Many years ago, well before the start of the ancient DNA revolution, someone made the very clever inference that the N-Tat Y-chromosome marker was closely associated with the expansion of Uralic languages.
Since then, N-Tat has been renamed several times over, to the point that I no longer know what it's called, but the aforementioned inference has turned into a very solid consensus backed up by a wide range of studies focusing on modern and ancient DNA.
Nowadays, Y-haplogroup N-L1026, a subclade of N-Tat, is seen as the main genetic signal of the Uralic expansions, along, of course, with Nganasan-related genome-wide genetic ancestry.
A
recent paper at
Science Advances by Gnecchi-Ruscone et al. featured the first ever genome-wide samples from the Sargat horizon, which is an Iron Age archeological formation in western Siberia normally associated with the Ugric branch of the Uralic language family. Surprisingly, and disappointingly, the authors failed to investigate this widely accepted connection.
If we go by the Y-haplogroup classifications in the paper, which may or may not be the smart thing to do, at least two of the Sargat horizon males belong to N-L1026, and one also to the more derived N-Z1936 subclade, which has been found in the remains of Hungarian Conquerers from Medieval Hungary. Of course, Hungarian is an Ugric language generally thought to have been introduced into the Carpathian Basin by the Hungarian Conquerers who originally came from western Siberia.
That's probably enough to corroborate the association between the Sargat horizon and the spread of Ugric/Uralic languages, but let's also take a quick look at the autosomal DNA of these Sargat individuals. Firstly, here's a Principal Component Analysis (PCA), based on
Global25 data and produced with the
Vahaduo G25 Views online tool. The results are self-explanatory.
Interestingly, I can't get a decent statistical fit when I try to reproduce the four-way qpWave/qpAdm model done by Gnecchi-Ruscone et al., probably mostly because my right pops or outgroups are different. This suggests to me that there's something important missing in their model.
Sargat_IA
MNG_Khovsgol_LBA 0.203±0.045
RUS_Ekven_IA 0.183±0.044
RUS_Sintashta_MLBA 0.545±0.014
TKM_Gonur1_BA 0.068±0.013
chisq 16.805
tail prob 0.0186971
Full output
So how about if I replace RUS_Ekven_IA with kra001, the oldest Nganasan-like individual in the ancient DNA record (see
here), and MNG_Khovsgol_LBA with KAZ_Mereke_MBA, to add a more local stream of ancestry?
Sargat_IA
KAZ_Mereke_MBA 0.135±0.017
kra001 0.301±0.007
RUS_Sintashta_MLBA 0.499±0.023
TKM_Gonur1_BA 0.066±0.015
chisq 8.872
tail prob 0.262001
Full output
That's a better statistical fit and also, I'd say, a more realistic model, at least in terms of distal ancestry proportions. Note that Nganasan-related ancestry makes up 30% of the genome-wide genetic structure of the Sargat samples, which again corroborates the view that Uralic languages were spoken within the Sargat horizon.
Update 28/04/21: This is the best qpAdm model that I could find for Sargat_IA, at least in terms of the chisq and tail prob. It shows that the Sargat population was in large part very similar to that of KAZ_Pazyryk_IA.
Sargat_IA
KAZ_Mereke_MBA 0.032±0.016
KAZ_Pazyryk_IA 0.698±0.016
RUS_Sintashta_MLBA 0.236±0.021
TKM_Gonur1_BA 0.034±0.014
chisq 2.023
tail prob 0.958561
Full output
It's missing kra001, because KAZ_Pazyryk_IA packs enough kra001-related ancestry for the job.
KAZ_Pazyryk_IA
KAZ_Mereke_MBA 0.144±0.018
kra001 0.429±0.008
RUS_Sintashta_MLBA 0.378±0.026
TKM_Gonur1_BA 0.049±0.018
chisq 8.899
tail prob 0.259983
Full output
The fact that KAZ_Pazyryk_IA can be modeled with significant kra001-related ancestry isn't surprising, considering that its territory was located in Siberia.
However, my model doesn't necessarily prove that the Sargat population was largely or even partly of Pazyryk origin. Indeed, N-L1026 hasn't yet appeared in any Pazyryk remains.
See also...
The Uralic cline with kra001 - no projection this time
First taste of Early Medieval DNA from the Ural region
Hungarian Conquerors were rich in Y-haplogroup N
More on the association between Uralic expansions and Y-haplogroup N
It was always going to be this way
On the association between Uralic expansions and Y-haplogroup N
394 comments:
«Oldest ‹Older 201 – 394 of 394@vAsiSTha, there are a fair few different things we can do with qpAdm, depending on level of resolution you're interested in. E.g. Lazaridis, who invented it, used it to estimate Basal Eurasian ancestry using an African outgroup. I think using it for a low resolution look at where ancestry streams into the Steppe populations may lie on a general WHG->ANE/WSHG cline is one of those (WSHG is perhaps not ideal for this and a wider set of AfontovaGora3 samples would be better, but we have only one of those things).
I don't think pRight should have vastly different degrees of shared time depth with target populations, as this would seem to make CHISQ increase or decrease in ways that are biased with respect to the pLeft.
Btw, in case you are interested, I am using admixtools2 version of qpAdm and it lacks many of these options like allsnps, so this may affect outcomes. I believe using pre-computed f2 stats as I do just uses all SNPs common to all populations at the pre-computation step, which is given at the initial stage (so under my runs above, I think this would be 262608 SNPs with 211817 polymorphic for example.) I mention this only as this may be so for Rob as well if he is also running the admixtools2 implementation and this could set your discussion about methods at cross-purposes.
WSH needs to be modelled in qpADm primarily with a Hissar like source instead of CHG or Iran_N. CHG/Iran_N are historically used because they are quite differentiated and show up as 'basal' in ADMIXTURE runs. However, WSH will always model better with Hissar rather than CHG or Iran_N, and this really shouldnt be ignored, because it means Hissar shares DNA with Yamnaya not found in Kotias/Satsurblia, Iran_N or EHG.
These poor quality models are all based on those early ADMIXTURE runs in the first papers that came out, and ofcourse I have already move way past that.
"vAsiSTha said...
Interesting details from the Kak paper
"A most striking example of this are the figures on the Gundestrup Cauldron found in a bog in Denmark and dated to about 150 BCE that are clearly of an Indic origin: the Goddess being adored by two elephants (which is out of place in Europe),"
Hannibal crossed the Alps with elephants in 218 BCE. War elephants were also used in the Italian peninsula in the Battle of Heraclea which took place in 280 BCE.
The OOI crowd would hold on to every straw to prove that Yamnaya/CWC/Sredny Stog were formed in Central Asia instead of Eastern Europe. Thus, the tenacious and stubborn insistence that not only Indians or Iranians have/had Botai or BMAC, (contrary to Narasimhan 2018 and 2019), they even bend over backwards to claim that Yamnaya or Sredny did.
Off on a tangent here, it would be interesting to figure out where Nihali, Kusunda and Burusho came from, and whether Botai, Okunevo, Kelteminar or Steppe Maykop (WSHG, likely looked more Europoid than East Asian) - had any linguistic affiliation with Yenisseyan/Ketts and/or North American natives.
You can look at this map of J1a and see it fully supports what Davidski is saying. J1a is associated with CHG and it clearly expanded from West of the Caspian:
https://upload.wikimedia.org/wikipedia/commons/thumb/3/3d/Geographical_frequency_distribution_of_Haplogroup_J1-M267_%28Y-DNA%29.png/1920px-Geographical_frequency_distribution_of_Haplogroup_J1-M267_%28Y-DNA%29.png
Model with WSHG as 3rd source
left pops:
Steppe_Eneolithic
EHG - 37.2+-4.9
Georgia_Kotias.SG - 54.9 +- 2
Russia_HG_Tyumen - 7.9 +- 4
tailprob: 0.004
EHG = 3 samples from karelia and 1 from samara
steppe_eneolithic = 2 from progress, 1 from vonuchka, and 2 more: RK1007 and SA6010 which cluster together with steppe_en (discovered by Matt)
result file: https://pastebin.com/kKr9MzWL
Definitely shows some affinity to WSHG.
If I may add some more critical remarks about Carlos Quiles, here they are: it's ironic how he criticizes Davidski for being "just" a hobbyist, when he himself has merely studied economics, law and medicine, likes to learn languages in his spare time, and works as a medical doctor. So in what way is he a professional historical linguist, archaeologist or population geneticist? I'm not saying that hobbyists are inferior - what matters is expertise and the better arguments. But he wants to make the impression that he is more than that.
Few more fits with qpAdm on Admixtools2, using the 1240k file (which offers more available SNPs for ancients than the HO) - https://imgur.com/a/jl1IVJC
I don't think the pRight (outgroups) there are perfect by any means, and probably aren't specific enough to the test populations in many ways, but it's interesting how much they can pick up and what they're able to distinguish in terms of some of the main shifts and deep/broad scale differences.
@ Matt
Im using ADMIXTOOLS 1, for that very reason; although it is quite a learning curve :)
@ Vasistha
"Also please don't use Hotu sample in any model, it only has 32k SNPs.. literally all models which include it with allsnps NO will pass.”
Which is why I did it as a fun run
Khvalynsk, Yamnaya, and Progress/Vonyuchka, don't have real WSHG ancestry, beacuse for one, WSHG is like 90% EHG/ANE, it's very similar, and secondly, WSHG has East Asian ancestry, which Yamnaya, Progess/Vonyuchka, Khvalynsk, Volosovo, Karelia, CCC, Kunda, ect...don't have.
@matt
those runs look good, however the image resolution is very low. Would you post them as text in pastebin please?
@andrze
The Out of India crowd would hold on to every straw to prove that Yamnaya/CWC/Sredny Stog were formed in Central Asia
I chuckled for a good 5 mins.
@epoch
Your qpadm also has this:
"Israel_Natufian_published f4: -0.001816 Z: -3.287836"
True. ideally it means the source needs more of Natufian. However i realized that Natufian has the lowest snps of all my right population at 190k instead of 700k+ for all the rest so i decided to remove it for my latest runs, instead added Barcin_N.
"Also, if you consider the tail prob from Robs bad, yours should also be considered bad: 0.00122112461"
Yes indeed. None of my models for steppe_eneolithic work. Which is why it intrigues me.
"The second you posted, with Samarian HG, is that with allsnps yes?"
All my models are with allsnps: Yes. the qpfstats is first run on the list of samples (left + right together) with allsnps:Yes and then that output is given as input to the qpAdm file instead of the whole eigenstrat file as input.
This has the advantage of running qpAdm models with any subset of populations which were part of qpfstats run (qpf run takes 30-40 mins for me for 20-23 labels) and results come instantly.
Earlier each qpAdm model would take 20-30 mins as it computed fstats fresh each time.
Wshg looks like a descendant of Ma1, along with Botai. These pops are well disconnected from the West Asian cluster from which Yamnaya emerges.
Steppe pops really dont model with any ANE over what they get from EHG, the 10% (max) or so in models seems to be noise.
This looks like some kind of deep LGM type of seperation.
vAsiSTha: Image versions should be OK if looking on computer and not mobile, but here are pastebin: https://pastebin.com/fT4uUDLx and https://pastebin.com/CXepneRH
Should convert neatly into a tab separated .csv file. It might look a bit of a mess (hard to read I mean) unless you load it in your spreadsheet software of choice as a .csv.
I think that in some ways the pRight I've used are likely to be insensitive to some of the differences between source populations, so this is preliminary and would take any improvements offered. I've kept them fairly basic. (I'm not totally sure how to get it much better though; splitting some of the left populations to be in the right seems like a problem, particularly when there are low sample sizes, and other samples that could be in right as informative populations are guys like Natufians or Iran_Hotu who have low sample size / low SNPs / contamination or damage. I might try the fits again with Yana_UP added to see if that does anything useful). Also might not be exploring all the options that admixtools2 has to treat the data properly, so will take any suggestions on that. The main interesting thing to me was that the relatively limited outgroup set picked up a surprising (to me) amount of what we'd expect for main transitions and proportions.
@Simon_W,
That is funny.
Carlos Quiles is much like the phd on a plan who calls himself a doctor when someone is seriously injured.
Matt, nice output it's good to read. However, as Vasistha mentioned, you can try it with more differentially related proximal right pops like SehGabi, BMAC, Botai etc and others, that would be interesting, to see how it affects the fits. Those outgroups are really far from the target and source. I reckon WSHG would fall out with closer right pops.
@matt thanks
Right pops look mostly fine.
You should add mbuti or shumlaka as 1st right pop.
Optionally you can add MA1, irula/onge, czech vestonice/kostenki, villabruna to see if you get additional insights
@Davidski
More eastern I imagine more EHG/ANE?
