between them than their autosomes, and in fact mildly the opposite. Can't explain how that one works.

But then is this not doable using the Y Chromosome -limited to male samples, obviously- that is 100% inherited from the paternal side and therefor it would have higher confidence? Too low number of SNPs? (but even then, low number of high confidence SNPs seem preferable to larger number of lower confidence ones? But maybe I'm missing something about this all).

In a sense you're right and I think people have made comparisons of the Fsts between the y-chromosome for two populations and their autosomes. Like if Basques tend to be much lower differentiation from Irish in their y-chromosomes than in their autosomes, seems like male biased geneflow is happening there, right?

But I think it's not as straightforward to interpret because the y-chromosome is non-recombining. With the y-chromosome, it'll get copied from father to son, and it'll accumulate random mutations at random times, and sometimes variants of that y-chromosome could go extinct with enough time, if they pick up a bad mutation. But they basically don't recombine, and that's why two members of a population can hold y-chromosome variants that have totally different mutational history, for like, 20,000 years or something.

Whereas with the X because it's a recombining region where it's present 2/3 in females and in females is where the recombination happens, it's in some ways easier maybe to interpret in comparison against other recombining regions (the autosome). For one should be less liable to founder effect, and repeated founder effects through history, than a non-recombining region. For another, I don't think ADMIXTURE would really work or make sense on the y-chromosome, which as it's non-combining can be defined by a tree of mutations (but you need to sequence the whole chromosome to get all the mutations).

Chad Rohlfsen said...

That's not really the case. Download Admixtools and see for yourself.

Samuel Andrews said...

@Chad,

My most important I don't mean over 50%.

Yamnaya was roughly 50-60% EHG, 10-15% EEF, 25-40% CHGish and MN was roughly 20-30% WHG and 70-80% EEF. Lets assume Northern Europeans are 50/50 Yamnaya/MN, then EEF is their most important ancestor out of the four. Even if Yamnaya wasn't 10-15% EEF, it is still the most important ancestor. The EEF numbers, if you include EEFish ancestry from Western Asia, for Southern Europe are usually over 50%.

Matt said...

@ Shaikorth, yeah, I think someone could extend as a method via Chromopainter / Globetrotter methods as well. Albeit with the whole computationally intensive nature of Chromopainter to deal with for that.

Tobus said...

@MaxT: Based on aDNA, EEF is dominant genetic ancestry of Europe. Y-DNA does not say much about admixtures.

But it *does* say a lot about where that ancestry comes from. Women provide half the aDNA, so if Y-DNA is primarily Steppe then it's highly likely that it's the EEF women who provided the bulk of the EEF ancestry, while the EEF men essentially "died out", like Karl suggests.

ArtemisVentus said...

FunnelBeaker was an I2 culture, Gokhem 9 is I2. There is direct continuity from FunnelBeaker to Unetice to Urnfield (Lichtenstein cave samples) to even related Remedello. No replacement there, just migrants from a culture which wasn't the Yamnaya culture settling in the surrounding area and probably trading cows for FunnelBeaker wives. Learning Indo European language and customs from the FunnelBeakers, since as we know, every aspect of Bronze Age and Indo European culture is found thousands of years earlier in Neolithic Europe than anywhere else.

With the exception of riding the horse.

The cart, the wheel, the language, animal domestication, all this from the Neolithic.

But the first guy who sat on the thing and rode it around without a saddle may have lived on the Steppe.

At the time it had no practical application because the horses were too small, hence 1000 years later, after the Yamnaya, the chariot was invented.

https://www.academia.edu/12858044/Corded_Ware_Coastal_Communities_Using_ceramic_analysis_to_reconstruct_third_millennium_BC_societies_in_the_Netherlands


------------------------------------
6.4 Corded Ware Culture: A new narrative

“a wandering race of hunters and pastoralist. They appear as pre-eminently
martial folk: yet, whether by plunder or trade, they were able to secure products of
distant lands…These conquering battle-axe wielders exerted a profound influence
wherever they went” (Childe 1929: p.158).

The quote by Childe has proven inconsistent with the observed situation, and
should therefore be altered to read:

“communities of farmers, cattle breeders, fishers, hunters and gatherers. They
appear as pre-eminently of local origin: yet exchanged products and ideas with
those from distant lands…These beaker owners had regional traditions as well as
a desire to locally express supra-regional affinities.”
------------------------------------
and more
------------------------------------

An analysis of the differences between the Corded Ware Culture and preceding
cultures has led some authors to postulate a discontinuity (e.g. Vander Linden,
2001a). Others have stressed instead that several elements show continuity
(Sørensen, 1997, Furholt, 2003a, Hübner, 2005, Larsson, 2009). For example,
Vander Linden (2001a) has pointed to the differences between the Funnel Beaker
Culture and the Corded Ware Culture, but Sørensen (1997: p.228) and Hübner
(2005: p.753-754) state that important elements of the Corded Ware Culture,
such as battle axes and single graves, were already present during the Funnel
Beaker period. The amphorae of the preceding Globular Amphora Culture are
seen as direct precursors to the amphorae of the Corded Ware Culture in the
same region (Larsson, 2009: p.61). Particularly the Strichbündel amphorae have
been interpreted as a development of the Złota-group out of the preceding Kugel
amphorae (Furholt, 2003b: p.19).

