Thursday, November 29, 2018

The Uralic cline in the Global25


The Uralic cline is a concept that was discussed in some detail in the recent Lamnidis et al. palaeogenomics paper on the origin and spread of Siberian ancestry in Europe (see here). It pertains to the most northerly genetic cline that links the populations of West and East Eurasia, and is largely made up of Uralic-speaking peoples rich in Y-haplogroup N1c.

This is what the Uralic cline looks like when inferred from my Global25 Principal Component Analysis (PCA) data. Note that the plot features most of the Lamnidis et al. ancient samples, and they're all more or less sitting along my version of the said cline. The relevant datasheet is available here.

Admittedly though, as pointed out by Lamnidis et al., the Bolshoy samples probably aren't those of Uralic speakers because they're dated to 1523±87 calBCE, which predates most linguistic estimates of the spread of known Uralic languages into the Kola Peninsula. So the important question is why do they cluster along the Uralic cline and 2/2 of the male samples belong to N1c?

The most logical explanation, I'd say, is that the Uralic cline actually represents an older, pre-Uralic contact zone between the east and west. Nevertheless, I think it's likely that the Proto-Uralic language formed somewhere in this ancient contact zone, and the early Uralic-speaking peoples used it to their advantage to spread rapidly both east and west, especially during the Late Bronze/Early Iron Age, when they, and their N1c, finally reached the East Baltic region (see here).

See also...

Corded Ware people =/= Proto-Uralics (Tambets et al. 2018)

How did Y-haplogroup N1c get to Bolshoy Oleni Ostrov?

The mystery of the Sintashta people

122 comments:

  1. These pops are now in all of the Global25 datasheets...

    Bolshoy_Oleni_Ostrov
    Chalmny-Varre
    Levanluhta_IA
    Levanluhta_IA_o

    The links are always the same, see here...

    https://eurogenes.blogspot.com/2018/10/cimmerians-scythians-and-sarmatians.html

    ReplyDelete
  2. Bolshoy Oleni Ostrov people were not Mesolithic hunters but belonged to an advanced Bronze Age culture and had contacts with the south:
    - Bolshoy people made advanced sledges/boats with tarring
    - Bolshoy people were aware of metal making techniques
    - Bolshoy people were aware of the creamation techniques as cremations were found in Bolshoy
    - in the QpAdm, Bolshoy people are c. 22% Yamnaya
    - one of the mtDNA haplotypes is rare T2d1b1 which is today found in Khanties, Komis, Persians and Haryana Brahmins.

    More information on the burial site is found in this paper: https://www.academia.edu/31826165/Kola_Oleneostrovskiy_Grave_Field_A_Unique_Burial_S ite_in_the_European_Arctic

    These are of particular interest:
    Most of the bodies had been buried in wooden, boat-shaped, lidded caskets, which looked like small boats or traditional Sámi sledges (Russian kerezhka). It seems that the boards of the boats were made of thin wooden planks and were probably tarred.

    Intensive caulking of joints between the boards and lids and meticulous sealing of hulls and knotholes on the bottom of the boat all suggest that funeral boats (or sledges) were used for transportation over the water. Such caulking is necessary only for sailing, and on the other hand, moving boats on dry land would damage the tarring. Therefore it is likely that the ‘coffins’ were first moved on dry land to the shore, after which they were carefully sealed with tar, and then rafted in tow across the bay to the burial place on the island.

    A few fragments of a crucible with drops of bronze on the inner surface were recovered from the destroyed part of the gravefield. A small bronze dagger was found in burial VIII. A bronze blade was found in burial 15. Additional evidence suggesting contacts between the Kola Peninsula and Southern Scandinavia is provided by the chemical composition of a bronze dagger, a small metal blade and droplet on the inner surface of a crucible.

    Three cremations have been recorded during excavations of the grave field. Particular attention should be paid to the burial investigated by Shmidt (burial X), as everything in it contradicts the typical traits of the local burial rite. The burial consists of an incomplete, charred skeleton – like a failed attempt to burn the body – whose orientation to the west deviates from all other burials.

    My understanding on the basis of the above is that the Bolshoy people are related to the Garino Bor culture of the Urals.

    As a sidenote I would like to add that according to the paper "it can be suggested that the Bolshoy population suffered from a low level of insolation and consequent deficiency of vitamin D, cold stress and a shortage of plant foods." Therefore, it is not probable that the Bolshoy population had any relevant impact on the genetics of the modern populations.

    ReplyDelete
  3. @Kristiina

    As a sidenote I would like to add that according to the paper "it can be suggested that the Bolshoy population suffered from a low level of insolation and consequent deficiency of vitamin D, cold stress and a shortage of plant foods." Therefore, it is not probable that the Bolshoy population had any relevant impact on the genetics of the modern populations.

    Even if this is true, the existence of the so called Uralic cline suggests strongly that the N1c-rich close relatives of the Bolshoy population had a profound genetic impact across the north of Eurasia.

    ReplyDelete
  4. @All

    Oops, I have to run these samples again. I left out some of the chromosomes by mistake in my first run.

    The results won't be vastly different, but certainly more robust.