Also, a bit random but since this topic is about east -> west gene flow, I was wondering if anyone read this paper on Polynesia. https://www.nature.com/articles/s41586-021-03236-5
Here in open access: https://tfc-taiwan.org.tw/sites/default/files/upload/node/appeal/field_file/s41586-021-03236-5.pdf
In the extended data figure 10, the authors seem to be making the case that unidirectional East Asian gene flow post 21kya brought Denisovan admixture into Europeans. Anyone here has an idea how accurate that is?
Honestly, their entire model, including the inter East Eurasians relationship seem odd to me.
@Davidski
"Khvalynsk is, in turn, more eastern than Progress/Vonyuchka."
At least as far uniparentals are concerned, Khvalynsk might have some actual WSHG because it carries subclades of Q1a2, which would point to a more eastern source, on top of being more EHG or Afontova Gora. Then again, Tyumen were Q1a1 IIRC.
Anybody is being tested, but not Achaeans or Dorians.
@mzp1, thanks. I haven't done anything with Seh_Gabi, but I've tried some things with Gonur1 (largest BMAC set) and Botai, and some Central Asia populations (mostly the Steppe Maykop like ones, but also a few things with Sarazm_En), including seeing what Botai in the pRight does (though it's not strictly good practice), in case you want to read.
Pastebin version: https://pastebin.com/ixxkQKKC
Imgur version: https://imgur.com/a/aCCKxKB
Also tried what you suggested and re-running some of the Western Steppe type models from my last comment with Botai in the pRight: Pastebin: https://pastebin.com/rfXMNsBZ & Imgur: https://imgur.com/a/Kpos3c2
Adding Botai into pRight just seems to make the p-value and CHI-SQ lower, not to move the proportions around at all.
Let me know if there's anything else specific you would suggest I'd try. I'm wary of trying anything that would seem to cut SNPs down too far at the initial stage (e.g. inclusion of any populations with low SNP and/or single samples).
(For example in the above set I used Israel Chalcolithic in the pRight rather than the really early and less admixed guys just because the published samples have low SNP counts that limit down the initial step... It seems like there's lot of unpublished versions of these samples where Reich lab have invested in getting more dna out of them, and mitigating contamination, though I expect they still won't get too much out of 'em )
@vAsiSTha, I might try that again; I tried with ShumLaka ("Cameroon_SMA" in the data file) in the first models I ran (it seems normal to include an African outgroup, and Iberomaurusian doesn't strictly fit the bill) but it didn't seem to change the fits at all (contrary to what I'd expect, since everyone seems to do it!). But I will re-run things with it in there at some point to see what happens, in case I've made a basic mistake here.
I'm a bit wary about adding some of the ancients (like Villabruna or K14) who are single samples in case it reduces SNPs (hence using Sunghir set rather than K14 for example) but this might be an irrational inexperienced thing to imagine is a problem and I'll give it a try.
@gamerz_j
Y-DNA P, Q, and R, are ANE/ANS in origin. Q has also been found in EHGs from Volosovo, and EHG-admixed hunter gatherers from Mesolithic Latvia (Kunda?). Q1a may have also been carried by the ANE sample Afontova Gora 2, but that sample was low coverage, I can't find anything else online officially confirming that assignment. However, as stated before, WSHG populations like Tyumen, they have ENA (East Asian-Devils Gate) related admixture. Khvalynsk doesn't have this East Asian admixture, and neither does Yamnaya, Srendy Stog, or Vonyuchka/Progress. I think some of the Devils Gate samples even had ANE/ANS admixture in them, though I could be wrong.
@ Vasishta
Even with allSNP:Yes, the Tail prob is >0,5
left pops:
NorthCaucasus_Steppe_Eneolithic
Russia_HG_Samara
Georgia_Kotias.SG
right pops:
Cameroon_SMA.DG
Turkey_Epipaleolithic
Russia_MA1_HG.SG
Israel_PPNB
Russia_Kolyma_M.SG
Iran_GanjDareh_N
Brazil_LapaDoSanto_9600BP
Russia_Kostenki14.SG
China_Tianyuan
best coefficients: 0.608 0.392
Tail Prob 0.094
Adding WSHG doesn't improve the fit.
These were probably fisher-hunter-gatherer groups on the Volga-Caspian steppe
@ gamerz_J
There were two different types of Denisovans. One Southeast Asian and another much further to the North. Which population contributed to Europeans if any ?
@ gamerz_J
What is interesting for me is that a very close relative of the Ancestor of Neanderthals, Denisovans and Modern Humans was found in Spain yet people keep on speaking about a East to West migration of these Archaic type Ancestry...I would like to see some kind of genetic evidence of Ancient samples showing how Neanderthals, Denisovans and Homo Sapiens ancestry gradually separated from one another. Samples between 700 000 and 400 000 years old will be very informative of what shared Ancestry really looks like and what actual Admixture looks like genetically.
"Adding Botai into pRight just seems to make the p-value and CHI-SQ lower, not to move the proportions around at all."
This usually hints at some botai affinity to the target.
@rob
Idk mate, I don't get anything above 0.01, and neither did Matt
@vAsiSTha, oops may have misspoke there, should have said "Adding Botai into pRight just seems to make the p-value lower and CHI-SQ higher" (i.e. fits involving WSHG as ancestor get a little bit worse) "not to move the proportions around at all.".
Extending the number of Prights, not much scope for extra WSHG
left pops:
Russia_Steppe_Eneolithic
Russia_HG_Samara
Georgia_Kotias.SG
Kazakhstan_Botai_Eneolithic.SG
right pops:
Cameroon_SMA.DG
Turkey_Epipaleolithic
Russia_MA1_HG.SG
Israel_PPNB
Russia_Kolyma_M.SG
Iran_GanjDareh_N
Brazil_LapaDoSanto_9600BP
Russia_Kostenki14.SG
China_Tianyuan
Turkey_N.SG
Russia_HG_Tyumen
Russia_MN_Boisman
best coefficients: 0.594 0.382 0.024
Tail Prob: 0.0878869
left pops:
Russia_Steppe_Eneolithic
Russia_HG_Samara
Georgia_Kotias.SG
best coefficients: 0.635 0.365
Tail Prob 0.122
https://pastebin.com/6STLzYLM
@vAsiSTha: I tried out a set of systematic tests with Cameroon_SMA_Published as an extra population and the first population when computing the f2 stats, then when running qpAdm.
Pastebin: https://pastebin.com/ndEZr9sv
Imgur: https://imgur.com/a/2NMCfTQ
It didn't really seem to change the p-values or CHI-SQ very much. Maybe there is some difference in how admixtools2 computes things and it is less important under this software, *or* maybe the idea that an basal outgroup is essential is more of a legacy of the first qpAdm models used, which were trying to work on inferences from modern day pRight, and it just doesn't contribute that much to most models of intra-Eurasian ancestry shifts when you have a meaningful and diverse set of ancient Eurasia pRight to use?
I don't think there's any harm in including Cameroon_SMA_Published and in my models it had no major effect on SNPs (where as we know, I think, from Harney's work there's generally isn't too much of an issue in relatively minor shifts of 5% less SNPs and things at numbers of +300k SNPs and such?). But it also didn't seem essential to me here. Models that work without it seemed to generally fit with it; there's some small shifts in the CHI-SQ, although this can take a passing p-value down to non-passing (on a note on that, in general, from the way these are behaving it seems like the order of magnitude of p-values tells us more about how close a model is than actually whether its 15% or 60%, or 7% or 20%).
Following on from my qpAdm results, though I'd see how the models I've been using with those outgroups compare to G25: https://imgur.com/a/8T7UhZD
It looks like G25 and my qpAdm results (best fitting only, p-value over 5% with their outgroups only) are rock solid agreement on the placing "East-West" on G25 West Eurasia PCA... But there are more differences where the G25 true samples for Steppe_LN / EMBA and Europe_MN/LN are projected "North" on the PCA compared to where their qpAdms lie. (Not really the case for Steppe_Maykop / Khvalysnsk which are even more rock solid in agreement). Maybe this is because the populations that contributed more real ancestry to Europeans are projected in a more European position (i.e. north) on PCA. The alternative is that the qpAdm is underweighting HG shift somehow... but only for Steppe_LN and Europe_MN, not Khvalysnk or Steppe_Maykop.
Rob
All my outputs are here, and I also paste result file link with each post of mine, maybe you should have checked that.
I don't know what you're doing wrong and I don't have time or interest to check your work, but suffice it to say that the model of ehg+chg does not work. Harvard still hasn't commented exactly what the chg/Iran like ancestry is, for that same reason.(although dzudzuana sample should have helped them understand)
https://pastebin.com/u/misnomer001#
@matt
Yeah I understood that you made a typo, but point remains that if p- value drops with additional right pop, you should probably move that from right pop to left and check.
As far as African outgroup is concerned it's just convention that's all, even I haven't seen it change results too much.
Neither CHG nor Iran_N is a good source for Yamnaya *on their own*
Laziridis on twitter, sep 2017
I think can't really distinguish between CHG and Iran Neolithic as Yamanya source.
Iain Mathieson - sep 2017
I mean Yamnaya=EHG+(Either CHG or Iran_N or both or something similar)
Mathieson Sep 2017
I don't know their updated positions, certainly hasn't been any new published data to change this position.
So Rob, the matter is only sorted in your head.
@rob
Nick Patterson on the issue, from a post on this blog.
"I am much less sure about the origins of Yamnaya than
the blog seems to think.
There's no doubt that Yamnaya has lots of "Iran-Related" ancestry but so does ancient CHG. I'm thinking about this right now. Is there a paper describing a scenario for Yamnaya formation? If so I don't remember it."
From
https://eurogenes.blogspot.com/2019/08/did-south-caspian-hunter-fishers-really.html?showComment=1565880175620&m=1#c1508598484628353588
Really let's not belabour the issue, fact is that we sorely lack samples from places and eras of interest to figure out what's going on. Need more chg samples, need dzudzuana published and need 5000-6000bce samples from cultures like jeitun, east of Caspian.
Matt,
I think of it like this. WHG and MA1 are older North Eurasian populations. EHG then moves North, later followed by Steppe groups.
So there are basically 3 'clusters'. If this is right the older MA1 would be a better source for some of your models instead of WSHG.
Basically WSHG has no impact on anything West of it, but any additional ANE (on top of EHG) in these Western pops that can be modelled is just old affinity from before WHG separated from ANE.
In this case it would make sense that the fit would get lower once Botai is introduced in pRight, because it is close enough to WSHG.
So it would be interesting to see what those models show if you replace WSHG with MA1, which is much older than the targets and WSHG.
@ Vasistha
Lol was just paying you lip service; I don’t need you to check my models; your Pc keeps crashing; your self professed expertise verges on arrogance; & you seem to have a proclivity to distort the data. Hence your qpAdm models are empty.
As I told you , the debate about CHG<->Iran & Eneolithic steppe (whilst interesting on its own accord) is actually irrelevant to the issue of PIE because they’re both peripheral to the issue. And I say this as someone who was quite open to the possibility
There’s nothing in Matt’s models which contradicts this reality, he’s simply experimenting here.
For interest:
Standard model
left pops:
Russia_Kalmykia_EBA_Yamnaya.SG
Russia_Steppe_Eneolithic
Ukraine_Globular_Amphora
(With same pRights)
best coefficients: 0.867 0.133
TP 0.0591061
Historic model
Russia_Kalmykia_EBA_Yamnaya.SG
Smyadovo_C_o
Russia_Steppe_Eneolithic
best coefficients: 0.644 0.356
TP 0.325472
SNPs 228511, transversions only
@ Vasistha
“ Need more chg samples, need dzudzuana published and need 5000-6000bce samples from cultures like jeitun, east of Caspian.”
They would be great but irrelevant . These CHG people in the Volga-Caspian steppe were fisher /hunters; not SCA dry-land farmers -pastoralists
No rob the issue is that your model is trash.. you see, if it works for you does not mean it is valid. It should work for all reasonable sets of right pops..
As far as your comments about my computer crashing is concerned, it is nothing but pitiable and irrelevant ad hominem very expected from someone like you.
@ Vasistha
The CHG/ EHG model isn't 'mine', its the received convention. But whatever the exact composition / flavour of these hunter-gatherers was, it wont change the overall picture.
Y-hg provides the sanity check. The Khvalynsk-Caucasus steppe Eneolithic features J1 & R1b-V3616, both of which are found in quasi-contemporaneous groups from the Caucasus & even East Anatolia, whilst they are lacking in the Turan region. How do you explain that ?
@vAsiSTha: OK, I'll try some the models reversing the positions of WSHG and Botai and see what happens. I think when you add Botai into the outgroups, it's actually that the models with WSHG fit slightly worse because Botai has WSHG ancestry, and that models with Botai and WSHG switched would become worse because Botai has that extra East Asian ancestry that most WSHG descended populations that have Steppe_Eneo ancestry (or blend) do not have. But I'll see what happens and print it up.