-------------------------------------------


martin said...

Vous suggèrez que les mâles se reproduisaient entre eux ? Certainement que après avoir passé quelques générations sans femelles, ils se jettent sur les femmes indigènes et les indo-européennisent))

Ariel said...

Reich in his last conference was saying that yersinia pestis was found in the steppe but not in neolithic Europe, and the bronze-age population replacement could have been similar to the one in the americas.

Rob said...

Ariele

I'm not Convinced yet about plague. They presented a handful of dates which don't actually fit any scenario
I recall that the earliest date (? 2200 BC) for plague found in Central Europe was long after the Neolithic demise

A handful of typed Yersinia iant "the plague"
It needs critical mass and propper causality - chronology sequence

Rob said...

Artemis

We need to be specific here. For ex:

* the 'continuity' between TRB and CWC, could be simply the idea of the beaker; not all population elements .

* Similarly, the CWC seems to have actively displaced the preceding GAC in an oppositional manner. There is only 'fusion' in Zlota, as you point out. And here is also were earliest CWC ceramics appear in Poland, perhaps via the earliest pastoralists arriving in Hungary

* the terminal Eneolithic - EBA seems to be a dearth in horse symbolism. In Yamnaya but a handful of horse burials have been found, virtually none in CWC.
The advantage of CWC was simply economic- a specialised cattle pastoralism but also adaptable to hunting
Yamnaya - same but with the added benefits of being procurers & traders of Caucasian metallurgy.
They all probably mobilised with wagons and Ox rather than horses

Perhaps these advantages, and their special social code, enabled them to simply outcompete other males in Northern Europe, which at the time consisted of scattered self-sufficient hamlets (as per TRB) after an apparent cycle of mass-violence (late LBK).

We obviously need more aDNA data for Italy Iberia and Balkans before we can explain social processes there



Gioiello said...

@ Rob
"We obviously need more aDNA data for Italy Iberia and Balkans before we can explain social processes there"

Of course it seems that the Stanfordists (the authors come all from there and from the same linked group) are more interested in making their pet theories rather that to test these places, which would be able to resolve the question very easily...

Gioiello said...

rather than... I don't understand why my PC Always writes "that" when I write "than"!

Samuel Andrews said...

https://3.bp.blogspot.com/-DaalKaB8WZM/V-8Zbik2-YI/AAAAAAAAE6s/GNZJ8X_zAUUfutZj_qxg0FYm3J_IryDEQCLcB/s450/Goldberg_Fig_3.png

The third image is deceiving. Nothing in the paper suggests 5-14 times more Steppe men migrated than women, all it suggests is admixture between Steppe and MN the male was Steppe 5-14 times often than female. There's a strong presence of Steppe mtDNA in LNBA Central Europe(see link below). IMO Steppe populations made settlements with whole populations(male, female, all ages), that could of been mostly young men, in Europe

https://docs.google.com/spreadsheets/d/1GToouHMsUnoUpGSY91OIMA3bjc86bhNRSzknZ_BKLYE/edit#gid=462265126

Quote from paper,
"We find evidence of ongoing, primarily male, migration from the steppe to central Europe over a period of multiple generations"

Or more likely there were continuing waves of MN women marrying Steppe men and moving into their husband's villages.

Gioiello said...

@ Samuel

I didn't read the paper of course, I am almost tired of these Stanfordists, but from what you say it seems that the migration was to Central Europe, not to Western Europe. It is all to demonstrate that they carried R-L51 and subclades etc etc

Gioiello said...

I'd say that people of Samara were satem IE speaking, thus they could be the westernmost slavic people, but not any centum speaking ones.

Gioiello said...

Of course there is no limit to madness. Look at how these people, arbitrarily segmenting the Stele of Lemnos, undoubtedly belonging to the Etruscan stock, say that it is on old form of Slav:

Ambrozic, A. (2005). The “Warrior” Stele from Lemnos. In Proceedings of the 10th International Topical Conference Ancient Settlers of Europe (pp. 107-120). Ljubljana: Založništvo Jutro. http://www.korenine.si/zborniki/zbornik05/ambrozic_warrior.pdf

But where has Nirjhar007 gone?

Davidski said...

@Gioiello

How is it that you still don't understand that there was a massive population shift in Italy during and/or after the Bronze Age?

We have Oetzi the Iceman and we have the Remedello Copper/Bronze Age samples, and they're obviously not like any modern Italians except some Sardinians.