    ReplyDelete
  5. @ Davidski

    Yes, but the core area of the Siberian ancestry is not in the far north but in Western Siberia in the cultures containing a lot of Siberian mtDNA, see V.I. Molodin et al., "Human migrations in the southern region of the West Siberian Plain during the Bronze Age: Archaeological, palaeogenetic and anthropological data" (https://www.degruyter.com/downloadpdf/books/9783110266306/9783110266306.93/9783110266306.93.pdf).

    Forest Volga is still completely unsampled, so we have no data from the core area of the Uralic languages.

    ReplyDelete
  6. @Kristiina

    Forest Volga is still completely unsampled, so we have no data from the core area of the Uralic languages.

    If that's indeed the core Uralic area, then the samples from here will cluster along the Uralic cline, but, I suppose from what you say, east of the Bolshoy population.

    On the other hand, we'll obviously have a problem if they don't cluster along the Uralic cline. But they probably will, considering how the Mari, Udmurts and Mansi cluster.

    ReplyDelete
  7. @ Davidski

    The core area of the Siberian ancestry as found in modern Uralics is with all probability in Western Siberia, but these populations only introduced the Siberian subtrate(s) of the modern Uralic varieties. The Proto-Uralic probably developed in the forest Volga in a close relationship with the Bronze Age IE languages.

    ReplyDelete
  8. @Kristiina

    I'm gathering that you are more in favor of the pre-proto Uralic speakers being EHG like, rather than WSHG like.

    You are probably aware of the studies from linguists like Haakinen claiming a Siberian origin of Pre-proto Uralic based on shared morphology with Siberian and "Altaic" languages.

    Do you have a reason for disagreeing with these conclusions? I agree that proto Uralic itself is most likely derived from Garino-Bor or nearby, but I don't currently think it has a direct genetic relationship with PIE nor did it develop originally out of the local EHG population of the forest Volga.

    It will be exciting to see aDNA from the Neolithic and BA Volga-Kama region certainly!

    ReplyDelete
  9. @All

    As far as I know, I've updated all of the Global25 datasheets with the new, more robust coordinates for the ancient samples in question. But please let me know if you spot anything strange.

    @Kristiina

    What Synome said.

    ReplyDelete
  10. I think "Para-Uralic" (cf. Kallio 2015) would be a good term to describe what language the Bronze Age Kola Peninsula inhabitants could have spoken.
    But is there a good reason not to assume that they were speakers of a lost dialect of "West-Uralic/Pre-Finno-Saamic" proper?
    If they disappeared before the arrival of Saami speakers in Fennoscandia c. 2000 years later, I guess we wouldn't expect them to leave any trace in terms of place names, etc.

    At least the very old loanwords probably from the Germanic branch, where the only parallels are attested (looking almost PIE in structure, and then continuously and increasingly at later stages of Pre-Proto-Germanic, and up into Proto-Germanic, Proto-Norse, Old Norse, and Swedish), would suggest that early stages of Finnic and Saamic ("Pre-Finnic/Saamic") were already in Finland or nearby around the Bronze Age, along with the presence of South Scandinavian material culture and burials on the coastal part of Finland from c. 1800-1600 BCE (see e.g. various papers by Mika Lavento).

    Of course it is difficult to show that with aDNA without any surviving skeletons from Bronze Age Finland. :)

    ReplyDelete
  11. I dont see a reason why Uralic has to be related to any other major language family. It seems like once past the Pre-Proto stage things become too pseudo sciencey for my tast. An Altaic family linking Estonian with Japanese and an Indo-Uralic family linking Nganassans with Spanish both seem silly to me.

    That being said I believe you will see N1c and WSHG or EHG+Nganassan like people in the Volga-Kama during the early Bronze age speaking PU. Prior to that I don't know but most likely West Siberia.

    ReplyDelete
  12. Why modern Uralic speakers, noh Indo-Europeans, are the closest to Yamnaya out of all modern populations?

    ReplyDelete
  13. Very interesting, many thanks, D. How would WSHG and Sintasha_o3 look like in the PCA, vs. the "Uralic cline"?

    ReplyDelete
  14. @ Synome "
    Do you have a reason for disagreeing with these conclusions?"

    Yes,

    http://eurogenes.blogspot.com/2015/09/support-for-linguistic-macrofamilies.html

    Figure 1., http://rspb.royalsocietypublishing.org/content/277/1693/2443

    https://en.wikipedia.org/wiki/Indo-Uralic_languages

    ReplyDelete
  15. @ Mikkel

    If they disappeared before the arrival of Saami speakers in Fennoscandia c. 2000 years later, why should we think that they spoke a Western Uralic language?

    One of the Bolshoy genomes has been uploaded to Gedmatch: SH8213729 - Bolshoy_Oleni_Ostrov BOO001 (MDLP K16 identified as Mansi)

    Mansis speak an Eastern Uralic language and their original home area is presumed to be West of the Urals in the area of influence of Garino Bor. It would be logical to think that the Bolshoy language was related to the proto-Eastern Uralic.

    ReplyDelete
    Replies
    1. I meant fennoscandia in the sense "north of Finland". So what I tried to suggest was that they could have been a linguistic early extention of, e.g. early pre-saami/west uralic north of Finland. But in any case I was speculating.