@mzp1: I'll try it with MA-1+AG3 as a source after trying vAsiSTHa's model; I'll see if using that as a merged population still allows for enough available SNPs that I'm happy running the models.
@Rob and vAsiSTha: Yeah, I am just experimenting with different models at the moment. I'm not seeing in the models that there is too strong a basis to prefer purely EHG+CHG models for Steppe_En or models for Yamnaya that are definitely Steppe_En+influences from the west to ones that have different blends. This could be a limitation of the outgroups I've used, particularly due to low levels of sampling for some different groups. However they generally seem to give results consistent with G25 and published material for comparable models (some differences with G25 raising questions to me about what PCA projection differences might emerge for groups that are actually ancestral to modern people). It seems likely to me still that a safe proposal is some combination strong basis on the steppe+Volga Delta in maybe different groups there and also general wider connectedness increasing to a small but noticeable level over time.
Obviously if what Davidski has said before and the Sredny-Stog from 2-1kya earlier are consistent with being 100% ancestral to Yamnaya, then that would refute some of the specific historical models that have come up and some of the ones I've run for the Steppe_EMBA populations. It should be easy to test if those Sredny-Stog samples are autosomally the same as Yamnaya or just in a superposition on PCA by running a 1-pop qpAdm with a good set of outgroups (Like the 1-pop qpAdm I was running in the examples of testing with and without Cameroon_SMA, where passing p-values happened when modelling Yamnaya as 100% Afanasievo but we started to get failing values when looking at modelling North Caucasus culture, Steppe_Eneolithic itself, etc as 100% Afanasievo - the same thing should be possible in principle with these Sredny-Stog samples. It's basically the same thing as we maybe could do with qpWave but I think its easier to understand and compare doing with qpAdm).
https://www.yfull.com/tree/r1b/
Basal R1b is found in Dzharkutan sample I4315
@vAsiSTha: I tried the models again that I ran earlier, flipping WSHG and Botai between pRight and pLeft.
Pastebin: https://pastebin.com/4JWmEkHZ
Imgur: https://imgur.com/a/zqIv6UQ
In general, models with the Botai in place of WSHG are all worse, and in some senses this pushes them to be not the best models in competition with models without Botai. The model fits for Steppe_Maikop which has a lot of WSHG like ancestry tend to get a lot worse in p-value and blow-up in CHI-SQ. Some models with Botai still work without too much p-value difference, but these are models where they are assigning like 5-7% WSHG (e.g. since Botai is like 92:8 WSHG:Extra _East_Asian its a difference between a model with like 0.5% difference in East Asian ancestry, so doesn't make anything blow up too much, just a little worse).
I ran these on a different f2 precomputation run than the previous set which I linked for mzp, and in some instances this changed the best-model from one with WSHG to one without. That kind of shows that some of these things are on the edges of whether they are better or worse or no difference I guess...
@Simon Stevin
"Y-DNA P, Q, and R, are ANE/ANS in origin. Q has also been found in EHGs from Volosovo, and EHG-admixed hunter gatherers from Mesolithic Latvia (Kunda?). Q1a may have also been carried by the ANE sample Afontova Gora 2, but that sample was low coverage, I can't find anything else online officially confirming that assignment. However, as stated before, WSHG populations like Tyumen, they have ENA (East Asian-Devils Gate) related admixture. Khvalynsk doesn't have this East Asian admixture, and neither does Yamnaya, Srendy Stog, or Vonyuchka/Progres"
Has Volosovo been tested for East Asian-related admixture? One can model EHG in Global25 without it, but using Tyumen_HG, they get some. I really don't know how accurate that is. Lazaridis (2018) had EHG be 3% Han-related tho p-value of the qpAdm output could have been greater.
I am not aware of ANE in Devils_Gate_N but wouldn't be surprised if a bit was in there. I am not saying there is East Asian in Yamnaya and Sredny Stog, I am just wondering if there is any in EHG or not. I can't run anything more advanced than G25 to check.
@Ric Hern
"There were two different types of Denisovans. One Southeast Asian and another much further to the North. Which population contributed to Europeans if any ?"
It appears that the northern type. Whatever was in Han anyways, they seem to get some Papuan admixture bringing southern too, but in the Salkhit paper the Denisovan in Han was different than that in Papuans so idk.
Anyone knows of any drawbacks in studies using SFS?
@Rob
"That’s not basal, but unresolved . You’re a muppet"
Wow i think i hit a nerve. I guess you yourself are on the R1b line? Thats why it hurts? lol
I have the snp file and Y calls of I4315 (because i know how to do that and check for myself). See for yourself whats unresolved.
https://drive.google.com/file/d/1ZGfTxqo7nKaqqnHXDdExJMYbH1AWDyf9/view?usp=sharing
There is not a single valid call below R1b which is +ve. All R1b1 (10 snps) and R1b2(3 snps) calls are -ve.
This is why the samples has been put by Yfull under R1b and not R1b*
European R1b lines, including M269, are rooted in Europe since the Paleolithic.
So the R1b in Yamnaya, Bell Beakers, etc. has nothing to do with Central Asia.
@Rob
"It’ll be PH155/ 200"
Keep whining crybaby. because it is negative the equivalent 3 SNPS at R1b2 level of PH155
rs779989989 G>A Found G
rs978669961 G>A Found G
rs1055222056 G>C Found G
@matt
"In general, models with the Botai in place of WSHG are all worse"
Yeah i dont know which targets you were testng. But general principle as per Harney is that if an addition to right pop makes fits significantly worse then it means that that pop is 'PROBABLY RELATED' to a source which is missing in the left pops. Thats all i meant to say
Bronze Uzbekistan Dzharkutan [I4315 / UZ-JAR-004, Jarkutan 4b-85, Grave 60 ] 1609-1465 calBCE (3255±15 BP, PSUAMS-2518) M R1b1(xR1b1a) pre-PH155 R2+13500+195 Narasimhan 2019
@mzp1, I merged the MA1 and AfontovaGora3 samples into a single population label, then tried the previous run again with them included. It cut down the SNPs quite a lot (from "342608 SNPs remain after filtering. 309804 are polymorphic" to "286742 SNPs remain after filtering. 259142 are polymorphic."). I then tried running the same 5-way models that I tried before (with source Georgia_CHG.SG, Serbia_IronGates_Mesolithic, Russia_HG_Karelia, Russia_West_Siberia_HG, Serbia_En, for targets Russia_Steppe_Eneolithic, Russia_Khvalynsk_Eneolithic, Russia_Steppe_Maikop, Russia_Afanasievo, Ukraine_N).
It didn't give anything like swapping "MA1_AG3" fraction for WSHG and improving fits, but gave quite different and worse fitting results and led to lots of infeasible results with a negative "MA1_AG3" and a very high Karelia_HG fraction. Also the case when tried limiting populations and using 3-way models and the like. It seems like with these outgroups MA1+AG3 is distinct from just being WSHG without as much of the ENA ancestry in ways that cause some problems.
Looks fairly clearly like some dead end pre-PH155 to me, plenty of ancestral calls under R1b1 so it's irrelevant for European R1b.
http://sprunge.us/n5RsDk
Evidence for early dispersal of domestic sheep into Central Asia
Shnaider et al April 2021
The development and dispersal of agropastoralism transformed the cultural and ecological landscapes of the Old World, but little is known about when or how this process first impacted Central Asia. Here, we present archaeological and biomolecular evidence from Obishir V in southern Kyrgyzstan, establishing the presence of domesticated sheep by ca. 6,000 BCE. Zooarchaeological and collagen peptide mass fingerprinting show exploitation of Ovis and Capra, while cementum analysis of intact teeth implicates possible pastoral slaughter during the fall season. Most significantly, ancient DNA reveals these directly dated specimens as the domestic O. aries, within the genetic diversity of domesticated sheep lineages. Together, these results provide the earliest evidence for the use of livestock in the mountains of the Ferghana Valley, predating previous evidence by 3,000 years and suggesting that domestic animal economies reached the mountains of interior Central Asia far earlier than previously recognized. Archaeological and biomolecular investigations of ancient sheep remains from the site of Obishir V in southern Kyrgyzstan reveal that domestic livestock and Neolithic lifeways reached the heart of Central Asia by ca. 6,000 BCE, thousands of years earlier than previously recognized.
https://www.researchgate.net/publication/350741147_Evidence_for_early_dispersal_of_domestic_sheep_into_Central_Asia
Yes, I know, and according to vAsiSTha these sheep carried Indo-European languages into Europe.
You can, of course, joke, but the facts are as follows. At the Orlovka sites of the Varfolomeevka and Oroshaemoe cultures, the first domesticated animals were only sheep in the sixth millennium BC. Almost simultaneously, at the beginning of the fifth millennium, domesticated animals also appear on the Don on RYar, but quite different - cows, pigs and goats. I think these are two different sources of domestication. In the Volga region from Central Asia, and on the lower Don, possibly from the Balkans or the Caucasus.
But the fact is that European subclades of R1b are rooted in Europe since the Paleolithic.
Also, there were hunter-gatherers rich in CHG-like ancestry on the European steppe, and Yamnaya doesn't have anything to do with Central Asia or even Khvalynsk.
So there's no need to bring Central Asian sheep into the argument. Let's just stick to human DNA from Europe.
Im completely lost what this R1b in a BMAC sample proves. They all have fairly recent ancestry from central asian foragers, probably more of a Kumsay profile than straight up WSHG.
Those populations have steppe_eneolithic like ancestry, and WSHG proper has EHG like ancestry.
A divergent R1b clade carried by someone with ancestry from cousin populations of EHGs, gee wiz what a big deal.
@ Daviski
''So there's no need to bring Central Asian sheep into the argument. ''
And there isnt, because Vladimir's understanding of the facts is rather... sheepish
He claims that ''Varfolomeevka and Oroshaemoe cultures, the first domesticated animals were only sheep in the sixth millennium BC. ''
Vybornov, in his article claims
A production economy is a particular feature of the second group of sites, which can be dated to the end of the 6th millenium BC.
Specifically referring to ''Remains of domesticated animals were also recently found at Prikaspiiskaya sites, preceding Khvalynskaya culture''
The Varfolomeevka bone isnt even carbon dated, whilst in the article's table the earliest sheep bone dates to 4724–4557 [UGAMS–23059] from the site of Oroshaemoye, which is multiphased, with Khvalysnk overlying Precaspian phase.
In reality, this is just an Eneolithic sheep. Nothing new
''according to vAsiSTha these sheep carried Indo-European languages into Europe.''
They apparently flew from Krygysztan to Ukraine with Caucasian Hunter Gatherers.. because his super-dooper qpAdm model says so
and the reason why sheep/ goats characterise the Volga-Caspian steppe is because they are more suited to its more arid conditions, whilst a greater variety can be sustained further west
Looks to me if these sheep were ancestral to Tibetan sheep and maybe related to the Hungarian Zekel type sheep which arrived with the Magyars. European sheep are clearly distinct from these sheep. My bet is that sheep in and near the Pontic Caspian Steppe were of the Northern European Shorttailed type and more specifically ancestral to the Romanov breed.
Sheep were taken into Central Asia by Afanasievo people, and these were European sheep.
So the migration of sheep between Europe and Central Asia went from west to east, same as that of people.
Lol davidski where do you get this bs from?
Jeitun has domesticated sheep already by 6000bc. Afanasievo is dated to when? Lol
"Im completely lost what this R1b in a BMAC sample proves."
It was a response to Robs question. It was meant to prove nothing but that some basal lineages of R1b did exist is central Asia. From which date? That is unknown.
As soon as I gave the example, rob started insulting, because he has no more questions to ask.
MA1 is too undrifted, AG3 is the only proper ANE sample that has drifted.
@vAsiSTha
Unlike you, I'm actually familiar with the genetic phylogeny of sheep breeds.
Like I said, at 3,000 BCE European sheep moved into Central Asia, while Central Asian sheep never moved into Europe.
Haha.
I found this video by D. Anthony on where the CHG component of Yamnaya comes from, though you all probably have seen it already. He seems to suggest to think it's from near the Volga someplace (having presumably moved north from the Caucasus) , though he also says more samples are needed.
https://www.youtube.com/watch?v=-MOrEA84qvo
@vAsiSTha
This question?: "The Khvalynsk-Caucasus steppe Eneolithic features J1 & R1b-V3616, both of which are found in quasi-contemporaneous groups from the Caucasus & even East Anatolia, whilst they are lacking in the Turan region. How do you explain that ?"
A MBA sample with basal R1b/pre-PH155 is not really an answer to that question to be honest.
Sorry correction: Zackel/Racka Sheep ...arrived in Europe with the Hungarian Language. This is the only Central Asian related sheep most probably related to Tibetan sheep which arrived in Europe I think.