Modern Italians carry genetic components that Oetzi and the Remedellos lack, and these components come from the Russian steppes and the Near East.

This is a fact. It can be easily demonstrated, and it's not something that you will see reversed when more ancient DNA from Italy is published.

You'll have to accept at some point that modern Italians are not like Neolithic or Bronze Age Italians, and the reason for this are multiple migrations to Italy from the east.

Time to get used to this idea and accept it.

Gioiello said...

@ Davidski

I thank you for your explanation, and hope we may continue to discuss freely, but may I say that I am not convinced of what you say?
1) The problem I examined from the beginning was the origin of hg. R1b above all. You know that there was a theory that said that R1b came from Middle East. I have written from ten years that it seemed to me that weren't true: the oldest subclades of R1b were in Italy: R-V88, both in the oldest subclade M18 and V35, and the oldest one: Sexton in the Isles and Marchesi in Italy etc.
So R1b1-L389+, R-M335 etc, till R-M269* and R-L51*. You said on 1 may 2016 that R1b wasn't in Italy, but 2 May was published Villabruna.
2) After that the Middle Easterner origin failed (no R1b has been found so far in Middle East except recent subclade from Southern Russia as F38), the Stanfordists wed Gimbutas, Anthony etc, and became Kurganists from the Levantinists they were before. I have written a lot about R-L23 (which is also my haplogroup) and even R-M73 etc.
3) May I desire that more samples from aDNA from Italy beyond these 5 samples (Oetzi, Remedello and Villabruna) are tested at least at the level of the tens and hundreds of other countries?
4) I quoted a paper on linguistics because in that field I may demonstrate how the truth is manipulated more than with the autosome...

Davidski said...

It's not possible to manipulate anything.

This stuff is easy to test. I could teach a 12 year old to do this properly with a bit of time.

Italians are not the same people that they were during the Copper Age. A million genomes from Copper Age Italy won't change this.

zardos said...

If these results get verified by additional studies and so far everything seems to support the conclusion, the enslavement or killing of large parts of the Neolithic males and the forced marriage or enslavement of a large part of the Neolithic women is the only plausible explanation. You don't get such results over such a wide range of territories in such a relatively short time otherwise. Simple exogamous marriage patterns can't explain it.
The Neolithic male lines became reduced to almost nothing in a couple of generations, that's dramatic and explains the profound cultural change.
Polygamy is a necessity for that pattern, but there is not just monogamy and polygamy. There is also the relationship of a free or noble man to unfree women and slaves. That is not the same as a legitimate marriage to a free woman, but it doesn't change the genetic outcome.

Rob said...

@ Gio

By "more aDNA from Italy", I meant to get a complete picture of the timing and character of the changes. Otherwise I very much agree with everyone else that there was a large change throughout Europe in the LN-EBA .
And no, i don't think evidence points to PIE originating in Italy

Gioiello said...

@ Rob

I think having demonstrated that Western European R-L23 doesn't derive from the samples found in Yamnaya, except perhaps R-CTS7763. Certainly Western European R-M73 doesn't derive from Yamnaya, because in Asia and eastern Europe we have the subclade R-M478 but not so far R-M73* as in Western Europe. Even the mt H32, found in Iran and Middle Easterner aDNA, very likely migrated up there from Western Europe, lacking there the ancestor H-152C! and all the sister clades present in Western Europe, above all in Italy, and much more.
About Indo-European languages I said that they evolved amongst the clan Villabruna, of course with many vicissitudes that we'll understand next.
Certainly more and more linguists think that Celt expanded from Italy to Iberia and from there through the Atlantic coast Northward...

Richard Rocca said...

I uploaded RISE489 - Remedello Culture, Italy (2908-2578 BC) Gedmatch ID T135721 a couple of weeks ago. The best fit population mix is as follows:

71.6% Sardinian + 28.4% Basque_Spanish @ 4.88

Of course that is of no surprise to anyone, as the link between I-M26, Paleo-Sardinian and Proto-Basque has already been established. Since Sardinia has always been a recipient of geneflow from Tuscany, the probability of Tuscan Copper Age results being full of L23 and the missing components found in modern Italians (like EHG) is nil. Then there is Otzi, who's G-L91 is most common in northern Sardinia and southern Corsica. Where have we seen a Neolithic mix of I-M26 and G2a before?...the EpiCardial Avellaner (Spain) and Cardial-derived Treilles (S.France) samples all have G2a and/or I-M26. So, these are very good proxies for Italian Neolithics as well, since the Cardial Neolithic went from Italy to France and Spain. While additional Copper Age Italian samples are more than welcome, expecting that they will change everything we know is completely unrealistic. Citing irrelevant mtDNA data from a male dominated period and minor dead-end R1b samples that had no impact on modern Europeans is grasping at straws.

Grey said...