      Interesting if the bolshoy individuals are closer to Mansi than to West uralic soeakers. Did not know that.

      However, I'm not sure it says anything against pre-saamic arriving in Finland (without reaching the Kola Peninsula) in the bronze age?

      Delete
  16. @Kristiina

    There is T2d1b1 in Volga-Ural magyars too.

    https://docs.wixstatic.com/ugd/0735a8_fe67b071f7964118aad4b1237282caba.pdf?fbclid=IwAR2iFY6x0d04D4h6iFhGh_AIrA7KK5Np09atDKh8tXH_RvlT6_g2jv31cZ0

    Page 218

    ReplyDelete
  17. @Dopa

    Why modern Uralic speakers, noh Indo-Europeans, are the closest to Yamnaya out of all modern populations?

    They're not.

    ReplyDelete
  18. @Huck Finn
    "Very interesting, many thanks, D. How would WSHG and Sintasha_o3 look like in the PCA, vs. the "Uralic cline"?"

    There are two Sintashta_o3, I marked also Sintashta_o2:

    https://i.postimg.cc/d3x2XZLX/screenshot-455.png

    It looks like Sintashta outliers may have something to do with ‘Uralic cline’. Very interesting.

    ReplyDelete
  19. Here's the main plot with West_Siberia_N and Sintashta_MLBA_o3 highlighted (btw, anyone can do this with PAST).

    Uralic_cline_Global_25_PCA2

    ReplyDelete
  20. @Davidski

    Mordovians are closer to Yamnaya than any European is, and they speak an Uralic language.

    ReplyDelete
  21. @Kristiina - "http://eurogenes.blogspot.com/2015/09/support-for-linguistic-macrofamilies.html"

    Doesn't that support @Synome's point?

    I would say I think it's plausible that Eurasiatic family consisting of Indo-European, Altaic and Uralic whose common source is ANE.

    ReplyDelete
  22. If Sintashta was Proto-Indo-Iranian, then Sintashta outliers could well be Proto-Uralic (or was it rather Proto-FU?) people, because all FU people share early Indo Iranian loanwords.

    ReplyDelete
  23. Sintashta Proto-Indo-Iranians probably had contacts with all sorts of North Eurasian groups once they moved into the south Ural area, including early Uralics.

    And I'd say that most of these North Eurasian groups would sit along the Uralic cline in my PCA.

    So the Sintashta_MLBA_o3 outliers might be early Uralics, or they might be, say, early Yeniseian speakers. Or a mix of both.

    ReplyDelete
  24. @Dopa

    Mordovians are closer to Yamnaya than any European is, and they speak an Uralic language.

    The differences between many Central/Eastern/Northern European populations aren't statistically significant in this respect.

    All you can really say is that Mordovians are one of the Eastern European populations that show the highest affinity to Yamnaya.

    But Mordovians obviously have a lot of Indo-European ancestry, because they share basically all of their uniparental markers with Balto-Slavs.

    However, unlike most Balto-Slavs and Yamnaya, they also show excess eastern ancestry from Siberia that they share with Siberian Uralic-speakers, hence this is the genetic structure that can be said to be associated with their Uralic language.

    So I don't know if you're just trolling here or what...

    ReplyDelete
  25. Many thanks, D. Looks pretty good, in my opinion. I'd guess that both WSHG and Sintasha_o3 land somewhere next to Udmurts in the PCA?

    ReplyDelete
  26. 1.6727"

    Finnish_East

    Levanluhta_IA,31.5
    Avar_Hungary_Szolad,23.7
    Baltic_BA:Turlojiske3,21.5
    Nordic_IA,18.1
    EHG,3.1
    Narva_Lithuania,2.1

    =1.4527"

    Saami

    Levanluhta_IA,64.5
    Nordic_IA,11.9
    Baltic_BA:Turlojiske3,10
    Nganassan,4.6
    Khanty,3.5
    Narva_Lithuania,2.9
    EHG,1.8
    Chukchi,0.7
    Yakut,0.1

    ReplyDelete
  27. People talk about how ancient DNA has challenged traditional ideas of homeland, structure of human diversity, etc, etc. Obviously they're mostly referring to Europe when they say this.

    However, if you think about Europeans are all very similar and have a very similar history. Since the Mesolithic most humans in the socially constructed location of Europe have been similar. When outsiders migrated into it, the outsiders (Anatolia, East Euro Steppe) impacted the whole continent making everyone similar once again.

    Ancient DNA is showing a much more diverse history for people in many locations in Central Asia & Northern Eurasia (including Finland, Russia). Some however have a simple history, like Ulchi.

    ReplyDelete
  28. @Huck Finn

    I'd guess that both WSHG and Sintasha_o3 land somewhere next to Udmurts in the PCA?

    No need to guess, you can download the datasheet and check with PAST (I'm out and about right now, so won't be able to do that for a while).

    But even if so, it might be somewhat coincidental because you're only looking at 2 PCs. You should try modeling the ancient ancestry of West_Siberia_N, Sintashta_MLBA_o3 and Udmurts using Global25 data to see how they differ in terms of ancestry sources and proportions.

