A genetic probe into the ancient and medieval history of Southern Europe and WestAsia

In the paper there are models for each sample using CHG, EHG, Levant_PPN, SRB_Iron_Gates_HG and TUR_Marmara_Barcin_N.

I simply took the name of the samples and their relative ancestry proportions, and put them on Vahaduo. Technically it is not a real calculator, I was just using it to see the distances to the other samples.


Their relative ancestry proportions based on what? G25?
 
Angela: Ok, I was not sure what he meant. I think Bell Beaker culturally was related to the Corded Ware, so I think your interpretation is correct. I would think while they all picked up IE languages via Corded Ware expansion which overlapped Bell Beaker culture, the genetics I don't think are the same. The last big paper on the Bell Beakers was that Olade et al 2018 Nature paper and there are 226 Bell Beakers and the paper did note high heterogeneity among the samples. I went back and quickly looked at as I like you did not get to invested in the paper since of the 226 Beaker samples, only 6 really interested me, the 3 Northern Italian Beakers and the 3 Sicilians, but I think due to quality control/coverage, only 1 of the Sicilians had full genome analysis.

But your hunches, as I already alluded to, were correct regarding Beakers:

From the paper

"Individuals associated with the Beaker complex are notably heterogeneous within the European cline along an axis of variation defined by Early Bronze Age Yamnaya individuals from the steppe at one extreme and Middle Neolithic and Copper Age Europeans at the other extreme (Fig. 1c; Extended Data Fig. 3a). This suggests that genetic differentiation among Beaker-complex-associated individuals may be related to variable amounts of steppe-related ancestry"

The Beakers that went into Great Britain had very, very high Steppe and replaced like 90% of the pre-Steppe Neolithic population. In Iberia, some of the Beaker Iberians were culturally Beaker but lacked any Steppe and were similar to the Neolithic Iberians, although some Iberians had Steppe ancestry which brought IE languages in. But to be honest, there is no clear admixture model in the main text as the samples vary too much.

I found this statement in the Supplement (p. 152)


"Steppe ancestry in Beaker-associated individualsWith PCA, ADMIXTURE and f-statistics, we learned that our newly reported individuals reside along the Steppe Early Bronze Age-European Neolithic axis of genetic differentiation. Thus, we tried to model them as a mixture of Steppe_EBA + Anatolia_N + WHG (Table S4). These values were used for Fig. 2a. Many populations can be explained by a mixture of Anatolia_N + WHG without any contribution from Steppe_EBA, indicating a lack of Steppe-related ancestry."


On page 167 (Table S.4 is the admixture model), the Beaker Italians do not have Predominate Steppe either, 2 of them have none and the 2 that have Steppe have 25.6% and 29.7% respectively (I assume 1 of these is the Sicilian Beaker. Some UK samples have 60% or more Steppe, the Germans > 50%, etc.

So if Northern Italian Bell Beaker types represent the population that brought in Steppe ancestry and IE languages in the North of Italy, which is what Raveane et al 2019 suggested, they were most definitely not Corded Ware types genetically. I used Jovialis's K8 model to see what it looks like and it works I think quite well for those 4 Beakers (3 Northern Italian, 1 Sicilian). 2 of them, similar to the reported results in Olade et al 2019 Supplement have Steppe (1 the Model hits dead one) and the other 2 had zero (which is effectively what Sicilian Beaker and I2477 Northern Italian Beaker have in the Model). I remember you often saying the Steppe in Northern/Central Italy was mediated through Beaker culture and I think the evidence supports what you said in another post in another thread.



Target: Olade_etal_2018:I2478_Bronze_Age_Beaker_Northern_Italy
Distance: 1.9609% / 1.96094933
60.6Remedello
29.3Yamnaya
8.9Minoan
1.2Bolshoy_Ostrov



Target: Olade_etal_2018:I2477_Bronze_Age_Beaker_Northern_Italy
Distance: 1.2065% / 1.20646925
54.2Remedello
30.4C_Italian_N
9.7Minoan
5.4Iberomaurusian
0.3Yamnaya


Target: Olade_etal_2018:I1979_Bronze_Age_Beaker_Northern_Italy
Distance: 4.9469% / 4.94690417
42.1Minoan
37.7Remedello
18.1Yamnaya
2.1Bolshoy_Ostrov


Target: Olade_etal_2018:I4930_Bronze_Age_Beaker_Sicily
Distance: 1.6558% / 1.65579333
76.8Minoan
18.5Remedello
1.9C_Italian_N
1.4Yamnaya
1.3Iberomaurusian
0.1Bolshoy_Ostrov

Anyway, sorry to go deep into the weeds on the Steppe impact and spread of IE languages on Italy but since it is a hot topic, I have joined in more on the Steppe non Steppe source of Proto-Indo European language and which type of groups spread IE into Italy and Greece.