Ariel

"Reich in his last conference was saying that yersinia pestis was found in the steppe but not in neolithic Europe, and the bronze-age population replacement could have been similar to the one in the americas."

I can believe in an epidemic of some kind but if pestis is carried by marmots i don't see a mechanism for it spreading with IE unless they carried pet marmots around with them.

"Go for the eyes Boo - go for the eyes!"

I'd have thought a disease from domesticated animals would be more likely.

.

Rusty on this but isn't the way the Akkadians are supposed to have taken over Sumer is they initially spread as herders into and along the marginal land within the settled Sumerian farming population creating an initially symbiotic relationship with lots of bilingual inscriptions and such?

(A model like this is interesting re. the Hittites.)

In a situation like that if some major epidemic affected the herders the farmers still couldn't move onto the herder's marginal land whereas if a major epidemic hit the farmers the herders could move onto the vacated/weakened farmer's land.

(nb not arguing against the warrior elite thing as the final chapter - just not convinced the earliest chapters were the same)

sds said...

David, what kind of time frame are we looking at? Or better put, what kind of time frame was between these three successive migrations?

Davidski said...

Migrations associated with Corded Ware, eastern Bell Beaker and then Unetice. No idea about precise time frames.

Samuel Andrews said...

@Richard Rocca,

Your overall point is correct; Italy faced migration after the Copper age. However, nothing suggests Remedello has a connection to Basaque. If you compare Remedello's results in ADMIXTURE tests based on modern DNA to other CA/MN Europeans you'll see they come out as a Basque/Sardinian mix as well. The reason is Sardinians score lower in WHG-heavy components and Basque's 20-30% Steppe ancestry doesn't differentiate them from Neolithic Europeans because WHG(raises score in Northern components) but their Steppe is too low for them to score anything in West Asian components which would differentiate them from Neolithic Europeans.

Richard Rocca said...

@samuel Andrews,

That doesn't mean that some of the same shared pre-Bell Beaker autosomal components didn't survive in both locations. And yet the only two places I-M26 survived is Sardinia and north east Iberia including Basque Country. In continental Italy, it was only found in 4 of a whopping 800 men. And there is a link between Paleo-Sardinian and Proto-Basque. The reference is easy enough to find.

Davidski said...

By the way, it's very difficult to accurately check whether Yamnaya X chromosomes are enriched for West Asian ancestry relative to the autosomes.

I did some Admixture runs in the past and they were inconclusive. A bit messy actually.

I also did some double outgroup D-stats. Graphs based on these stats suggested that Yamnaya X chromosomes do look more Near Eastern on average, while the autosomes are generally more Eastern_HG, but that was only using ~30K SNPs, and I'm not confident with the outcome.

Don't think I can use Chromopainter with so many missing markers in the ancient samples. I'm not even sure it's possible to run Chromopainter with data from the X.

So I think I'll leave this to someone else, or just wait for more ancient steppe and Caucasus samples to see what the uniparental data show.

Karl_K said...

@Rob

"I too wouldn't link M26 and Sardinian with Basque . The cultures show no links"

Whether or not the cultures show links, they have an obvious genetic link, so they had some kind of major genetic input from a similar (or identical) source population.

Gioiello said...


@ Richard Rocca

That I wrote on YFull FB page:
Has Sexton arrived?
R-V88 FGC21059/Y7781 * Z30230/Y7770 * V88/PF6279+72 SNPs formed 17300 ybp, TMRCA 10200 ybp info

R-V88*
id:YF07201 new

And for your "Bildung":
The R-M269-PF7562* from Caucasus (he is a Laz living in Turkey) is a first migrated from the Italian Refugium and has nothing to do with the subclades from PF7563, centred in Italy.

Ryan said...

Aren't Sardinia and the Basques linked through the whole Megalithic phenomenon?

I find it interesting that the areas often pointed too as under Vasconic influence are places where Y haplogroups found in WHG samples predominate (ie R1b and I2). Though I suppose it's easy enough to look at a map and see a pattern where none exists.

Why is it taking so long to get samples from Bell Beaker men in Spain?

Rob said...

@ KarlK

As Rich wrote, but did not intend :
"The ... Sardinians score lower in WHG-heavy components and Basque's 20-30% Steppe ancestry doesn't differentiate them from Neolithic Europeans because WHG(raises score in Northern components) but their Steppe is too low for them to score anything"

Are exactly the mild but very significant differemces that will bring forth linguistic implications

There's no archaeological link between copper age Franco-Cantabria and Sicily (over and above a broad and archaic European middle Neolithic)

Atriðr said...

@Gioiello
About Indo-European languages I said that they evolved amongst the clan Villabruna, of course with many vicissitudes that we'll understand next.

We understand plenty on linguistics already, but as it takes too long for most to learn linguistics (let alone languages), most don't even try.