    ReplyDelete
  29. Btw, off topic but, anyone following the debate around the new Neanderthal ancestry paper on Lazaridis twitter? Basically about basal dilution of Neanderthal in West Eurasians vs East Asian enrichment from extra Neanderthal pulses.

    One thing that I did find quite interesting is that the supplement for the paper seems to have a lot of direct population estimates, using the direct Vindija-Altai Neanderthal based estimates: https://pbs.twimg.com/media/DtLiPH-XgAYx7an.jpg:large.

    Now these are too noisy to really take seriously as individual population estimates, but I figured that by taking them and regressing with the G25 you might get some decent noise free estimates. (Taking G25 as a very strong model of modern day population structure).

    I actually found G25 to have slightly too many dimensions (independent variables) as is optimal for this, slightly detailed population structure, and the results were slightly overfit (e.g. there were some slight tendencies in regression estimated Neanderthal correlating with modern drift dimensions which seems false).

    But, still, one thing that did come out was that the early Neolithic Near Eastern populations - Levant_N, Natufian, Iran_N, Barcin_N - could only be consistent with the all the modern populations and the Neanderthal values in the modern Near East (and East and North Africa) if they had about 80-90% of the Neanderthal of WHG/EHG, using 0 as a baseline, or 70% of WHG/EHG using Africa as a baseline*. That is, WHG/EHG about 4% Neanderthal, most extreme ancient Near East about 3.2-3.6% Neanderthal, West Africa about 1%. Quite interesting to me as hard to be consistent with models where Natufian / Iran_N have extremely high share of BEu, assuming BEu has 0 or African levels of Neanderthal?

    *Africa coming out about 1% Neanderthal considering estimates using Vindija and Altai which should be bias free, seems like an interesting line of evidence for Eurasian backflow to Africa in some form?

    ReplyDelete
  30. @ Matt

    My question is. If two populations split from a common Ancestor, shouldn't they have more in common with each other and gradually loose their commonality as time goes by and they evolve their seperate ways ?

    ReplyDelete
  31. @Matt,

    So you saying Meso/Neo Near East had only a little less Neanderthal than Meso Europe therefore their basal Eurasian ancestry is low?

    ReplyDelete
  32. List of pops Levanluhta_IA_o is closest to in G25 PCA.

    Levanluhta_IA_o
    England_Anglo-Saxon 0.033049371
    England_MBA 0.033224732
    England_CA_EBA 0.033981037
    Beaker_Britain 0.03404148
    Scotland_MBA 0.035419263
    Netherlands_BA 0.035662186
    Icelandic 0.035841696
    Nordic_LN 0.036475701
    Norwegian 0.036505981
    Ireland_EBA 0.036777007
    Swedish 0.037089514
    Unetice 0.03718517
    Beaker_The_Netherlands 0.037557766
    Scotland_CA_EBA 0.038088509
    BenzigerodeHeimburg_LN 0.039579922
    Scotland_LBA 0.040985565
    Welsh 0.041332174
    Nordic_IA 0.041539132
    Poland_BA 0.042167923
    Czech_EBA 0.043478443
    Halberstadt_LBA 0.044084427
    Orcadian 0.044337081
    Irish 0.044841601
    Scottish 0.044898587
    England_Roman 0.045093164
    England_IA 0.045676854
    Dutch 0.046549305
    Shetlandic 0.04781032

    ReplyDelete
  33. Nordic_LN:RISE98 3.332508
    Nordic_LN:RISE97 4.797899

    ReplyDelete
  34. Amoung modern populations he's closest to: Icelandic, Norwegian, Swedish. Danes aren't included in G25. It's safe to assume Levanluhta_IA_o and Norse are connected somehow. Also, maybe can explain ancient branches of yHG I1 in Finland unless he was purely an immigrant not a settled population.

    ReplyDelete
  35. @Ryan "I would say I think it's plausible that Eurasiatic family consisting of Indo-European, Altaic and Uralic whose common source is ANE."

    I do not believe in that ANE narrative. The highest ANE percentages are in Native Americans and Eskimos, and Yeniseians come next. In that macrofamilies paper, Yeniseians are linguistically as far from IEs as possible. In the other paper, Native American languages are very far from IE, Uralic and Turkic languages.

    ReplyDelete
  36. I tried dissecting ancient components in Bolshoy_Oleni_Ostrov like I did for modern pops. My estimate is.

    42% ENA, 35% ANE, 13% WHG, 10% Sintashta (because Mansi has Sintashta admix and gives Bolshoy good fits, might just be more ANE).

    Here's Udmurt. Unlike, early Saami, looks like a ANE-rich Siberian+Sintashta-like mix.

    2.5355"

    Udmurt

    Mansi,42.9
    Sintashta_MLBA,28.5
    Battle_Axe_Sweden,14.3
    Baltic_BA,11.2
    Lokomotiv_N,3.1

    ReplyDelete
  37. For the most half northern European, half ANE-rich Siberian? Sintashta ancestry might not fit in the picture.