"Haha."
This is what I come here for
"A MBA sample with basal R1b/pre-PH155 is not really an answer to that question to be honest."
The question itself is stupid because steppe_en has a sum total of 2/3 male samples. To prove or disprove anything based on sex chromosome markers is fruitless on that sample size.
So what happened to the Y-chromosomes and sheep of all of these Neolithic Central Asians who migrated into Europe?
How come there's no sign of anything Central Asian in Yamnaya, Corded Ware, Unetice etc. and European sheep don't derive from Central Asian sheep?
Did these questions slip your mind?
@davidski
"Sheep Mitochondrial DNA Variation in European, Caucasian, and Central Asian Areas "
https://academic.oup.com/mbe/article/23/9/1776/1014282
"Four (A, B, C, and D) highly diverged sheep lineages were observed in Caucasus, 3 (A, B and C) in Central Asia, and 2 (A and B) in the eastern fringe of Europe, which included the area north and west of the Black Sea and the Ural Mountains."
https://academictimes.com/researchers-find-evidence-of-domestic-livestock-in-central-asia-far-earlier-than-previously-thought/
"The Obishir V sheep (6000bce) were most similar to SW and central asian sheep breeds and mit haplogroup A, which is oldest haplogroup for domesticated sheep in the region"
From https://core.ac.uk/download/pdf/199416099.pdf
"Today, the most common sheep mitochondrial haplogroups are A and B, with A
dominating in eastern and southern Asia, while B is the most
common haplogroup in Europe and western Asia (Lv et al., 2015;
Pedrosa et al., 2005; Tapio et al., 2006)."
Where did European demosticated sheep maternal hg A come from? do you have a study conclusively proving the origin of this ancestry into eastern europe?
I thought Sheep domestication went from Iran to Europe. Europeans independantly domesticating sheep is not realistic given ancient Europeans are genetically North Eurasian and both North and East Eurasian ancient assemblages include Cattle and Pigs but not Sheep/Goats. Euro Neolithic sheep would derive from Anatolian with ultimate origins further East.
"MA1 is too undrifted, AG3 is the only proper ANE sample that has drifted."
It is undrifted, thats why MA1 should be the best source of ANE, AG3 is too drifted and seems to be much further away from West Eurasians than Botai, while MA1 should be the closest.
AG3 separates it from everyone else, it seems to be a dead-end if I'm not mistaken. Botai may have South Central Asian ancestry like CHG that keeps it more Central and less drifted.
@All,
I am going to post a youtube video, on the origins of Yamnaya's CHG ancestry in two days. I am going to reference this blog as my source.
@gamerz_j,
Yeah, that video is precious gold. Thanks for showing it so others can see.
Anthony has Neolithic DNA from Volga who resembles the most CHG rich Khavlnsky individuals. What Anthont suggests based on this data, is 6200 BC hunter gatherers from Northwest Iran migrated into Volga. *Yes, we are hearing about Iran again ;)
But I hope people don't take his theory of a 6200 BC migration too seriously. It has no hard evidence. He is too committed to the assumption, CHG had to of migrated to the Steppe shortly before the oldest CHG rich samples lived.
All we know is Yamnaya-like hunter gatherers lived on Volga far north of the Caucasus. This can be explained by multiple explanations.
It is just as probable that this means, Yamnaya-like hunter gatherers lived throughout PC Step since 8,000 BC.
Another quick qpAdm experiment, unrelated to anything previously discussed, on the topic that comes up of differentiation between East and West Europe today, using qpAdm to look at the question this time.
In brief, what I find is:
1) When using a deep set of pRight outgroups that is dated to the Early Neolithic or before (with addition of Eneolithic Steppe), Western and Eastern samples with similar proportions of WHG:ANF:Steppe (or combinations that can approximate each other), can model in qpAdm with high p-values as each other, irregardless of geography.
So this is what we expect if we think that they have more or less similar proportions of these underlying populations.
2) But, when I introduce a *single* Middle Neolithic farmer population, Spain_MLN, these fits immediately switch from high p-values to low, non-passing ones.
3) Introducing Afanasievo and Bell Beakers doesn't do too much to this, however introducing Latvia_BA as an pRight outgroup strongly confirms the difference.
Imgur: https://imgur.com/a/Gwr8evy
Pastebin: https://pastebin.com/y8RLjjEn
So that adds a bit more weight to me a bit more formally that it is *both* of differing affinities to Middle Neolithic farmers and drifts that we only see later in the Bronze Age (wherever and whenever they happened) that can today start to distinguish Eastern and Western European genetic paths. It's not all one thing or the other... Something like this might also suggest that we could look at the question of "Balto-Slavic drift" (and to some extent "Latin-Germanic-Celtic drift" or something) a bit more through formal models, if we wanted to, and it's not just reliant on PCA (even if that is an effective way to look at it).
A bit non-exhaustive, so more could be done to confirm and flesh out I guess.
@ Vasistha
'' It was meant to prove nothing but that some basal lineages of R1b did exist is central Asia. From which date? That is unknown.
As soon as I gave the example, rob started insulting, because he has no more questions to ask.''
LOL you still don't get it. There are no basal subclades of R1b in CA, Ph155 is a collateral clade.
If you don't even understand basic Y-hg phylogeny, then you certainly not in a position to claim that you overall platform is correct. You're simply full of yourself. The fact that matt agrees with you means nothing, in my book
And no, i'm not R1b , so all your insinuations are false. Don't throw stones if you live in a glass house, princess
@Genos Historia
I am going to post a youtube video, on the origins of Yamnaya's CHG ancestry in two days. I am going to reference this blog as my source.
Wait for the weekend.
I'm going to post everything that I know on the topic here and also include a juicy PCA that is much better than what David Anthony showed in his clip.
@ Genos
“ But I hope people don't take his theory of a 6200 BC migration too seriously. It has no hard evidence. He is too committed to the assumption”
Anthony’s perspective on the caspian region Mesolithic and Neolithic is good & informativ. He has noted the continuity of lithic industries since Mesolithic , even LUP. The only thing which changes is pottery uptake . This might or might not have entailed some form of shifts in settlement or social networks; but exactly which point on the caspian these came from isn’t going to be a game-changer
@Davidski,
Alright, if you do that, I'll make your post/theory the study piece of my video. How new is it going to be?
Good news, is I think I can get Survive the Jive to share my video. He shared one of my Viking videos and it got 12k views. But let's see.
It's going to have all the stuff that Anthony talked about, plus some details that he didn't mention, probably because nobody at Harvard told him, possibly because they don't know themselves.
@Rob,
I know Anthony is an archaeology expert, but I his genetic explanation doesn't fit the typical rules of human population genetics.
It would be weird if CHG migrated straight into the Volga in 6200 BC. And skipped over all of the Step land south of it.
If CHG ancestry existed in the Volga it probably existed everywhere south of the Volga all the way to the Caucasus. He seems to think this Volga site represents descendants of a CHG colony that was isolated in Russia.
Insulated colonies like this aren't attested in human population genetics. ANE ancestry didn't created colonies. Anatolian farmers didn't create insulated colonies.
Unless, the Volga Neolithic guys are like 80% CHG, we have no reason to believe the migration of CHG went directly into the Volga. Or that it happened recently, as he claimed.
Maybe Anthony still believes that Steppe Maykop represents the hunter-gatherers of the North Caucasus?
That was a really weird theory. Totally upside down to reality.
Yeah.......about his theory
Archaeologists don't have a good track record when it comes to predicting population genetics.
So when, David Anthony starts making suggestions based on archaeology, I'm going to take it with a grain of salt.
No offense Rob.
My theory on Anthony's theory is....
He is looking for a Volga origin of Proto-Indo Europeans. He probably thinks it goes.....
CHG directly to Volga Neolithic.
Volga Neolithic to Khvalynsk.
Khvalynsk to Sredny Stog.
Srendy Stog to Yamnaya.
They're still direct migration from the Caucasus, being what gave rise to the Proto-Indo Europeans. Even after a Eneolithic migration being excluded.
They haven't considered yet that CHG and EHG mixed people could have been indigenous to the Pontic Caspien Steppe.
@Genos Historia
Have you read David Anthony's works? His preference for the Volga and the Samara valley region is super apparent there, but it probably comes from him having done the most work in those regions. It isn't really a novel angle he is fishing for or anything.
Also I just have a feeling that the comment section of a video about the "CHG" ancestry in Yamnaya is going to be rife with nationalist screeching and serious misunderstandings about the topic.
@Sam/Genos, I suspect that in the end Anthony will be right that the Khvalynsk Culture and the EHG, did contribute to the later steppe cultures. Not a "dead side branch" of a true source that is true contributor to later cultures and that was exactly like Steppe_En and had been for thousands of years. And the Khvalynsk Culture does seem to be admixing between EHG and something else... The full set of Khvalynsk samples and their dating will probably tell us how this happened.
@Matt
Khvalynsk and Yamnaya seem to have different origins in that a large part of their hunter-gatherer ancestry is different.
The hunter-gatherer input in Yamnaya has more WHG, and I'll explain this in more detail later this week.
So Anthony was actually correct when he called Khvalynsk the "uncle" of Yamnaya, because the CHG in Khvalynsk and Yamnaya is probably from the same population, but this population diverged during the Eneolithic to form at least two separate groups and also clines.
There seems to be an obsession with getting Iran into the IE origin story by some academics but a lack of a justification just a need to get a component from as far south as possible.
Davidski, I think we'll need the full set of Khvalynsk cemetary samples that Reich lab has to accurately characterise them. The HG ancestry in the 3 Khvalynsk samples looks slightly more western than in the 3 Steppe Eneolithic to me so far from qpAdm. But those 3 K samples probably have quite a bit more EHG than the full set to be published.
@Palacista
Well, what David Anthony is doing is reacting to what David Reich, Nick Patterson and Iosif Lazaridis are saying, because he trusts them as top population geneticists.
So when they say that Iran is a very plausible source of the CHG-related ancestry in Khvalynsk and Yamnaya, then Anthony probably feels that he has to accommodate this somehow.
He can't check that CHG and Iran_N form different clades in admix graphs and fine scale PCA. To him this is Voodoo magic basically. So he even goes so far as to say that CHG lived in Iran.
The chances that his theory will work out are very slim. At best, and this is the very best scenario for him, the type of CHG-like ancestry that diffused into the steppe a long time ago came from a territory that roughly included a small part of what is now northwestern Iran.
Think Anthony called Khvalynsk "uncle" in the sense y-dna was different, not due to autosome, can't really check as presentation gone down.
@Matt
It makes no difference, because like I'm saying, Yamnaya doesn't seem to derive from Khvalynsk itself, but from a related, more CHG-rich population that also contributed to the formation of the more easterly Khvalynsk cline/cluster.
"Well, what David Anthony is doing is reacting to what David Reich, Nick Patterson and Iosif Lazaridis are saying, because he trusts them as top population geneticists."
Same reason you had Kristiansen going about Maykop being archaic PIE with middle-to-late PIE being transferred to steppe_EMBA populations by way of those Steppe Maykop populations.
It's unfortunate really. A completely unmecessary goosechase in my opinion, and they could have spent that time actually getting closer to solving more interesting things when it comes to the story of the Indo-European migrations.
The recent greek paper contains many strange thing but at least it is accurate in saying that pontic-caspian steppe groups that reached the greek peninsula were mostly of euro-hunter stock
IIRC they have EBA steppe modeled as 65/70 EHG and the rest CHG like.
This population existed in the pontic steppe pretty much already in the neolithic. In fact you have ukraine neolithic samples (5000 BC) that are roughly 80 per cent WHG/EHG and 20 Chg like. Obviously these groups are ancestral respectively to Sredni Stog and Yamnaya.
@Palacista,
Harvard's search for an Iranian ancestor of Yamnaya, I think began with this paper in 2016.
Lazardis 2016. Genomic insights into the origin of farming in the ancient Near East
https://www.nature.com/articles/nature19310
In the paper, you see they had a philosophy that Neolithic SW Asian farmers are the core of modern Western Eurasia. So, linking Iran farmers to Yamnaya was just natural.
@Davidski
"The hunter-gatherer input in Yamnaya has more WHG, and I'll explain this in more detail later this week."
Are you referring to the HG input unrelated to Progress?
Something like this: https://i.imgur.com/yBKyJFi.png
Many scientists from the MPI were supporters of the Anatolian hypothesis, in 2012 there were still articles coming from that corner in support of it.
I think they were talking about armenia/iran since like 2014, they definitely were doing so when Wolfgang Haak's article was released in 2015.
@CrM
Yes, the so called EHG in Yamnaya on top of what it got from its Progress-like ancestor was more western.