Palermo, do you have the coordinates for those four samples handy? Do you know if they're also available in K12b and Vahaduo?
 
Having spent some time going over the graphics I think I can answer some of my questions. If I'm going wrong here, people, please let me know. This is the graphic.

Yes, they included all the Antonio et al samples, northern, Anatolian like, Levant like etc. They also included, as you can see, all the other Italian groups from the Chalcolithic forward, and all of the Anatolian samples from "Ancient", which I assume is Neolithic, through the Bronze Age all the way to the Byzantine period.

There are not very many Iron Age samples from Anatolia.

What this team focused upon, which sample sets are indeed bolded, are the Antonio et al samples from the Imperial City of Rome, and what is labeled TUR_RomByz. I haven't checked every one of those samples to see what percentage was from the Roman Era and what percentage was from the later Byzantine Era. Does anyone know?

I suppose one could say that both groups are very "cosmopolitan". A few are quite "northern like", some are very Levantine like, while many have much less, and you even have a few who are Armenian like.

Lazaridis goes to some lengths to talk about the "outsiders" who could be found in Anatolia in the days of the Eastern Roman or Byzantine Empire.

I would submit that the same is true of the Imperial City of Rome: lots of outsiders who did not all necessarily contribute to the genetics of Early Medieval Italians, for example. At any rate, this rather broad, general analysis certainly wouldn't answer that question.

I think it's instructive, however, to also look at the Anatolian Bronze Age or even the TUR_Anc, which I believe is Anatolian Neolithic. Both Rome Imperial and Eastern Rom/Byzantium overlap a lot with those more ancient groups. How does that prove that all of the people from the Imperial City of Rome who overlap with all of those groups came to Rome specifically "from" Anatolia, in the first three centuries of the Empire?

I'm not saying it didn't happen. I just don't see how this proves it. I also think the conclusion is pre-mature.

As I pointed out in my first post here above to which I am responding, I think it would be crucial to include Aegean Bronze Age samples, and even more importantly, Aegean Iron Age samples, when or if they have them, and certainly the Greek Marathon sample, and the two Empuries samples, to see if they too could possibly be the source of this signal in Imperial Rome. Sicilian Bronze Age should also be in the mix, and, when we have them, samples from Greek colonists to Southern Italy.

I still believe the following:
"Using the term "lived" around Rome in the Imperial period is questionable, imo. We're talking about samples not just in Rome itself but also in Ostia, a port city where transitory seamen and merchants would "live", but only for limited periods of time."
of
The same would apply to Byzantium, of course, unless we are to believe that in Anatolia proper there were people with such high levels of Levantine ancestry. It's possible that the samples he's using for Byzantium are not, indeed, just from Anatolia proper. All of the samples would have to be examined to see what percentage are strictly from Anatolia proper.

Also, I still hold with the following:
"Indeed, in terms of Anatolia, there were, as we all know, numerous Greek settlements there, inhabited by people who were, I agree, probably a mixture of Greek and western Anatolian. Were the samples used to compare to Imperial Roman samples from this group of people?"

If his samples for EasternRom/Byzantium are mostly from those heavily Greek influenced areas, then are they precisely "Anatolian", or rather a mixture of Greeks and Anatolians, which is why, indeed they spoke Greek. That is, not because they adopted it as a lingua franca, but as the language of their ancestors, a la Herodotus himself, half Carian that he was.

So, was it 100% Anatolians who went in such large numbers to the City of Rome, or was it Anatolians, Greeks, Greek/Anatolians, and yes, some Levantines. Whether they all stayed and changed the genetics of, say, Lazio, permanently is another issue.

I also would like to address this statement that the Anatolians were the "engine" of the Roman Empire. Perhaps they were the engine of the Eastern Roman Empire/Byzantium, but I certainly wouldn't go so far in talking about the western Roman Empire. In fact, there would have been no Western or Eastern Roman Empire without the legions of Roman, which originally and for all the initial large conquests were made up of people from the Italian mainland and its islands. Nor could it have existed without the roads they built, which facilitated trade, although the larger part was probably by sea. Roman culture would certainly never have existed without the borrowings from Etruria, and, of course, from Greece itself, including certainly, Ionian Greek City-States from Anatolia, but hardly further inland, and with the most respect given, from everything I know, to Athens. The upper classes of Rome very early on taught their children Greek, indeed, but that was to understand Greek philosophy and literature and science. Yes, the people from the Aegean and all the areas where Greeks had established colonies spoke Greek, and so it became a common language certainly in the eastern part of the Empire and in parts of Italy itself, but let's not forget that the peoples of the conquered areas in Western Europe spoke Latin derivatives, not Greek ones, so let's not exaggerate.