    2.5505"

    Levanluhta_IA

    Baltic_BA,25.5
    Icelandic,22.7
    Bolshoy_Oleni_Ostrov,21.1
    Mansi,19.5
    Nganassan,8.7
    Narva_Lithuania:Kretuonas1,1.8
    Narva_Estonia:Veibri4,0.4
    Narva_Lithuania:Donkalnis6,0.3

    2.3128"

    Saami

    Baltic_BA,25
    Icelandic,19.6
    Bolshoy_Oleni_Ostrov,18.3
    Mansi,14.9
    Nganassan,9.4
    Nordic_IA,8.7
    Narva_Lithuania:Spiginas1,4.1

    ReplyDelete
  38. 0.9339"

    Finnish

    Icelandic,24.2
    Baltic_BA:Turlojiske3,19.4
    Baltic_IA,15.6
    Nordic_IA,9.8
    Swedish,8.9
    Baltic_BA,7.3
    Mansi,5.4
    Norwegian,4
    Nganassan,2.7
    Bolshoy_Oleni_Ostrov,1.6
    Narva_Estonia:Veibri4,1.1

    ReplyDelete
  39. I've got an interesting post coming tomorrow about the Proto-Uralics.

    Stay tuned...

    ReplyDelete
  40. Any chance the PU groups that migrated to Finland and Estonia and related groups had IE ancestry or were they plain Siberian? This convo has me confused all over again.

    ReplyDelete
  41. I'll explain tomorrow. But anyone who still thinks that the Proto-Uralics were anything like Corded Ware/Sintashta should find a good therapist.

    ReplyDelete
  42. Nothing new under the sun. It has been clear to everyone, except to those wandering with linguistic Volga Bend agendas, that the Siberian admixture, or origin if we suggest a large scale Siberian migration, came to Fennoscandinavia via the northeast home range of Asia. Very unhappy about the way how the gap between linguists and natural science has been narrowed by a scrappy and all-embracing interdisciplinary ideology and some geneticists prefer linguist ideas over their own.

    ReplyDelete
  43. @Sam, yeah, I'm saying that all of:

    1. "Iran N, Levant_N and Natufian have large shares of Basal Eurasian" (even 30-45% in Lazaridis's 2016 paper for example)

    2. "Modern Near Eastern / South Asian / North and East African populations are as descended from them as they appear to be"

    3. "Basal Eurasians had no Neanderthal ancestry"

    seem like they can't all be true at the same time, given the trends in Neanderthal ancestry in the 126 modern populations (and especially if including their 2 high coverage ancients) in Schraiber's paper.

    One's gotta break, and since it's unlikely to be 2, then either the early Near Easterners don't have much Basal Eurasian (1 breaks), or Basal Eurasians have some Neanderthal ancestry, if a bit less than Eurasians in general (3 breaks).

    As well as that, since it seems like Africans who are in theory unadmixed may come out as having 1/4 as much Neanderthal as Eurasians (1% vs 4%), and that's judged by Vindija vs Altai sharing, so shouldn't in theory be affected by any African archaics, implying an early Eurasian wave back (25% ancestry). If that's correct, that also inclines me to think it seems a bit less likely that Basal Eurasians had no Neanderthal ancestry.

    ReplyDelete
  44. Here's something interesting for everyone to ponder for now...

    Bolshoy_Oleni_Ostrov can't explain the eastern ancestry in Levanluhta_IA, Saami or Finns. At least not all of it.

    And neither can West_Siberia_N or WSHG.

    That's because their ratios of ANE/EHG to Nganasan-related ancestry are way off.

    What this of course means is that additional, post-Bolshoy_Oleni_Ostrov migrations from the east are needed to explain the present-day Fennoscandian genetic landscape.

    ReplyDelete
  45. @Sam and others, if you'd like to look - here's the basic correlations of the Neanderthal estimates from Schraiber's paper (as I read them in the figure) with G25 for populations that overlap: https://imgur.com/a/7Uw3qWB

    Here's the regression based on the above data with all G25 populations: https://imgur.com/a/dJxwboD .

    Regression looks pretty expected to me, only a few odd outliers - CHG, Iberomaurasian, Mari; those probably relate to a limited representation of North African, Caucasus and West Siberian data in the conditioning set for the regression and chance meaning they have erroneously low or high values. I suspect if you ran this with double the populations (go from 126 to 252 and perhaps some more high coverage ancients) outliers would tend to fade out. But it seems clear that there generally aren't any strong tendencies on dimensions where ancient Near East and ancient European HG are at polar positions and in theory characterised by rising BEu...

    ReplyDelete
  46. @Davidski, it doesn't need to BOO, although your PCA suggest so :) Neither Nganasan. Many FU groups lived in North Russia before Slavs and disappeared around 1000 years ago.

    ReplyDelete
  47. @M. Myllylä

    Not really sure what you mean there, but to get from Bronze Age Northern Europe to the earliest Uralics in the ancient DNA record requires quite a bit of eastern input, and in fact more eastern than just Bolshoy_Oleni_Ostrov...