That's why the whole Yamnaya cline doesn't sit perfectly parallel to the Khvalynsk/Volga cline, but it's actually tilted towards WHG.
@Davidski
Do you think any of the Y-DNA in Yamnaya and CWC derive from this HG population?
Yes, I think at least R1a-M417 is from this forager population.
@davidski, sure, I get the argument, but it would be easier to compare it with the full sequence of Khvalynsk that Reich lab have, that we saw in the presentation PCA...
Say you have a shift in the sequence from the cemetary from EHG like to more Progress like over time and this happens before Progress samples...
@Rob
So you think that proto IE is I2a?
@Matt “ I suspect that in the end Anthony will be right that the Khvalynsk Culture and the EHG, did contribute to the later steppe cultures. Not a "dead side branch" of a true source that is true contributor to later cultures and that was exactly like Steppe_En and had been for thousands of years. And the Khvalynsk Culture does seem to be admixing between EHG and something else... The full set of Khvalynsk samples and their dating will probably tell us how this happened.”
Cant be possible: Khvalynsk had WSHG/Q1a, which was clearly lacking in contemporaneous Sredny Stog and in later Yamnaya and CWC, while at the same time lacked all the EEF component. If anything, then it was a west to east migration of the Dnieper Steppe cultures into Samara and Khvalynsk area of Europe.
Andrzej: I don't see too much evidence of an large excess of WSHG that is inconsistent with passing models for ancestry of other populations, and we still have not much of the full set of Khvalynsk samples. Y-dna haplogroups can be lost/acquired and Anatolian farmer component can be acquired too. We'll see what proves to be possible with the fuller sample set...
I don't think there's any evidence at this time that cultures at Dneiper had CHG / CHG+EHG like genetics, so there is nothing at Khvalynsk genetic sequence that can look like migration from there at the moment. Maybe such samples will turn up, or not.
@Old Europe “
The recent greek paper contains many strange thing but at least it is accurate in saying that pontic-caspian steppe groups that reached the greek peninsula were mostly of euro-hunter stock
IIRC they have EBA steppe modeled as 65/70 EHG and the rest CHG like.
This population existed in the pontic steppe pretty much already in the neolithic. In fact you have ukraine neolithic samples (5000 BC) that are roughly 80 per cent WHG/EHG and 20 Chg like. Obviously these groups are ancestral respectively to Sredni Stog and Yamnaya.”
I cannot reconcile it at all with any scientific paper that came out since and including Lazaridis 2015 concluded that any WSH population that may be ancestral to any modern IE, namely Yamnaya, CWC and Sredny Stog - had an almost even mix of EHG : CHG.
Any pop tested consistently for these ratio, with and w/o the EEF and WHG introgression.
@Davidski “ So Anthony was actually correct when he called Khvalynsk the "uncle" of Yamnaya, because the CHG in Khvalynsk and Yamnaya is probably from the same population, but this population diverged during the Eneolithic to form at least two separate groups and also clines.”
Aren’t Khvalynsk and Yamnaya both (at least partially) from a Pogress/Piedmont/Vonyuchka pop that formed right at the foothills of the Caucasus and was already an amalgam of both EHG and CHG?
@Rob “ Anthony’s perspective on the caspian region Mesolithic and Neolithic is good & informativ. He has noted the continuity of lithic industries since Mesolithic , even LUP. The only thing which changes is pottery uptake...”
As much as I respect Anthony’s theories and share his Steppe origins of PIE point of view, I couldn’t but notice that lots of his notions are utter rubbish, and have been debunked time and again. His 2008 book “horses, steel and languages” had followed in Gimbutas’ footsteps in labeling GAC as Indo-European, he claims that Botai or BMAC were Indo-Aryan IIRC and other stuff. I acknowledge that he wrote his book 2 years before Reich and Patterson started using aDNA to actually source pop genetics, but...
And in 2019 he wrote that PIE was an EHG language, without any real basis (if I had used his reasoning then Ossetians, who inherited Scytho-Sarmatians from their maternal mtDNA but shared Y-dna with their non-IE Caucasus neighbors, should’ve spoken NW Caucasian instead of an Iranic branch of IE).
@Matt
Say you have a shift in the sequence from the cemetary from EHG like to more Progress like over time and this happens before Progress samples.
It might not be as simple as that, because some of the oldest Khvalynsk samples have the highest levels of the so called CHG ancestry.
Interesting new Ydna paper on NE Europe:
"Phylogenetic history of patrilineages rare in northern and eastern Europe from large-scale re-sequencing of human Y-chromosomes"
https://www.nature.com/articles/s41431-021-00897-8
@Matt
"I don't think there's any evidence at this time that cultures at Dneiper had CHG / CHG+EHG like genetics, so there is nothing at Khvalynsk genetic sequence that can look like migration from there at the moment. Maybe such samples will turn up, or not."
Well there is this I guess:
https://eurogenes.blogspot.com/2020/05/a-significant-finding.html?m=1
@Matt
Do you think there were WSHG populations that didn't have East Asian ancestry, unlike Tyumen and Botai? If so, would this hypothetical population resemble EHGs with a lot more ANE ancestry, like Steppe Maykop? Khvalynsk and Vonyuchka/Progress don't have literal WSHG ancestry, but they have ancestry from EHGs with a lot of ANE. These hypothetical EHGs could partly resemble what West Siberian Hunter Gatherer populations looked like before admixture with East Asian-derived foragers, which occurred sometime in the late Mesolithic and early Neolithic.
@Davidski, I don't think that's totally unexpected as we often find that high fraction samples come in before things homogenize (but aren't necessarily unadmixed when we find them), but yeah, OK. Anyway, let's leave that topic for now (maybe I will even avaoid until until the samples come out and we can say more about them).
...
On another unrelated topic, another thing with admixtools2, kind of an obvious observation, lots of complaint about the recent Greek paper's assumption of continuity, and some folks have responded to the authors that "You should look at high dimensional PCA and see the Balto-Slavic drift".
That's one response and it makes sense. On another tip, just looking at populations with similar deep ancestry proportions, and using admixtools2 to run large number of f4 stats comparing populations with similar (not identical but similar) deeper ancestry proportions: https://imgur.com/a/dOBWnez
It does seem relatively straightforward to get some pretty robust and strong Z scores off of f4 stats for populations comparisons like Latvia_BA, Spain_MLN, England_BellBeaker, for north/central European anyway.
The problem with this approach in Southern Europe might be finding comparison populations, because there's some differences in deeper proportions going from Italy to Greece and so on, and you might confounds between ancient and modern samples: https://imgur.com/a/lkzHiCD
But still it might be useful to look into this, as authors are often skeptical of PCA projection, so formal stats might provide a more solid way of dealing with continuity.
@andre, Old Europe,
"The recent greek paper contains many strange thing but at least it is accurate in saying that pontic-caspian steppe groups that reached the greek peninsula were mostly of euro-hunter stock
IIRC they have EBA steppe modeled as 65/70 EHG and the rest CHG like.
This population existed in the pontic steppe pretty much already in the neolithic. In fact you have ukraine neolithic samples (5000 BC) that are roughly 80 per cent WHG/EHG and 20 Chg like. Obviously these groups are ancestral respectively to Sredni Stog and Yamnaya.”
I cannot reconcile it at all with any scientific paper that came out since and including Lazaridis 2015 concluded that any WSH population that may be ancestral to any modern IE, namely Yamnaya, CWC and Sredny Stog - had an almost even mix of EHG : CHG."
The bolded is correct, see my earlier comment, and the below.
"...The Western End of this cline is actually represented by ancestry in Kumtepe_N, Maykop, Yamnaya Ozera, Yamnaya Bulgaria and related indivdiuals."
This will be the type of PC Steppe ancestry in Greece, Europe Neolithic. These are highly drifted PC Steppe populations.
"Obviously these groups are ancestral respectively to Sredni Stog and Yamnaya.”
They are not ancestral to Yamnaya, they left much earlier and therefore have since drifted quite a bit and therefore are quite far from Yamnaya and other populations. The separation starts very early with Kumtepe N furthest on this cline. I've already talked about this cluster here in the comments before.
But I am not sure how they can have WHG, because this cluster doesnt seem closer to any Euro HG or anything else afaik, they are just a drifted Steppe-like population.
@Davidski @CrM “
That's why the whole Yamnaya cline doesn't sit perfectly parallel to the Khvalynsk/Volga cline, but it's actually tilted towards WHG.”
Rob would salivate at this proposal: he has been entertaining the notion of PIE as a WHG language for years!
Which reminds me that Drago used to post hypotheses in which PIE might’ve come from the Tripolyan language, or zardos thinking that GAC were speaking a WHG and not an Anatolian Farmer one because of their y-dna I.
It would be cool if any WHG or Farmer language exerted a strong influence on PRe-Proto-PIE.
@ Andrze
@ Vlad
“ So you think that proto IE is I2a?”
''Rob would salivate at this proposal: he has been entertaining the notion of PIE as a WHG language for years''
Yes, yes you're both comedians. hg I2a is at least as old as the Ice Age, and I don't have any PCT or nostraticist convictions for PIE, hence wouldnt make such an equation
But i have pointed out that some lineages of i2a present in Transylvania & Ukraine served as key mediators of EEF influence upon the steppe, and these are the very same lineages found in Bulgarian Yamnaya. They therefore represent a primary element in the genesis of the kurgan culture. If these same lineages are in turn found in western Anatolia, then i will be proven right.
as I have stated previously, EHG, WHG, CHG are useful heuristic labels for the Epipaleolithic & Mesolithic, but not really for the Eneolithic. More specific terms like Dnieper basin 8000 BP, or Volga basin 5000 BP would be appropriate
The key thing is to understand the phases
A) formation / coalesce of steppe groups; various HGs nested around river basins with clinal EEF influence (5500-4000 bc)
B) range expansion, & YDNA bottle necking
- cernavoda expansion 3700 bc
- usatavo 3300
- Yamnaya 3000
- CWC & offshoots 2800
- Fatyanovo 3500
Any attempted model should cover all these aspects & obviously have a conman day of linguistics; archaeology , Y phylogeny and Gw models. Good luck
@Copper Axe, I don't know we can get from models that fit low levels of CHG like ancestry in Ukraine_N (and this agrees with some qpAdm runs I have done, although I'm not sure it worked better than some Anatolian, would have to re-check) to Progress like populations at Dneiper without a big stretch in hypothesis. There may have been I guess, it's just speculative right now, or I'm forgetting something.
@Simon Stevin, possibly, yeah.
There are Eneolithic samples with lots of CHG coming from the steppe near the Black Sea in Ukraine and from around the Don.
Hopefully we'll see em and when they lie in dates within Eneolithic period soonish.
Off-topic: Mokrin paper released - https://www.nature.com/articles/s41598-021-89090-x#Sec25
(It notes that the cemetary is from Maros Culture, which I overlooked before, so perhaps SRB_Maros_EBA or SRB_Maros_EBA_Mokrin would be a good label for them, to compare with HUN Maros?)
@Simon Stevin “ Do you think there were WSHG populations that didn't have East Asian ancestry, unlike Tyumen and Botai? If so, would this hypothetical population resemble EHGs with a lot more ANE ancestry, like Steppe Maykop? Khvalynsk and Vonyuchka/Progress don't have literal WSHG ancestry, but they have ancestry from EHGs with a lot of ANE. These hypothetical EHGs could partly resemble what West Siberian Hunter Gatherer populations looked like before admixture with East Asian-derived foragers, which occurred sometime in the late Mesolithic and early Neolithic.”
People tend to forget that CHG is 2/3 Dzudzuana (Anatolian farmer- or Gravetian- like) and 1/3 ANE.
Iran_Neo is 50:50.
Hence, WSHG signal could be some of the one third ancestry in CHG deriving from an AG3- related population.
https://becc2021.com/
------------
Dr Serena Aneli
Abstract Title: Pre-Roman human occupation of Northern Apulia: a bridge between two worlds
The geographical location and shape of Apulia, a narrow land stretching out in the sea at the South of Italy, made this region an important bridge between Western Europe and the Aegean world. Such movements culminated at the beginning of the Iron Age with the Iapygian civilization which consisted of three tribes: Peucetians, Messapians and Daunians. Among them, the Daunians left peculiar cultural heritage, with one-of-a-kind steles and pottery, but, despite the extensive archaeological literature, their origin has been lost to time.
In order to shed light on this, we collected human remains from three archaeological sites (Ordona, Salapia and San Giovanni Rotondo) geographically located in Northern Apulia and corresponding to the area historically inhabited by Daunians. According to the cultural and archaeological context, these remains may be as ancient as the 9th century BC; radiocarbon dates are in progress. These samples were sequenced at the aDNA laboratory of the Institute of Genomics (Tartu, Estonia) yielding 21 new whole-genome sequences at an average coverage of 0.03X.