From what I've seen on G25 Anatolia did receive European gene flow, quite of it actually; I suspect they missed it because of their focus on modelling with distal ancestries, whereas it's clear that it is better to use proximal ancestries.
Overall Greeks in Caria were around 50% Mycenaean-like and 50% Anatolia_BA
From around the black sea Greeks were 40% Mycenaean-like and 60% Anatolia_BA
As for Gordion samples, if we take them as "Phrygian", many seem to take around 20-25% Logkas-like ancestry, or 40% Mycenaean-like, but since they are dated in the late hellenistic period (150 BC), I am not sure whether it can be attributed to proto-phrygian ancestry from the Balkans or to Greek ancestry.


unknown.png

unknown.png
 
Angela: They are in Dodecad12B: I2478 is from the Vahaduo spreadsheet. The Sicilian Beaker I used the Gedmatch kit seen below. The other 2, I ran a while back and I have the coordinates but I don't have the GEDMATCH Kit that I used. Let me dig, or maybe I got the Kit from a post here. Sicilian_Beaker_I4930 (Kit #TZ9503361)

Beaker_Northern_Italy:I2478:Olalde_2018,4.97,0,0,0.15,47.73,28.87,0.74,0,3.49,0.84,12.34,0.86

Siclian_Beaker_I4930,15.32,6.54,38.77,12.25,22.74,4.38,0,0,0,0,0,0,0
 
Last edited:
In the paper they should have added Iran_N.

Red: paper
Blue: Global25

aJZK2eH.png

The authors made it clear that in their view, as we have said over and over again here, CHG and Iran Neo are almost indistinguishable for these kinds of purposes.
 
G25 seems biased, when trying to replicate fstat breakdown.

PCA related problem or ... ?
 
The authors made it clear that in their view, as we have said over and over again here, CHG and Iran Neo are almost indistinguishable for these kinds of purposes.

The same could be said for Barcin_N and Levant_PPN, but both were used.
 
I think I read somewhere in the paper that in many cases Iran_N produced unacceptable low p-values whereas the CHG models were robust.

So it's only logical for the methodology?
 
"When we
attempt to model Neolithic populations using
either Caucasus hunter-gatherers or Ganj Dareh
as a source population and the other as an
outgroup, we obtain good model fits for most
populations (further suggesting that neither
population is a better source than the other),
except (i) for the high Caucasus hunter-gatherer
ancestry individual from Aknashen, where the
Caucasus hunter-gatherer model is not rejected
(P = 0.46) while the Ganj Dareh one is (P <
0.001); (ii) the Azerbaijan and Mesopotamian
Neolithic for which both models are rejected
(P < 0.01); and (iii) the Barcın Neolithic for
which the Ganj Dareh model is narrowly not
rejected at the P = 0.01 level (P = 0.0142), while the Caucasus hunter-gatherer one is rejected
(P = 0.001). These results tentatively suggest
that Caucasus hunter-gatherer and Ganj Dareh
Neolithic are interchangeable for the purposes
of quantifying the amount of inland admixture,
although some populations may have a clearer
connection with one or the other (e.g., the Neo-
lithic of Armenia with the hunter-gatherers of
the South Caucasus rather than Iran, and the
geographically intermediate Azerbaijan and
Mesopotamia with both)."
 
Above poster beat me to it.

Also, in relation to the model for Yamnaya with PPN:

K=4 We were curious about the rather low p-value (0.05) for the fit of the Yamnaya cluster itself; this is driven by two outgroups (Ganj Dareh: -2.96 and AZE_N: -2.04). (Note that neither AfontovaGora3-related (Z=-0.51) nor Levantine PPN-related (Z=-1.58) is significantly underestimated by the 3-way model, and thus the low p-value in Yamnaya is of a different etiology than the rejection of the 3-way model for the Steppe Maykop and Piedmont Eneolithic discussed above). Thus, we also fit 4-way models for this population (Table S 25). The CHG/EHG combination is invariant in the fitting models, with the EHG proportion in the ~40-50% range in all of them. We note parenthetically that the model of (17) that includes CHG/EHG/WHG/Anatolian Neolithic ancestry fails in our framework (p<1e-10), and inspection of outlier f4-statistics indicates that it underestimates (Z<-3) shared drift with Levant_PPN (Z=-5.6), Natufians, Azerbaijan Neolithic, and Ganj Dareh outgroups. We note that Levant_PPN was not used in the set of outgroups of (17) which might be one reason why a successful fit was possible.
 