    Levanluhta_IA
    Levanluhta_IA_o 0.597
    Bolshoy_Oleni_Ostrov 0.241
    Glazkova_BA 0.162

    Levanluhta_IA
    Levanluhta_IA_o 0.564
    Bolshoy_Oleni_Ostrov 0.299
    Nganasan 0.137

    ReplyDelete
  48. It's hard to be too confident. I guess my feeling is that even the earliest Uralics in dna record are not very early (seems at best 2000 years after the proto-language, but likely a bit more), and there is ample time for the language family to have dispersed and admixed such that we wouldn't have a very good picture from a few samples that we could confidently assign. (It looks like know IE hopped to almost wholly different genetic backgrounds quite quickly, as another example, sometimes with replacement of y markers as well certainly within 2000 years). I'll await the post with interest though!

    ReplyDelete
  49. @Matt

    Regarding your Neanderthal trilemma, given the recent studies that found BE proportion in West Eurasians to be considerably lower than the original estimates given by Lazaridis, I'd vote for option #1 being most likely to prove untrue.

    Looking forward to reading and discussing more about Uralic origins.

    ReplyDelete
  50. @Davidski, Levanluhta outlier obviously wasn't a Baltic Finn and can't be a relevant representative of the European side. We still have to speculate about the Baltic Finn 1500-2000ybp, because we have not ancient samples. Maybe we see it in the nearest future. Of course there must have been "hard core" Siberians very close Fennoscandinavia and we don't even need the BOO. The whole White Sea region was populated by FU people and also Siberians. We don't need millions or even tens of thousands Siberians, just a few small packs of them, because even in the southernmost Finland there was only about a few thousands people.

    ReplyDelete
  51. @M. Myllylä

    Levanluhta outlier obviously wasn't a Baltic Finn and can't be a relevant representative of the European side.

    No, this sample is relevant. That's why the model works.

    Everything makes perfect sense, because Levanluhta_IA_o is obviously a Germanic who lived in Saami-speaking Finland before Finns got there.

    ReplyDelete
  52. By the way, this is how Levanluhta_IA and Levanluhta_IA_o cluster in Northern Europe based on Global25 data...

    Levanluhta_IA_G25_PCA

    ReplyDelete
  53. The Levaluhta region is known about Germanic Iron Age items (Vendel era). Those items have exact similarity with items found from Sutton Hoo/England and Vendel/Sweden. We can be sure that there was a Germanic connection.

    ReplyDelete
  54. Yes, we can, also because we now have a Germanic genome from that area and time.

    That's who the Saami in the region mixed with. They didn't mix with Baltic Finns, because they weren't in that area at the time or earlier.

    ReplyDelete
  55. @Davidski

    "@M. Myllylä

    Levanluhta outlier obviously wasn't a Baltic Finn and can't be a relevant representative of the European side.

    No, this sample is relevant. That's why the model works.

    Everything makes perfect sense, because Levanluhta_IA_o is obviously a Germanic who lived in Saami-speaking Finland before Finns got there."

    Yeah, it can be relevant for Levaluhta samples, but not for present-day Finns. Did we speak about the Siberian admixture in the present-day Finland? Maybe not, sorry my mistake. So far we have not ancient Finnish samples.

    ReplyDelete
  56. @M. Myllylä

    Yeah, it can be relevant for Levaluhta samples, but not for present-day Finns. Did we speak about the Siberian admixture in the present-day Finland? Maybe not, sorry my mistake. So far we have not ancient Finnish samples.

    Finns are a somewhat separate issue.

    The point I was making was that Levanluhta_IA (Saami Uralic) was a mixture of Levanluhta_IA_o (Germanic), Bolshoy_Oleni_Ostrov (?), and something like Nganasans.

    So if Bolshoy_Oleni_Ostrov weren't Uralics, then Proto-Uralic ancestry in that model is represented by Nganasans.

    I'll only be able to start tackling the issue of Baltic Finns and the ancestry of present-day Finns when I get those N1c-rich samples from the Tarand graves from Estonia.

    ReplyDelete
  57. Something to consider is that the two BOO men are N-L1026 which is the main Uralic branch and not that old. This would suggest BOO already had contact with Uralic or Para-Uralic speakers.

    ReplyDelete
  58. @Anthony Haken

    Something to consider is that the two BOO men are N-L1026 which is the main Uralic branch and not that old. This would suggest BOO already had contact with Uralic or Para-Uralic speakers.

    I'm not entirely convinced that BOO wasn't Uralic. These people may have spoken an extinct Uralic language.

    But whatever they were, they weren't eastern enough to explain the ancestry of Levanluhta_IA. This suggests that extant Uralic languages may have come from the east after BOO.

    ReplyDelete
    Replies
    1. @Davidski

      Depends how much you want to rely on archeology. BOO is connected to the Ymyyakhtakh culture from Yakutia meaning BOO is either a mix of two populations or N-L1026 was spread prior to Uralic.

      Also do you think Baltic_IA is a suitable stand in for the Tarand grave samples for now? It is closest to modern day Baltics but from the leaked chart the Tarand grave samples appear to be as well. It is currently the oldest N sample in the Baltic.