Preliminary explorations of their genetic variation together with published samples (projected principal components) show that our individuals fall within the genetic variability of their contemporaries from surrounding areas, with some interesting exceptions, possibly suggestive of a foreign influx. More work has still to be done to retrace the footprints of Daunians and ultimately, to reconstruct the tumultuous ancient dynamics shaping the genetic variability of people inhabiting those areas during the Pre-Roman period, as well as to explore the genetic continuity with modern times.
Virginia Veltre
Abstract Title: Quarto Cappello del Prete (1st-3rd cent. CE): genomic structure of a rural Roman community
Rome grew from a small City up to an Empire encompassing the whole Mediterranean area and beyond. Hence, it became a complex mosaic of people coming to the City from elsewhere. Indeed, Imperial Rome (1st -3rd centuries CE) could be referred to as the largest urban center in the preindustrial world. Indeed, Rome allured people into its walls to gain better life conditions. Thus, the identification of foreigners might be critical in the proper demographical estimations. The development of biomolecular techniques has remarkably deepened our knowledge about past populations by providing powerful tools to shed light on crucial aspects of individual biological characteristics. However, the genomic evidence for the population of one of the broadest Empire in antiquity has been very scanty until recently. In an attempt to contribute to elucidate the genomic structure of Imperial Rome inhabitants, we recruited 26 individuals from Quarto Cappello del Prete (QCP) necropolis. These samples were submitted to whole-genome sequencing to unveil the genetic component featuring the studied samples and the community's putative demographic structure. The preliminary results allow identifying heterogeneous landscape and providing reliable biological evidence about migration movements towards the Urbe. The genomic legacy with the south-eastern shores of the Mediterranean Sea and the central and western northern-African coast funerary influence pave the way to consider people buried in QCP resembling a Punic-derived human group. Continued efforts are mandatory, though the overall findings could provide a more comprehensive scenario of the events that shaped Imperial Rome inhabitants' genomic makeup.
Stefania Sasso
Abstract Title: Ancient DNA perspective on the origins and structure of a Merovingian population in Belgium
This project focuses on ancestry, kinship and genetic affinities of people buried in the dunes of the Early Medieval Merovingian burial site of Koksijde on the coast of Belgium. The Merovingian period (7th-9th ccAD), that followed the fall of the Roman Empire (476 AD) was the time of state formation in western Europe (Cook, 2002). The remains of 51 individuals were uncovered in archaeological excavations in 2017. Initial datings and skeletal examinations suggested that the burial ground was used by one or a limited number of family groups for a maximum of a century. Isotope analyses pointed to likely origins of the individuals from outside of Belgium, possibly from Scandinavia. To address the questions of genetic origins of the people of Koksijde and the extent of genetic relatedness at the site 36 best preserved remains of the 51 individuals were chosen for ancient DNA analyses using teeth, petrous bones and talus bones. For initial screening we produced low-coverage data that allowed us to run the first analyses of genetic ancestry and kinship. We investigated the endogenous DNA of each sample and analyzed mitochondrial DNA and Y chromosome haplogroups of the individuals, that showed levels of diversity that are comparable and not compatible with one or limited number of family groups. Autosomal data analyses cluster Koksijde individuals more closely with 15 individuals from a Late Medieval site of Sint-Truiden from central Belgium than with Scandinavians. Finally, the kinship analysis (Monroy Kuhn et al., 2018) done revealed 3 pairs of second-degree relatives.
@Davidski,
I'm looking forward to covering your post. I won't cover any of the schoolyard shittalking directed at Harvard lab though.
But can you email me the basic ideas you will have in your post in the next few days.
thepopulationgeneticschannel@gmail.com
I ask because I don't want to rush my video. I want to have time to make my video good and time to create maps based on the post.
"I won't cover any of the schoolyard shittalking directed at Harvard lab though."
I meant this as a joke btw. I actually would cover it is hilarious.
@Genos Historia
I think what you're referring to are my comments about a recent phenomenon in which ancient DNA is used to argue things that aren't necessarily true, or even things that are completely false, but are seen as true by many academics because they mistake the interpretation of ancient DNA by certain other academics as hard science and hard evidence supporting these supposed truths.
This phenomenon results not only in a lot of ancient DNA papers with the wrong outcomes, but also persuades archeologists and linguists to find explanations for things that never actually happened, and were never considered as likely before ancient DNA.
Now, many people might see this as a natural part of the scientific process, and a way to eventually reach a sensible consensus, but I think that this is a somewhat new phenomenon that has the potential to turn the scientific process upside down, to the point that it's pushed into pure fantasy.
I think that we've already seen the beginnings of this with the search for the Indo-European homeland. For instance, see here...
https://eurogenes.blogspot.com/2016/10/dead-cat-bounce.html
And if you think that this sort of nonsense will correct itself overnight, then think again.
I reckon we'll be reading about Yamnaya coming from Mesopotamia for the next 20 years or more.
@ Davidski
persuades archeologists and linguists to find explanations for things that never actually happened
It works both ways. Geneticists often look for something, that is a product of imagination of linguists (early Anatolian split, early Tocharian split) or archaeologists (Slavic homeland). Or they pretend that they don't see certain things, because they're deemed to be "impossible" (Tollense, MX265 and others).
@ Arza
Well; Anatolian did split earlier; but perhaps not as early as if we take the tree at face value (populations splits to not mirror family trees)
Do you not think is a Slavic homeland ? Please don’t tell me you have acquired the nihilistic position of “there were no slavs; everybody was slavs”.
@ Dividski “ This phenomenon results not only in a lot of ancient DNA papers with the wrong outcomes, but also persuades archeologists and linguists to find explanations for things that never actually happened, and were never considered as likely before ancient DNA.
Now, many people might see this as a natural part of the scientific process, and a way to eventually reach a sensible consensus, but I think that this is a somewhat new phenomenon that has the potential to turn the scientific process upside down, to the point that it's pushed into pure fantasy.”
I think this is better than the position of some linguists (Jaska) that genes do not show the language and the position of some archaeologists that genes do not show material culture.
The Slavic homeland is only an ethnolinguistic concept. It is a geographic area where Slavic linguistic innovations were born. And it is not the area of Polesia, as some archaeologists try to prove, and which linguists definitely deny (the Baltic toponymy of Polesia). It has long been denied by genetics (the Baltic genetic substrate of Polesia), and today denies it archeogenomics - the vast majority of early Slavic genomes are most similar to contemporary Poles and Western Slavs, the vast majority of Poles and Western Slavs cannot be a mixture of genetic Belarusians and Restgermans (Kowalewko males).
@Vladimir
You're actually giving an example of a very similar problem, where instead of a true multidisciplinary approach, disproportionate weight is given to one method and little or no room for constraints from other methods.
So what we need to see is neither one or the other approach, but a more mature multidisciplinary approach, where ancient DNA interpretation is seen as potentially fallible, and data from all of the disciplines are used as constraints.
@Arza
Those Merovingians from Koksijde being like samples from St. Truiden are interesting. It is fairly sure that the part of Flanders quite close tot Koksijde was settled by Frisians. This is derived from toponyms such as Diksmuide and pottery finds. St. Truiden o.t.o.h. was surely settled by Franks. So it would be interesting to see when these samples get available whether the used term "cluster more closely than with Scnadinavians" would still allow for Frisians.
@Arza
Mind you, the settling could have been - and likely was - far more complex than Frisians here and Franks there. But the area where Frisian are proposed were tidal areas and that kind of environment was not unlike Frisia itself.
@All
At vAsiSTha's request, the 3D tool has been updated.
https://vahaduo.github.io/3d/
To add a line click HIGHLIGH button first. It'll switch to LINE. Next button cycles trough ADD, DEL and OFF. When set to DEL clicking a point deletes all lines that go through it. When set to ADD clicking points adds line segments.
After the first point is clicked LINE changes to CANCEL, so that the current line can be easily deleted.
After the second point is clicked ADD changes to OK. Clicking OK ends the line drawing, so another can be added.
Third button cycles through colors (same as for highlight, but independently selected). CLEAR clears all lines.
Bug: after adding the second line, you can no longer click the labels and to remove them you must click the point to which they are attached.
Other changes:
The tool now takes both, scaled (default) and raw coordinates. You can switch between them in the header of the modal window.
Styled scrollbars.
I also renamed the coordinates from "Non-scaled" to "Raw" in g25views.
Let me know if you see any bugs.
@ambron
The Slavic homeland is only an ethnolinguistic concept. It is a geographic area where Slavic linguistic innovations were born. And it is not the area of Polesia, as some archaeologists try to prove, and which linguists definitely deny (the Baltic toponymy of Polesia). It has long been denied by genetics (the Baltic genetic substrate of Polesia), and today denies it archeogenomics - the vast majority of early Slavic genomes are most similar to contemporary Poles and Western Slavs, the vast majority of Poles and Western Slavs cannot be a mixture of genetic Belarusians and Restgermans (Kowalewko males).
Cannot Proto-Slavs themselves be a mixture of a Baltic-like population and early Germanics?
@Onur Dincer
“Cannot Proto-Slavs themselves be a mixture of a Baltic-like population and early Germanics?”
This is impossible:
https://i.postimg.cc/bYxfWzPZ/screenshot-81.png
@Old Europe “
"The recent greek paper contains many strange thing but at least it is accurate in saying that pontic-caspian steppe groups that reached the greek peninsula were mostly of euro-hunter stock
IIRC they have EBA steppe modeled as 65/70 EHG and the rest CHG like."
What is the implication? Furthermore, the BA Aegeans had already significant CHG admixture would this CHG not lower the embedded EHG and increase the embedded GHG in Log4? Besides is Log4 really close to Scots/ (Lithuanians) as it appears from f3?
@vahaduo
Thank you
Onur, I understand you're asking in a genetic context... Theoretically, it would be possible, but then the pre-Slavs would look genetically like today's Kashubians. Some of the pre-Slavs could actually look like Kashubians, because that was the genetic makeup of the Legedzine population, which came to Ukraine originally from northern Poland.
@EastPole
Your graph is little different from Turbo-Slav shits. Proto-Germans obviously existed before Proto-Slavs.
@ambron
My question was more about possibility than probability, thanks for your answer. I do not know modern Kashubian genetics BTW, what can you tell me about it?
@Onur Dincer:
Are you surprised? I've come to expect stuff like that by now.
Be thankful that no Pole started "we wuzzing" as a Suebi here... yet.
Speaking of "early Slavs were like modern Western Slavs"... which ones? I'm currently quite fond of the Krakauer Berg samples and they look rather East Slavic.
That brings me to the topic of how the Poles acquired their Germanic ("Oh nooooooooo!") admixture.
I remember reading some comments here how it is actually Danish rather than German. While Danish admixture certainly makes sense (you know, with all the Vikings who have been running around here), can we be sure about that?
I mean Germans as a whole might be easy to differentiate, but what about Northern Germans compared to Danes? Many of the Germans moving east would have been Saxons and I would expect them to be a bit harder to distinguish from Danes.
@ Rob
There is no proof that Anatolian spit earlier, because reconstructions are not data.
prof. Gąsiorowski:
Thus, protolanguage reconstructions are not “data”. They are forever provisional and hypothetical. Using them as data is a category error.
It's reconstructed as such, but reconstructions can differ. And they can change dramatically.
My prediction about the Anatolians was always that they were barely IE from a POV of genetics. That they were just locals from Anatolia or the Balkans, and that one day they'll be used to undermine the steppe theory, because "everyone knows" that they are nearly synonymous with Proto-Indo-Europeans.
Archi's prediction, if I understand correctly, is that they were Pre-Greeks. So if Archi is right, they should be quite local (as I predicted), but similar to non-IE layer (Minoan, Cycladic, whatever) in early Greeks (hence their linguistic divergence, similar case as with the divergence of Tocharian due to the contacts with Uralics, as per the newest research on this topic), and not to the pre-IE inhabitants of Anatolia.
Do you agree?
And what's your prediction when it comes to their autosomal composition? Because Y-DNA, especially the one present also in non-IE contexts, is not everything. Yamnaya? Neolithics from Bulgaria? Can you mark a spot on the West Eurasian PCA and post it?
Re: Slavic homeland
I told you everything ~4 years ago. Balto-Slavic not-a-cline. It's still there and it won't go anywhere. Just check what's on the other side, and you'll know how far south you need to go. Helves on AG just passed this IQ test.
The only thing that changed, is that back then I thought that Baltic_BA originated in the north-east and moved into inner Carpathians in the Bronze Age. Now I know that it was the opposite. Volosovo is nearly identical to LVA_MN, so no surprises in the area between them are to be expected. Pre-Fatyanovo rolled over the supposed Slavic homeland. And what? Nothing. New samples from the Balkans/Carpathians on the other hand are packed with the "Balto-Slavic drift"...