I'm curious if you base this thought on a prerequisite for your ideas, or from a linguistic departure point.

In a world of infinite possibilities, everything is worth considering. In a world of probabilities, the chances of Indo-European languages starting with Villabruna - that's a challenge. Where do you place the clan if you do not refer to the location?

Gioiello said...

@ Atriðr

1) I give for demonstrated that there has been an "Italian Refugium" (or call it "Alpine Refugium" etc.), and clearly another Refugium in Southern Italy, Sicily above all (mt hg. HV is a witness also from aDNA from Favignana).
I am sure that in this refugium there were not only R1b, possible also R1a, certainly hg. I and I think also J (I have written that I think that the oldest J1 will be found in Italy, older than Satsurblia). Of course some hg. G are very old in Italy, and, having found them in Middle East, doesn't mean anything, having been a migration after the Younger Dryas from Italy in all directions.
2) IE has two "subclades": centum and satem. Centum is older. I think that satem languages may have been those of South Russia, perhaps linked to R-L23, but I think that L23 came from West. After the expansion to Asia and perhaps also this back migration to Central Europe whose these Stanforfdists spoke about the chromosome X. They could be the "Slavs" of Central Europe, from where they expanded everywhere more recently.
3) Italy has the most varied presence of IE languages. It is written also from Mallory&Adams. Schrijver, a great Celtist, thinks that Celts expanded from Italy to Iberia and from there through the Atlantic route Northward. Lepontian wouldn't be a Celt come back, but a primitive form of Celt, for not speaking of Ligurian.
4) I thought that Etruscan, that Alfredo Trombetti defined intermediate between Caucasian and IE, was from the Alpine zone, where we find Rhaetian, but recently, answering just to a paper of the same Schrijver who hypothesized that the language of the first agriculturalists from Anatolia were Hattic, I said that it would have been more likely to think to Etruscan, which could be their language.
5) Mazzau wrote about a link of Etruscan and Sardinian, and the link, that many found now between Sardinians and Basques at some common hgs., may be due to the migration of the first agriculturalists from Italy 7500 years ago (Zilhao).
6) When we'll get a more defined picture of the hgs, I think we'll get also a more precise picture of the languages.

Karl_K said...

@Rob

"Are exactly the mild but very significant differemces that will bring forth linguistic implications"

I am not sure there is any real basis for this conclusion. We basically have no idea how the Basque or pre-Indo European Sardinians came to have their languages, or how their linguistics were related, if at all.

Onur said...

@Gioiello

The R-M269-PF7562* from Caucasus (he is a Laz living in Turkey) is a first migrated from the Italian Refugium and has nothing to do with the subclades from PF7563, centred in Italy.

No, not Laz. I know him personally. His paternal great-grandfather is from Trabzon province in Turkey. People of Trabzon mostly speak Turkish today, but a minority of them still speak Pontian Greek despite being Muslim. Genetically people of Trabzon are the same irrespective of language (Turkish or Greek) or religion (Islam or Greek Orthodoxy) and the majority of them have no Turkic genetic influence.

Gioiello said...

@ Onur

I thank you for your information, but it seemed to me that he said being of Laz origin, anyway the important thing is that there has been a migration from old times along the Mediterranean shores of many hgs. We find there also the R-M269* of the "Sardinian type" with YCAII=17-23 and DYS385=12-14. We'll understand next where its origin was, if in Sardinia/Italy or in some other point along the Mediterranean Sea.

Onur said...

@Gioiello

I thank you for your information, but it seemed to me that he said being of Laz origin

Yes, I know. It is because in modern Turkey everyone from northeastern Turkey is wrongly assumed to be Laz despite the fact that Trabzon and many other provinces of northeastern Turkey have no Laz population.

Rob said...

@ KarlK

"I am not sure there is any real basis for this conclusion. We basically have no idea how the Basque or pre-Indo European Sardinians came to have their languages, or how their linguistics were related, if at all."

But that's my point. You're the one who posited a genetic link between Basques and Sardinians, which I also took as your tacit support for said linguistic links

But really, Basques are merely the tail end of a central-western European continuum, whilst Sardinians are not on such an axis.

So we have
- different autosomes clusters
- A wholly different set of Y lineages
- totally different Bronze Age facies

So the burden of proof of proof lies in the court of those favouring such definite links

Atriðr said...

@Gioiello

On your points:

2) Although there is a propensity towards claiming centum the oldest, there is no proof. So I wouldn't use this as a rock-solid argument. Also, k-s sound change not understood by most, even linguists. Worth the study.

3) Diversity of which languages? Latin entered the peninsula from outside. There is not that much diversity in the context of all I-E languages. And diversity as argument, as with biological genetics, can also be misleading.

4) Etruscan works against your position.

The avenues of exploration for the Urheimat in Villabruna are very slim, but as all avenues worth the pursuit (we are 7 billion on the planet after all), I'd think you'd have more luck trying to combine your position with Neolithic Farmers/Anatolia.