      Delete
  59. @ Davidski

    So, you claim that all ancestry in Leväluhta Saamis that is not from Bolshoy or Nganasan/Glazkovo (i.e. almost 60% of their ancestry) is from Germanics because one in seven samples is Germanic like. You seem to claim that this Germanic sample engendered the 60% of the ancestry of the Saamis buried in the same lake. If that’s really the case, shouldn’t the Leväluhta population speak Germanic then?

    ReplyDelete
  60. @Anthony Haken

    Depends how much you want to rely on archeology. BOO is connected to the Ymyyakhtakh culture from Yakutia meaning BOO is either a mix of two populations or N-L1026 was spread prior to Uralic.

    BOO is certainly a mix of two different populations, because BOO006 is way more eastern than the rest. Take a look at the PCA.

    But this is a female with mtDNA D.

    Also do you think Baltic_IA is a suitable stand in for the Tarand grave samples for now? It is closest to modern day Baltics but from the leaked chart the Tarand grave samples appear to be as well. It is currently the oldest N sample in the Baltic.

    It's probably very similar to some of those Tarand grave samples. I'll have a look what it can do for Finns and Estonians.

    ReplyDelete
  61. @ Anthony "BOO is connected to the Ymyyakhtakh culture from Yakutia"

    C. Carpelan, an archaeologist who is one of the most respectable specialists in pottery, refutes that connection. This is from the report of the burial ground:
    "In discussing the cultural connections of the population buried in the KOG, it is necessary to mention the "waffle"-stamp-decorated vessel from burial X. Already N. Gurina noted the similarity of this vessel and the Neolithic pottery of Yakutia and Chukotka. V. Shumkin suggested that the appearance of "Waffle" ware on the Kola Peninsula may indicate the infiltration of Ymyyakhtakh culture from Siberia. However, according to C. Carpelan, the "waffle" Ware of the Kola Peninsula belongs together with Vardoy Ware, which was distributed over northern Norway and northern Finland. He also points out that such pottery decoration is not known in the area between northern Fennoscandia and the Taimyr Peninsula."

    Moreover,

    ReplyDelete
    Replies
    1. @Kristiina

      Even if Carpelan is right, at the end of the day half of BOO's ancestry is Siberian which requires an East to West migration from beyond the Urals.

      Delete
  62. @Kristiina

    If that’s really the case, shouldn’t the Leväluhta population speak Germanic then?

    The models I posted don't exactly show that almost 60% of the ancestry of the Leväluhta Saamis is from a Germanic source.

    What they show, of course, is that Levanluhta_IA_o happens to be a good surrogate for the western ancestry in these Saamis.

    So a portion of this ancestry might not be from a Germanic source, but considering the stats, most of it probably was.

    I can't answer the question of why these Saamis didn't speak Germanic. But I remember seeing a post by Mikkel here pointing out that there was significant Germanic influence on Saamic. So things seem to make sense more or less.

    ReplyDelete
  63. I'm sure this will become more clear with the upcoming post on Uralic peoples, but personally I wouldn't expect the proto-Uralic people to be completely Nganasan-like.

    Is there no room for some of that puported Germanic ancestry to really be Eastern European CW type ancestry?

    ReplyDelete
  64. @Synome

    Is there no room for some of that puported Germanic ancestry to really be Eastern European CW type ancestry?

    Yes, there is, and it's a lot more relevant than CW, but I'll explain this in my upcoming post.

    ReplyDelete
  65. The paper its self clearly suggests waves of eastern influences coming to Northern Fennoscandia leading up to 1500BC.

    From Lamnidis et al.

    "with the introduction of asbestos-mixed Lovozero ceramics during the second millenium BC, and the spread of even-based arrowheads in Lapland from 1,900 BCE. Additionally, the nearest counterparts of Vardøy ceramics, appearing in the area around 1,600-1,300 BCE, can be found on the Taymyr peninsula, much further to the east. Finally, the Imiyakhtakhskaya culture from Yakutia spread to the Kola Peninsula during the same period".

    Even if BOO belonged to a local culture it must have had contact with those just entering Kola peninsula. Which seems to be supported by the aDNA.

    ReplyDelete
  66. @Davidski

    From what I have seen on other forums Baltic_IA and BOO006 work well for modeling Levanluhta_IA with some sort of Germanic stand in leaving no need for extra Nganassan.

    ReplyDelete
  67. @Anthony Haken

    From what I have seen on other forums Baltic_IA and BOO006 work well for modeling Levanluhta_IA with some sort of Germanic stand in leaving no need for extra Nganassan.

    Yes, but BOO006 is a female outlier in the BOO group, which suggests that this type of ancestry wasn't yet widespread in Northeastern Europe at that time.

    I think that it became widespread with the Uralic expansion which came soon after.

    ReplyDelete
  68. @ Anthony

    Lamnidis et altri are geneticists and not archaeologists, therefore I rather believe C. Carpelan who is a Finnish archaeologist specialized in ceramics of northeast Europe.

    ReplyDelete
    Replies
    1. @Kriistina

      Yes but genetics can also alter our idea of population origins vs. archaeological assumptions, not to mention that archaeologists often in recent times assumed there were few, if any, population movements ever. For example, it was, and in some circles it still is, widely believed that the Indo-Europeans did not migrate in large numbers and that it spread by elite dominance or, worse, cultural diffusion; these are often the same kinda people who claim IEs are not an ethnic grouping but a linguistic one only.