@ Onur Dincer
Cannot Proto-Slavs themselves be a mixture of a Baltic-like population and early Germanics?
They can't. Basic PCA-reading skills are enough to know why.
Your graph is little different from Turbo-Slav shits. Proto-Germans obviously existed before Proto-Slavs.
You have absolutely no idea what you've commented on. You just made a complete fool of yourself and proved how biased you are. You don't even know which theory this diagram illustrates, whose theory it is, why it looks like that, in whose book it was published, and why it's a proof that Slavs are not half-Balts, half-Germanics. But you saw a giant circle with "Slavic" written in the middle and it immediately triggered you, because if it's "Slavic" then it should be a tiny, unimportant circle somewhere in the corner.
@Davidski
What happened to the last post about the Greeks? Cant find it anymore.
@Copper Axe
Apparently it was deleted because it posed a malware/phishing threat. That's what Blogger emailed me.
I'm guessing the threat was of the Trojan kind...haha.
But seriously, not sure what's going on there, so let's move on from that topic for now.
I'll hit it hard again when new samples come out from Classical/Medieval Greece.
The Truth always wins out in the end :)
Re: Greece
To me this suggests that most present-day Greeks harbor significant levels of Slavic ancestry and some sort of recent Cypriot-related ancestry
PCA based on the D-stats from the post
That's the reason why I'm using Cypriots as a stand-in population for pre-Slavic Greece "since always". They are situated perfectly at the end of the cline. A massive genetic shift must have happened between the Mycenaean period and the Middle Ages.
@Arza
You have absolutely no idea what you've commented on. You just made a complete fool of yourself and proved how biased you are. You don't even know which theory this diagram illustrates, whose theory it is, why it looks like that, in whose book it was published, and why it's a proof that Slavs are not half-Balts, half-Germanics. But you saw a giant circle with "Slavic" written in the middle and it immediately triggered you, because if it's "Slavic" then it should be a tiny, unimportant circle somewhere in the corner.
Have I said anywhere that Proto-Slavs can or cannot be a mixture of a Baltic-like pop and early Germanics? Have I criticized anyone for making either of those statements? I have only asked a simple question: "can Proto-Slavs be a mixture of early Germanics and and a Baltic-like pop?" and have not given an answer myself. My criticism to the graph EastPole posted was because it did not make any sense to me linguistically, not because it was used to prove that Proto-Slavs cannot be a mixture of a Baltic-like pop and early Germanics. Additionally, my statement that Proto-Germans predate Proto-Slavs is the consensus view. Finally, no sane person here would talk positively about Turbo-Slav crackpots, I could not care less if my negative remarks about them would offend someone. In no way these show that I am biased against Slavs, I am only biased against stupidity. FYI, I am someone who is much more descended from Slavs than Germanics (it is even disputed if I have any Germanic ancestry at all). It is your touchy reaction to me that gives the feel of bias. Real men do not act like that.
Onur, Kashubians are those Slavs who are at the top of the Polish cluster at the NE PCA. Of all Western Slavs, only they can be a mixture of Balts and early Germans, because they lie on the PCA between Lithuanians and the males of Kowalewko. But their genetics have nothing to do with the Middle Ages Slavic migration; it was already like that in the Iron Age, according to a study by Margaryan.
The graphic presents the Schmidt / Lehmann wave theory, which - unlike the language tree - adequately reflects the process of Indo-European linguistic differentiation. (Schmidt is a German linguist, Slavist). And the Slavic dialects are of course much more archaic than the Germanic dialects.
@ Onur Dincer
The consensus is that the proto-languages in the absence of written records cannot be reliably dated.
For no reason you've called the work of two Indo-Europeanists (Lehmann & Schmidt) and Colin Renfrew "little different from Turbo-Slav shits".
And now you continue the trolling with "Real men do not act like that."
EOT
@ambron
Onur, Kashubians are those Slavs who are at the top of the Polish cluster at the NE PCA. Of all Western Slavs, only they can be a mixture of Balts and early Germans, because they lie on the PCA between Lithuanians and the males of Kowalewko. But their genetics have nothing to do with the Middle Ages Slavic migration; it was already like that in the Iron Age, according to a study by Margaryan.
Thanks for the explanation.
@ Rob
I wrote that they'll be barely distinguishable from the locals (similar enough to question the steppe theory under the assumption that they represent some primordial PIE population). That's as far from the "Slavic R1a" as possible.
And don't worry, I know that they'll be "I2a-rich".
Genetic continuum from Latvia to Volosovo eliminates area east of the Baltics as a source of the WHG admixture in Baltic_BA. That's all.
“ New samples from the Balkans/Carpathians on the other hand are packed with the "Balto-Slavic drift"...”
And we also know they represent demographic & cultural dead ends; not ancestral to the protoSlavic expansion
@ambron
And the Slavic dialects are of course much more archaic than the Germanic dialects.
Slavic itself might be more archaic than Germanic, but that does not mean Proto-Slavic was spoken earlier than Proto-Germanic, just means that pre-Germanic diverged from Proto-IE more than pre-Slavic.
@Arza
The consensus is that the proto-languages in the absence of written records cannot be reliably dated.
All the IE trees I have seen so far that show the times of branchings converge all the Germanic branches at an older time than all the Slavic branches. But this can be inferred even from the levels of intelligibility among the modern Germanic languages and among the modern Slavic languages.
@Arza, re; whether Latvia_BA (and Baltic_BA generally) would have to have EHG admixture if their extra HG admixture is "home-grown".... (I saw your blog post on this, was thinking about replying there but will reply here instead).
It would be very reasonable to suggest this, *but* I would comment that the Latvia_MN set seems more diverse than the EHG like Comb Ceramic. The first "Latvia MN" paper published was two samples, one of whom is identical to Latvia HG and the other is a basically EHG CCC (Comb Ceramic Culture). Both around the same time.
It's not unreasonable to expect they admixed, but also far from unreasonable to think that these survived as separate cultures...
The sample set has built up since then with a paper by Mathieson mainly, and we have a few samples who are WHG and EHG like in basically the same sort of time period, within a 100-50 years of each other: https://imgur.com/a/BAyN5gB
(As an aside, I'm not sure why Reich Lab label the more WHG like samples as Latvia_MN and the Comb Ceramic samples as Latvia_MNo3. I could guess that this is because Latvia_MN represents a larger set of samples they have which are unpublished and the Comb Ceramic are actually outliers, but this would be pretty speculative.).
With admixtools2 I'm finding its very easy to get passing models for Latvia_BA with the two earliest CWC samples in Baltic (I4629 and Plinkaigalis242, who I label Baltic_CordedWare_Early), Globular Amphora and various HG populations. These include the two-samples Reich Lab labels as Latvia_MN - I4437 and I4627. These two-samples are pretty similar to the Central European Mesolithic guys like IronGates_Meso.
Here are some examples: https://imgur.com/a/8xVCi8o
So long as the EEF, Steppe and HG proportions can move freely, things work OK.
Despite SNPs getting down to only 265k with 240k polymorphisms for each population due to the amount of samples I included in this run, I do think they can preserve signal pretty well and are rejecting the HG populations they should. They also seem to replicate fairly well with the allSNPs option on the 1240k data (having finally figured out that it's an available parameter in ADMIXTOOLS2, if you run directly on genotypes).
A two way model of early CWC and Narva (Latvia_LN+Narva) that you mention on your blogpost on the topic as in Mittnick's paper may fail due to a lack of enough EEF input; can't get the ratios quite right.
Again I'm not saying that your Central European newcomer model (if I'm understanding it rightly?), for them is wrong and precluded at all (after all "Uncanny proportions" apparently persist in Central Europe), just noting that the idea of lack of persistence of these more WHG like populations in the Baltic region is not so clear and models with some of the chronologically latest Latvian HG do seem to work well to me. From the admixtools2 perspective, Latvia_BA seems to me to work with either SE-Central HG and also some of the latest Baltic HG to be about. So I don't think the argument of "Presence of Comb Ceramic in late Latvia_MN precludes local admixture" is quite a slam-dunk.
Further samples and bigger sample sets in the intervening periods will clear this up.
@Davidski
"But seriously, not sure what's going on there"
I got that message too, very strange indeed. Maybe someone flagged your website? Greek conspiracies...
Btw why did you get suspended this time on AG? The silly debate about how common sense, sanity checks and healthy skepticism applies to all language families except Uralic?
Too little factuality and too many wishes and desires in that debate anyways, and that doesnt just go for the people who argue for a European homeland imo.
@Arza
"hence their linguistic divergence, similar case as with the divergence of Tocharian due to the contacts with Uralics, as per the newest research on this topic"
If you're talking about that article regarding the deviant typological profile of Tocharian, that article in no way refuted the early split of Tocharian because they place the contacts with Pre-Tocharian and Pre-Samoyedic/Proto-Uralic to the late 4th and early 3rd millenium b.c and attribute it to the Afanasievo culture, which arrived in the Altay region before the estimates of LPIE breakup.
Rob is right, Anatolians did split earlier. As did Tocharians, although not nearly as early. Linguistics shows it, archaeology shows it, and so will aDNA when we finally get genomes from Anatolian and Tocharian speakers.
@matt
How would be a model for early baltic CWC using three sources?
Progress
GAC
Ukraine Mesolithic
Early Polish CWC is even more Yamnaya-like than early Baltic CWC.
@ Arza
Substratum effects operated everywhere; nothing unique to Anatolia . You’d have it from Ireland to India; because IE weren’t moving into vacuums
So we have a tangible basis for the distinction of Anatolian languages- slightly earlier divergence;& not the exclusively R1b-M269 & R1a-Z645 lineages
@ Copper Axe
The fact that they stick to the old belief that Afanasievo = Tocharians is irrelevant. Interactions between IE and Uralics will be surely detailed in upcoming studies and it'll be rather late 3rd - 2nd millenium BCE, not 3rd-4th. Since there is a Z280 in Srubnaya, maybe even some non-Z93 clades will be found there, as in Tarim mummies (I know that's just a PCR, but on the other hand the chance of contamination with "non-Asian" R1a in a Chinese lab is rather low).
As for the question whether Afanasievo could have been proto-Tocharian see this discussion:
https://languagelog.ldc.upenn.edu/nll/?p=47645
TL;DR Tocharian, besides some quirks, is similar to European languages. If it would be derived from Afanasievo, it should exhibit much greater level of divergence.
And they assume that Late PIE = Yamnaya. With the Corded Ware the expected difference would be even greater.
In other words we don't have to even bother with aDNA in this case, as Afanasievo can be rejected on a purely linguistic basis.
@Old Europe, the two samples I'm calling Baltic_CordedWare_Early seem happy enough (happier if anything) in qpAdm with 3-way models with Ukraine(Neo/Meso)+Steppe_Eneolithic+GAC_Poland: https://imgur.com/a/bIQV1n6
It seems that the outgroups I've got are for some reason not too sensitive to the choice of how the model gets to the right EHG:IronGatesHG like ratio, even with the Ukraine samples in the pRight, and lots of models that can get there work OK. (It might be that this really is the case, or maybe my strategy of using *relatively* recent outgroups is having some problem here.)
I'm using the Russia_Steppe_Eneolithic as is from the 1240k IND file, in these models so that's the 2 Progress and 1 Vonyuchka.
poz81 (Polish Early CW) not included in the above B_CW_Early because it is shotgun (as well as not being from Baltic) and didn't want to add that as a complicating factor as think the other two aren't.
(On another topic, another quick attempt at amusing myself by fitting a bunch of populations into 3-way model, like been done many times before: https://imgur.com/a/9mI44yd).
@matt
thank you
I think going by these numbers it would be parsimonious to conclude PIE is a Progress kind of language.
Even tough Progress so far has not given the right Y line IIRC.
I'm confused.
No, not Progress.
Take a look at this PCA. It shows where the Sredny Stog population really clusters.
https://1.bp.blogspot.com/-RJFDsjkdGsM/YKAfi08v0dI/AAAAAAAAJ_A/l_vh55vv5VsxbPRtpMJ7z5AnvspCL36rACLcBGAsYHQ/s1218/PCA_of_ancient_West_Eurasia.png
Ok I see that is right but then I would like to know how to model
Sredni Stog with
EEF
PROGRESS
UKRAINE NEO/MESO
got it
sample": "Ukraine_Eneolithic:I4110",
"fit": 3.4225,
"Globular_Amphora_Ukraine": 42.5,
"Progress_Eneolithic": 30,
"SHG": 19.17,
"EHG": 7.5,
"WHG": 0.83,
"LBK_N": 0,
This is only one sample but it seem predictable that most will be
30 EEF
35/40 PROGRESS
30/35 UKRAINE HG (WHG+EHG)
we go back to PIE from Ukraine HG (DNEPER/ DON FORAGERS)
Ukraine_Eneolithic I4110 is nothing like Sredny Stog.