"When we
attempt to model Neolithic populations using
either Caucasus hunter-gatherers or Ganj Dareh
as a source population and the other as an
outgroup, we obtain good model fits for most
populations (further suggesting that neither
population is a better source than the other),
except (i) for the high Caucasus hunter-gatherer
ancestry individual from Aknashen, where the
Caucasus hunter-gatherer model is not rejected
(P = 0.46) while the Ganj Dareh one is (P <
0.001); (ii) the Azerbaijan and Mesopotamian
Neolithic for which both models are rejected
(P < 0.01); and (iii) the Barcın Neolithic for
which the Ganj Dareh model is narrowly not
rejected at the P = 0.01 level (P = 0.0142), while the Caucasus hunter-gatherer one is rejected
(P = 0.001). These results tentatively suggest
that Caucasus hunter-gatherer and Ganj Dareh
Neolithic are interchangeable for the purposes
of quantifying the amount of inland admixture,
although some populations may have a clearer
connection with one or the other (e.g., the Neo-
lithic of Armenia with the hunter-gatherers of
the South Caucasus rather than Iran, and the
geographically intermediate Azerbaijan and
Mesopotamia with both)."

Thanks, just wanted to post that part but you was faster.
 
You are welcome, it was in the Mesopotamia paper that Eupator shared with us.
 
I think as others here (not all) that the wording of the Lazarides paper about the input of "Anatolian-like" DNA in Imperial Rome is a bit misleading. I read it to quickly? The Greek strong input and maybe the concept of Greekization of Anatolians of the time for a part could be received better? Partly kind of a Greeks mediated immigration?
 
Angela: Here are the Coordinates for the I2477 Northern Italian Bell Beaker. The First 3 are from Gedmatch Kits GQ2993310, T210092, and ZZ8725178. The 4th kit is from one that I had picked up from blog site and I honestly don't remember where and I don't have the Kit number. But they all do show zero Steppe which is what they should show per the Olade et al 2018 paper. The 4th set of coordinates are what I used in the previous post. For I2478, the first set of coordinates same deal as the 4th Kit for I2477, got it a while back ran the coordinates and don't have them. The 2nd set for I2478 is Gedmatch Kit M782936

Olade_etal_2018:I2477_Bronze_Age_Beaker_Northern_Italy,0,0,3.28,0,64.45,5.74,0,0,7.51,0,19.02,0
Olade_etal_2018:I2477_Bronze_Age_Beaker_Northern_Italy,0,0,1.83,0,68.82,5.11,0,0,6.71,0,17.52,0
Olade_etal_2018:I2477_Bronze_Age_Northern_Italian_Beaker,0,0,1.83,0,68.82,5.16,0,0,6.7,0,17.49,0
Olade_etal_2018:I2477_Bronze_Age_Beaker_Northern_Italy,0,0,6.01,0.21,60.29,7.43,0,0.07,7.60,0.08,18.30,0

Olade_etal_2018:I2478_Bronze_Age_Beaker_Northern_Italy,5.34,0,0,0,47.45,28.85,0.64,0,3.59,1.22,12.27,0.65
Olade_etal_2018:I2478_Bronze_Age_Beaker_Northern_Italy,2.79,0,0,0,51.85,32.24,0,0,0.54,0,12.57,0


Some of those Kits are I still listed on the Eupedia Gedmatch Kit number thread, which is where I first got them. But all of the Gedmatch kits provide coordinates that produce results that are in line with the Olade et al 2018 paper. None of the I2477 show any Steppe when you run them and use Jovialis K8 model (Yamnaya used to capture Steppe), the I2478 kits (both) show Steppe ancestry which is what the published paper shows.

Hope this Helps, PT
 
Last edited:
The same could be said for Barcin_N and Levant_PPN, but both were used.

Ed. OK, everybody beat me. You guys must not have moved from your computers for hours. :)

No, it couldn't. They make the difference clear in the paper on Mesopotamia which we haven't yet explored to any great degree.

For most purposes, CHG and Iran Neo can be used interchangeably. The same is definitely not true of Barcin and Levant PPN.

https://www.science.org/doi/10.1126/science.abq0762
 
I think as others here (not all) that the wording of the Lazarides paper about the input of "Anatolian-like" DNA in Imperial Rome is a bit misleading. I read it to quickly? The Greek strong input and maybe the concept of Greekization of Anatolians of the time for a part could be received better? Partly kind of a Greeks mediated immigration?

Well, I certainly agree with that, as I hope I made clear in my post #76. :)

Of course, the reality is probably that the Greek colonists and Greek traders and artisans etc. were also heavily Anatolian Bronze, as I think we all get. That's why I don't get all the Anatolia chest thumping sort of language in the paper.
 

This thread has been viewed 37161 times.

Back
Top