This is not my position of course, quite the opposite, but not out of unwillingness.

Gioiello said...

@ Atriðr

I think no serious linguist doubts that centum languages are older than the satem ones. We may have "s" from "Ky", never "Ky" from "s", without counting all the intermediate positions as Albanian "th". You could say that a centum language is the Easternmost, as Tocharian, but we have the law of the "aree laterali" (Bartoli). You could put the question of Hittite and of the laryngeals, but also Albanian have them and the IE substratum of the Balkans should be explained. Many Greek words come from substratum.
That Latin has come from elsewhere Italy is just the theory of my opponents. Which proof? Certainly Latin was the language of the pile dwellers, and the oldest are in Italy, even 4500 years BC.
I don't understand why the Etruscan question is against my theory, both if it came from Anatolia with the agriculturalists and if it came from Central Europe or the Alpine region.

Simon_W said...

I agree with most commenters here that we desperately need samples of late Bronze Age and Iron Age autosomal DNA from Italy, to finally answer questions like: what were the early Latins like? Did they already have that strong south Italian component that modern central Italians have? Or were they still Iberian-like? And what were the Etruscans like? Is there any evidence for admixture with an Aegean population? And what were the early South Italians like? When did they acquire their strong extra CHG? Very early on or late? And what about the Sicels, did they already have the strong Near Eastern signal modern Sicilians have or were they very different? Etc, these would be very interesting things, but I fear we'll have to wait for long until something gets published.

Simon_W said...

Also worth investigating: what were the Padanian Gauls like after admixture with the locals, and how much continuity was there after the Roman conquest? It seems like Italian geneticists are not interested in this sort of questions.

Matt said...

@ Davidski, thanks for the info on your experiments. Maybe one day the quality of the samples will be enriched heavily and can be tested.

I might suggest that if the Yamnaya's X can't be tested reliably with low SNP count, there might still be some trace in the X of the populations who are descended from them, where these are richer in CHG than the autosome.

(Though I'm not sure about this due to the "X swapping" than would've happened with Europeans, e.g. if X in Yamnaya was richer in CHG than autosome, but then those X are "swapped" out for European EEF X, then descendants may end up with more CHG on the autosome.)

@ All: Connected to the X chromosomes, new adna study on the genesis of Polynesian model:

https://www.sciencedaily.com/releases/2016/10/161003130908.htm
http://www.archaeology.org/news/4884-161003-polynesian-genome-sequenced
http://www.sciencemag.org/news/2016/10/game-changing-study-suggests-first-polynesians-voyaged-all-way-east-asia

(I can only find news links, so if anyone has a link to the proper study, that would be cool).

a) Suggests a basic model essentially pretty similar to the European model -> where farmers fan out across the area without picking up much, then male mediated admixture happens, with essentially the Austronesians being like the EEF of Europe and Oceanian people taking on an equivalent role of to the Yamnaya. Crucially, the X chromosomes of recent people is relatively depleted of East Asian segements and enriched in Oceanian segments.

b) Relevant to Eurogenes type interests because tropics adna. May mean practical adna is closer for India and Southeast Asia and Arabia. This one is a particularly good early test case because the relatively shallow time depth is about 3000 YBP.

Shaikorth said...

Assuming Lapita spoke Austronesian there was no language shift associated with the genetic shift. Matrilineality allowed for male-mediated Papuan admixture and then mixed Austronesian groups expanded over the unmixed ones further in the Pacific?

Matt said...

@ Shaikorth, good question and interesting idea. I haven't checked out the ancestry estimates yet. Just from a quick look the Tongan group has about 25% ancestry from presumably mainly male Oceanian people, so it may be that a lack of language shift is just something like the Basque case? Where there's a clear impact that visually looks around 25% but the Basque language may be something the EEF brought.

Full paper's here with links to public data - http://www.nature.com/nature/journal/vaop/ncurrent/full/nature19844.html.

In PCA and f3 stat the Lapita group are somewhat remote from their closest proxies today (Atayal and Kankanaey), and in f3 may show some of that extreme tightness of f3 shared drift I think seen in European Upper Paleolithic-Mesolithic HG and the Boncuklu Epipaleolithic.

Atriðr said...

@Gioiello

Well, it's safer not to use "consensus" as an argument, especially when you seek to promote a view that no consensus remotely believes.

The sound change you describe is not set in stone. Also, by that same approach, s always -> h, but never h -> s. This hasn't stopped the distinguished from dating Avestan older than Sanskrit in one instance of thinkers going against an absolute. Linguistics is a science, but still evolving. Well, actually it's kind of dying because few study it... but my point is that things are still up for debate.

Tocharian languages don't fully fall into the centum-satem split quite like the other groups.