Sredny Stog resembles the Yamnaya samples with elevated WHG.
Other ukraine eneolithic samples from western sredni stog area
Ukraine_Eneolithic:I5882
Ukraine_N 66.85
Trypillia 23.2
CHG 8.6
Yamnaya_Samara 1.35
EHG 0
[1] distance%=2.8944 / distance=0.028944
Ukraine_Eneolithic:I5884 (R1b-Z2103)
Ukraine_N 71.4
Trypillia 28.6
CHG 0
EHG 0
Yamnaya_Samara 0
they are overpacked with local hunters as everyone can see.
Someone is cooking the book in genetic laboratories about Yamnaya dna ...
@Old Europe “ we go back to PIE from Ukraine HG (DNEPER/ DON FORAGERS)”
Do you mean PIE came from Dnieper Donets?
@Genos Historia @old europe What’s the source of EEF/Anatolian Farmer in Sredny Stog and Yamnaya? One model gives precedence to GAC whereas another one prioritizes Tripolye instead.
@Davidski “ Ukraine_Eneolithic I4110 is nothing like Sredny Stog.
Sredny Stog resembles the Yamnaya samples with elevated WHG.”
I’m starting to suspect that Khvalynsk and Progress did not speak anything close to PIE.
PIE is likely a Dniester creation rather than a Volga-Caspian one.
PS: Modern day Ukrainians and Balts seem to have an equal ratio of WHG : ANF (both 20% - 25%).
Sredny Stog is basically like Yamnaya, but since there are many more Yamnaya samples, then they're slightly more varied.
I know G25 won't show it, but Meshoko might have some Progress-related ancestry (which makes historical sense IMO)
left pops:
Russia_Caucasus_Eneolithic
Russia_Steppe_Eneolithic
Georgia_Kotias.SG
Turkey_TepecikCiftlik_N.SG
best coefficients: 0.281 0.517 0.202
Full
@ Andrze
''PIE is likely a Dniester creation rather than a Volga-Caspian one.''
IMO it likely expanded from that region after 4000 bc. But organically/ geologically, it is not originally from there. The bulk of its ancestry is a lot more eastern.
But you see, these are all different 'dimensions' of its genesis
Yes, in my analyses the Ukraine Eneolithic samples come up nothing like Yamnaya or Sredny Stog either. So can we say that Sredny Stog has more eastern origins, like e.g., the Volga-Caspian steppe or the Caucasus steppe, and largely replaced the Dnieper–Donets people in the Ukrainian steppe and then Yamnaya emerged from Sredny Stog in the Ukrainian steppe and replaced the populations in the Volga-Caspian and Caucasus steppes?
@Davidski
Is there EEF in Sredny Stog like how there is in Yamnaya?
@CrM
Well, there must be, because they're basically identical.
@Davidski
Could it come with the forager ancestry?
If the models are correct we can conclude that the ukraine eneolithic population was wiped out completely in a way that matches for example what happened in Britain and some other wetern european places in the final neolithic.
But that seems impossible.
IMHO there is also another explanation. genetic labratories are just lumping together all the EHG dna together with CHG to create an artifact called steppe ancestry
WHAT IF
You just have to take all the EHG in Yamnaya and lump it with WHG rich Ukraine hunter and than Yamnaya would be bring back to some kind of a more population similar to Ukraine eneolithic.
The greek papare is an indication that this could be the case.
I will e-mail the authors about that
@ Onur
''Yes, in my analyses the Ukraine Eneolithic samples come up nothing like Yamnaya or Sredny Stog either. So can we say that Sredny Stog has more eastern origins, like e.g., the Volga-Caspian steppe or the Caucasus steppe, and largely replaced the Dnieper–Donets people in the Ukrainian steppe and then Yamnaya emerged from Sredny Stog in the Ukrainian steppe and replaced the populations in the Volga-Caspian and Caucasus steppes?''
Firstly, it remains to be seen that the samples being referred to in the kindly 'leaked' talks are actually Sredni Stog, in the sense that S/S should refer to the 4800 - 4000 bc phase, and those after 4000 bc are western 'post-Stog' groups like Cernavoda, Mikhailovka, etc.
At present, we arent analysing apples with apples, as we are using 5500 BC samples form DDII to model Yamnaya in 3000 bc. The shifts will indeed be huge, but possibly a red herring.
True Sredni Stog groups might have retained considerably more Ukraine_N ancestry as well as absorbing EEF. It probably arose from the Skelja population in the Dnieper-Donetz region, and we have to remember that DDII is the intruder here on a population which was originally more EHG-rich. An example of the Skelja group, but at the extreme range of diversity, is the Smyadovo R1b sample. Even though it is of low quality at present, I don't think it'll change much. When when you take populations like this, and mix them with Progress-Khvalynsk groups, then you get Yamnaya, and this will be the new model for Yamnaya genesis broadly 50/50%, instead of the currently awkward model of Progresss 85% + GAC 15% (which fails the sanity check). Then it'll all align with Y-DNA, archaeology, paleozoology, etc
@Rob
"Yamnaya, and this will be the new model for Yamnaya genesis broadly 50/50%, instead of the currently awkward model of Progresss 85% + GAC 15% (which fails the sanity check). Then it'll all align with Y-DNA, archaeology, paleozoology, etc"
I agree with this take here, I'm expecting something similar. Obviously it doesnt work all the time but we are quite often seeing that populations were 50/50 mixes of two relevant populations rather than 30% X, 35% Y 35% Z.
@Rob, potential problem with that Caucasus_Eneolithic run.
I see in the full run you've got Russia_Caucasus_Eneolithic with 2 samples in. In the HO Origins and 1240k file, these two samples are I2056 (who is our Darkveti-Meshoko sample), and then SA6010... the artist formerly known as one of "Yamnaya Caucasus", who appears to be another Progress_En like person and also a relative of a Steppe_Maykop person at the same site, apparently, according to the .anno. They've redated him based on his relationship to the Steppe Maykop woman.
So if you're running a qpAdm with one Meshokan and one Progress_En, you're gonna get a model with "Progress_En" ancestry, but this isn't really Progress ancestry in both samples, just a lot in one of them.
I don't know why Reich Lab have chosen to do this; it doesn't seem to make any sense. I've relabelled SA6010 under the slightly awkward "Russia_Caucasus_Eneolithic_Steppe" in my IND file, in case I ever want to run the Meshokan separately. Though he could be aggregated in with Russia_Steppe_Eneolithic if we so chose.
Sadly, out of 3 DVM samples, all of them are a set of 3 siblings; I2056, I2055, I1722 - so they've separated them out into 3 labels, keeping only I2056 in the "real" label. Looks like they got some good stuff out of a family grave, but then nothing else. I suppose we could merge all these for qpAdm purposes, as long as we're doing that, but it would create disasters for estimating Fst. They've done this because I2056 has by some distance the best coverage out of the two brothers and one sister.
Some Yamnaya models with and without Khvalynsk:
https://drive.google.com/file/d/1etLxO9BHLMlzUcYTJ-NMXMbcKS2hu9Xo/view?usp=sharing
Having a go at systematically comparing some Global 25 runs to a qpAdm run using similar outgroups as I've generally run*.
Here: https://imgur.com/a/5hMVmos
Generally a lot of similarities between them, but finding that G25 gives quite a few populations more HG ancestry than the qpAdm suggests they should have, while more EEF and to some extent Steppe ancestry in G25. Because there's a lower estimated HG proportion, it makes me more sympathetic to the idea that it is difficult to model populations like Belarusian and Ukrainian on a cline from Latvia_BA, whatever the G25 fit seems to be... Still not sure how to reconcile these differences and what different sources of biases in each models could make them go awry. The models in qpAdm do seem to be giving systematically less HG to modern Europeans relative to G25, while the ancients are the same in either case. Not sure why this could be.
*This run doesn't provide acceptable p-values, but I think this is because I've used a lot of HG outgroups with complex relationships, and the best-fitting proportions still tell me something meaningful.
@ Matt
Right you are, it is mislabelled.
In that case, Meshoko is a mix of CHF and some southern ancestry. Most optimal
Russia_Caucasus_Eneolithic
Azerbaijan_Caucasus_lowlands_LN
Georgia_Kotias.SG
0.481 0.519
0.875683
What's even more interesting is that Majkop is quite distinctive to Meshoko, whilst in Wang they essentially suggest that they were continuous (an odd conclusion given the archaeological / cultural differences between the 2)
@Rob “ In that case, Meshoko is a mix of CHF and some southern ancestry. Most optimal”
Do you mean CHG?
Andrze - yes , CHG
@ Matt Re: last comment
f3(Mbuti; row, column) maxmiss=1
,Brazil_LapaDoSanto_9600BP,Cameroon_SMA.SG,Georgia_Kotias.SG,Germany_Mesolithic,Iran_GanjDareh_N,Israel_C,Mongolia_North_N,Morocco_Iberomaurusian,Russia_HG_Karelia,Russia_Sunghir.SG,Turkey_N,Ukraine_Mesolithic,Ukraine_N
Russia_Steppe_Eneolithic,0.0684,0.0163,0.0699,0.0728,0.0687,0.0674,0.0629,0.0507,0.0747,0.0634,0.0669,0.0754,0.0746
Russia_Caucasus_Maikop,0.0661,0.0168,0.0687,0.0676,0.07,0.0727,0.0619,0.0519,0.0682,0.0616,0.0697,0.0723,0.0726
Serbia_IronGates_Mesolithic,0.0641,0.0159,0.065,0.0882,0.0628,0.0665,0.0607,0.0517,0.0765,0.0674,0.0685,0.0803,0.082
Serbia_EN,0.0627,0.016,0.0664,0.0716,0.0653,0.073,0.0596,0.0531,0.0673,0.0633,0.0753,0.0701,0.0702
Belarusian,0.0615,0.0155,0.0663,0.0712,0.0628,0.066,0.059,0.0498,0.07,0.0632,0.0684,0.0701,0.0707
Ukrainian,0.0614,0.0157,0.0664,0.0712,0.063,0.0663,0.059,0.0499,0.0699,0.0633,0.0686,0.07,0.0705
It seems that in f-stats there is a very high level of drift shared between the Iron Gates and DEU_Meso, UKR_N and UKR_Meso (shared recent ancestry? ROHs?). This probably sets apart IG from other left pops and the target.
@arza, sure though f3 stats generally seem really high between these "WHG" type populations. I.e. it seems expected. Do you reckon it's too high and there's some problem that influences the 3-way model compared with Global25? If I get time later today I will have a look and see if there are any viable good sets of populations I can swap in or out for each of them without screwing up the numbers of SNPs (should be OK if I can find another WHG like population with at least 2 or 3 good coverage individuals).
@arza, matt, others,
"f3(Mbuti; row, column) maxmiss=1"
Simple measures of genetic distance like Fst, f2, f3 are very useful for population analysis. However, in f3 calc there is a need for an outgroup, in this case, Mbuti. The problem is some populations will be closer to Mbuti than others and hence these pops will have have lower f3 (f3(Mbuti; pop, x)). This then will not be a good measure of distance/shared drift b/w pop and x.
Therefore, it could be better to use F2 or maybe fst. I dont have access to either of these so I still use f3 but f2 I think is available in Admixtools2 and fst I dont really know if its as useful. Something to think about.
@ Matt
I wouldn't call the stats "too high". They're exactly as they should be. But at the same time they show a different picture than the one we see in a PCA (not only in the G25) and we need to find the reason why it happens.
This strong connection between IG and WHG-admixed right pops certainly influences the outcome of modelling in qpAdm. Adding more IG would make the targets more similar to DEU_Meso, UKR_N or UKR_Meso than they really are, so IG is avoided.
@ mzp1
Fair point, but if the distance to Mbuti (e.g. due to the basal Eurasian admixture) is the reason behind this, I'd expect "normal" stats between IG and WHG-rich pops, and deflated ones between IG and basal-admixed. But as you can see f3 with other pops look "normal", and with WHG-rich are elevated well above the average for this set.
Arza, yh i just meant in general about f3 calc, not your calcs, i dont think its an issue for you in that calc, but it can be when looking at more diverse pops globally.
@All
As far I can see, this thread unfortunately somehow evolved into one big off-topic mess, turning into some Greek-Slavic complexes and phobias.
Coming back to the topic of this post I have a question about the arrival of the N haplogroup to the Baltic Sea area:
Can this be related to the appearance of Seima-Turbino on Altai and its route through the Ural to the Baltic Sea?
What's your opinion on the origin of the Pannonian Avars? They had 80% kra001-related ancestry, and they built a long lasting empire in the Carpathian Basin, where an Uralic language is now spoken.
Are the Avars really mostly Kra001, or are they mostly Baikal/Mongol BA?
Post a Comment