The intrusion of the Latini into the peninsula is attested by their own myths, nevermind the areal features of Etruscan onto the language. It also has less declensions than other I.E. languages, another strike against its antiquity. And where do you get this 4500 BC proof of Latin? I've never seen so I'm curious.

Shaikorth said...

Don't think Basque situation is comparable even if we presume I-E gave them Y-DNA but no language. Papuan Y-DNA and autosomal influence is all over Polynesia but Austronesian dominates the Pacific - even Melanesia - and Papuan languages are restricted. Basque is an isolate and surrounded by I-E.

Gioiello said...

@ Atriðr

"The intrusion of the Latini into the peninsula is attested by their own myths, nevermind the areal features of Etruscan onto the language. It also has less declensions than other I.E. languages, another strike against its antiquity. And where do you get this 4500 BC proof of Latin? I've never seen so I'm curious".
1) By a genetic point of view, there isn't any difference between the Etruscan world and the Italic or Ligurian ones. Look at the percentages of R-U152, thus the differences of languages are due to recent vicissitudes that very likely didn't change the previous genetics.
2) Everybody is against me when I say that Italy is very few tested as to the aDNA: only 5 cases (Oetzi, 3 Remedello, Villabruna) apart the old samples from Ostuni etc. We are waiting that other tests come, for instance a Villanovian from 2700 ya and more Rinaldone, Balzi Rossi and possibly samples from Mesolithic at least.
3) Having many cases isn't a proof of evolution: IE had 7 or 8, to-day the IE languages much less. Etruscan is intermediate between IE and Caucasian (this was the thought of Alfredo Trombetti), and of course we don't know everything about this languages, just becasue a little documented.
4) That Latin language formed in a pile dwellers environment is said from Giacomo Devoto (the linguist who put Westernmost the origin of the IE) in Storia della lingua di Roma, 1940, p. 5, even though he says that they came from Thuringia. The hypothesis is of V. Bonfante, "Atti dell'Istituto Veneto", 97, 1938, pp. 53-70, just for words like "pagus", "pons", "portus".
Of course the origin of pile dwellers in Italy 6500 years ago is another thing, but you may find it also on the web. I posted all that in the past, many times and on many blogs. Of course that Latins have lived in pile dwellings isn't a demonstration that that happened at their beginning, in fact Devoto says that that happened when they arrived in the Alpine region coming from North, and also that is an hypothesis all to demonstrate.

Simon_W said...

Well, Latins from pile dwellers, or more precisely from the Terramare culture, that's a hypothesis at best. According to some archaeologists the Terramare facies penetrated the Tyrrhenian area fully. Latins via the Balkans at an earlier date would be another hypothesis. In any case for linguistic reasons they must have had a slightly different origin than the Sabellics. And as for their own myths attesting an intrusion from elsewhere, Dionysius of Halicarnassus claimed they were descended from Aborigines who had controversial roots. He wrote that some believed them to be autochtonous, that others believed they were descended from Ligurians from northern Italy (which would seem to fit with the Terramare theory), and that again others believed them to have originated in Greece. But I think the main point that has to be stressed is indeed that it's proven beyond doubt that there was considerable immigration into Italy that changed the Sardinian-like Neolithic population by enriching it with EHG and CHG admixture, and the expectation that future Italian samples will prove this to be wrong, is outlandish.

Simon_W said...

Or put differently, there was admixture with LNBA Europeans, or in other words a "northern" influence, plus excess CHG related admixture in the south. It would be an outright miracle if by chance a similar LNBA ancestry arose in Italy independently from the one in the north, and that South Italian HG were by chance similar to CHG and to Anatolian Chalcolithics. And anyway, Italy is an attractive country, it's a matter of taste of course, but many people (including myself) find it very beautiful, so why assume that it never attracted any immigrants?

Gioiello said...

@ Simon_W

My theory of an "Italian Refugium" was concerned about the origin of hg. R1b1-L389+ and subclades. I am certain so far, and I think having demonstrated that, that R-V88, R-L389+, R-M335 were here and very likely part of the clan Villabruna. I said that the R-L23 of Yamnanya came from Italy or at least from Western Europe. I am waiting that also R1a, J and I of course are found in the clan Villabruna.
About the origin of Latins and the other peoples of Old Italy we'll speak another time. There are 10000 years in between, and many things happened.

Simon_W said...

But Villabruna didn't have enough EHG ancestry to explain modern Italians, nor did they have enough CHG, they were basically straight WHG. So you'd need considerable autosomal admixture coming into Italy, without carrying any yDNA along, for your theory to work. And moreover it would be spooky if Neolithic Italy had CHG, the Caucasus had CHG, but early Neolithic Anatolia inbetween didn't have CHG.

I know your theory is about uniparental markers only, basically just about yDNA, but you can't ignore the evidence from autosomal DNA.

Gioiello said...

@ Simo_W
I posted that in another thread here and also on YFull page of FB. These uniparental markers seem demonstrating perfectly my theory. I doubt that a theory may be demonstrated from autosomal SNPs.

@ Richard Rocca
If you understand something of genetics, this is the demonstration as to R-V88 was born and its migration 7500 years ago from Italy to Iberia and 7000 years ago from Iberia to Sahara:

R-L754*
R-V88 FGC21059/Y7781 * Z30230/Y7770 * V88/PF6279+72 SNPs formed 17300 ybp, TMRCA 10200 ybp info
R-V88*
id:YF07201new [Saxton/Sexton from the Isles, Marchesi from Italy, that FTDNA tested M269 and I demonstrated being V88]
R-M18PF6372 * PF6319 * YP5453+23 SNPs
id:ERS256975ITA [IT-CA]
id:ERS256965ITA [IT-CA]
R-Y7777 SK2065/FGC21014/Y7777 * FGC21063/Y7784 * FGC20993/Y7786+14 SNPs formed 10200 ybp, TMRCA 7600 ybp info
R-Y7777* [
Spain
Els Trocs [I0410 / Troc 3]
M
5295-5066 calBCE (6217±25 BP)

528,396
R1b1c
M415+, M343+, [L754 equivalent: L774/PF6245/YSC277+, PF1144+, V88 eqivalent: PF6376+], M478-, PF6399-, L265-, L150-, M269-, V35-, V69-
T2c1d or T2c1d2


Haak 2015; personal comm Sergey Malyshev, review of Y-DNA raw data; Mathieson 2015; Lazaridis 2016]

R-V35 V35
id:ERS256961ITA [IT-CA]
R-Y8447 FGC20970/Y8447 * FGC20998/Y8458 * FGC21009/Y8452+1 SNPs formed 7600 ybp, TMRCA 7000 ybp info
[7000 years old are the African and Middle Easterner V88]

Onur said...

@Gioiello

I doubt that a theory may be demonstrated from autosomal SNPs.

A theory cannot be demonstrated ignoring autosomal SNPs.

Chad Rohlfsen said...

Villabruna isn't ancestral to shit. His ancestor split with the ancestor of L51 around 3k and maybe up to 5k years prior. You need to look further east, at the source of epiGravettian. End of story... Moving on.

Arch Hades said...

I really wish these geneticists would sample southern Europe too.

MomOfZoha said...

@Onur:
Is it just a coincidence that you happened to know that Trabzonlu donor or do you collect and analyze Anatolian/Turkish/Turkic DNA for academic/professional purposes?

Eger bu konuda arastirma yapiyorsan sana email ile soru sorabilirmiyim?

Onur said...

@MomOfZoha

Is it just a coincidence that you happened to know that Trabzonlu donor or do you collect and analyze Anatolian/Turkish/Turkic DNA for academic/professional purposes?

Eger bu konuda arastirma yapiyorsan sana email ile soru sorabilirmiyim?


I know that donor from Trabzon personally from Facebook as we were in the Facebook group of the same FTDNA project we are members of.

Yes, I currently manage a FTDNA project on the genetics of Anatolia, the Balkans and the Caucasus. You are invited to join it if you are on FTDNA. Here is the main link to our project:

https://www.familytreedna.com/groups/anatol-balkan-caucas/about

On that same page you can see my email when you move your cursor on the mail picture next to my name. You can direct your questions to me via that email (in English or Turkish, it does not matter).

If you are on Facebook, here is the Facebook group of my FTDNA project that you can join even without joining the FTDNA project:

https://www.facebook.com/groups/800912433320422/

MomOfZoha said...

@Onur:

Thank you for that information.

I just sent you a join request from the FB page (bisikletliyim). Unfortunately, we tested via 23andme and do not have FTDNA kits.

I'll also email you later today after I get through a bit more grading.

Tekrar sagol!

ser nam said...

See what nagging your husbands can do?

jv said...

Hello! First want to say, I really enjoy this blog! THANK YOU! Ok, regarding this male biased migration from the LNBA Steppe(Rib & R1a).........Did these men have cattle or sheep, animals to move to greener pastures?(because the climate in the Pontic-Caspian Steppes became colder & drier 3000-2000 BCE) AND for every 5-14 males, 1 woman from the Steppes migrated. I would like to know a lot more about this migrating WOMAN as my mtDNA is associated with the Steppe. Was she a sister to migrating brothers? Perhaps she was young and fit so she could make the journey? I wouldn't think she would have a wagon load of children with her! Perhaps she carried a small loom with her to make clothing? Yamnaya Culture used plant fibers but the Catacomb Culture used a combination of plant fibers and wool. Did she take her Steppe cord decorated pottery with her?( Yes, Corded Ware Culture)Hopefully, with time, we will find out a more about this migrating Steppe woman. I look forward to it as one of those woman is an ancient Grandmother of mine that migrated from the Steppes to the Balkans.