      This is now all proven wrong, and the data clearly shows a mass migration of steppe populations.

      That’s my counter argument, anyway, but my point is archaeology needs genetics as much as genetics needs archaeology.

      Delete
  69. @Davidski

    "Yes, but BOO006 is a female outlier in the BOO group, which suggests that this type of ancestry wasn't yet widespread in Northeastern Europe at that time.

    I think that it became widespread with the Uralic expansion which came soon after."

    Like you said earlier it seems that BOO is a mix of two populations based on BOO006 being a clearly more eastern outlier. The paper itself suggest there were non Uralic Siberian migrations to the area circa 1500BC. Both of the male western samples belong to N1c while BOO006 is female with eastern MtDNA. Also Netted Ware would have been presesnt in the area during this time which likely was Uralic speaking.

    If Baltic_IA is decended at least partially from Balto-Finns and Levanluhta_IA can be modeled with it along with BOO and Germanic then that should be the basis for Proto-Finnic as Germanic and BOO had to have been picked up by Proto-Saami in Fennoscandia.

    That being said I wont comment on Proto-Uralic as I fully expect to see some easterm admixture when samples from the BA Volga-Kama finally surface.

    ReplyDelete
  70. @Davidski, @Huck Finn
    "But even if so, it might be somewhat coincidental because you're only looking at 2 PCs. You should try modeling the ancient ancestry of West_Siberia_N, Sintashta_MLBA_o3 and Udmurts using Global25 data to see how they differ in terms of ancestry sources and proportions."

    I actually did something like that a few months ago with an earlier version of the datasheet and the on-line automated nMonte runner.
    Notably, I could not find any good fit for Undmurts. The best fit included a modern population and that was still bad.

    Udmurts (fit: 4.8355)
    CWC Baltic 0
    Mezhovskaya 70%
    Russian Voronez 13.33%
    Shamanka N 16.67%

    However the most notable difference between them and Sintasta_O3 is that Sintashta_O3 does not take any Baikal region ancestry, while Udmurts do, even if Sintashta_O3 itself also included.

    Udmurts (fit: 5.1082)
    CWC Baltic 12.5%
    Mezhovskaya 67.5%
    Shamanka N 15.83%
    Sintashta MLBA O3 4.17%

    vs.

    Sintashta O3 (fit: 3.9502)
    Shamanka N 0
    Sintashta MLBA 25%
    West Siberia N 75%
    West Siberia N Low Res: 0

    Again, the fits are bad, I am obviously missing something, but the main point is pretty clear even so.

    ReplyDelete
  71. @Synome, yeah, though the thing that tends to strike me is that those models have tended to suggest BEu is lower in EEF, which leaves open the possibility of higher levels in Iran_N / Natufian / Levant_N, which matches qpGraph. Whereas this suggests it is universally fairly low.

    The whole concept of Basal Eurasian becomes odd if BEu hasn't got the Eurasian Neanderthal admix - even a little - and Africans have about 1%, which would have to come from back migration by a post-Basal Eurasian population, but one which isn't particularly ancestral to any present day Eurasian population. It is something like Ust Ishim back to Africa.

    ReplyDelete
  72. Off topic:

    Some cool Tollense news from... Instagram:

    by @abenteuer_bronzezeit 25 November, 2018
    Yesterday at the BALTIC BRONZE AGE CONFERENCE we also talked about the Battle in the valley of River Tollense (Germany). They actually found 60 arrowheads, 50 made of bronze, and 10 made of flint stone. On some flint stone arrowheads they found blood cellls in good condition. Now they will try to extract DNA from this cells. On the pic you can bronze arrowhead, stucking in a skull.

    Circum-baltic Interaction in the Bronze Age (CIBA) - Universität Hamburg
    https://www.e-a-a.org/EAA/Publications/Tea/Tea_56/Announcements/EAA/Navigation_Publications/Tea_56_content/Announcements.aspx#56_CIBA


    Genes, Isotopes and Artefacts - upcoming conference in Wien.

    https://www.orea.oeaw.ac.at/veranstaltungen/event-detail/article/genes-isotopes-and-artefacts/

    Participants include Morten Erik Allentoft, Andrea Cardarelli, Claudio Cavazzuti, Edward Caswell, Peter Clark, Karin Margarita Frei, Catherine Frieman, Mario Gavranović, Jelena Grujić, Wolfgang Haak, Johannes Krause, Anthony Harding, Barbara Horejs, Reinhard Jung, Viktória Kiss, Corina Knipper, John Koch, Kristian Kristiansen, Gabriella Kulcsár, Anne Lehoërff, Andrew Millard, Barry Molloy, Janet Montgomery, Ron Pinhasi, Miljana Radivojević, Katharina Rebay-Salisbury, David Reich, Martin Sikora, Philipp Stockhammer, Benjamin Roberts, Natalia Shishlina, Robin Skeates and Marc Vander Linden.

    The event is free, but space is limited.



    Steppe and Iranian ancestry among Bronze Age Central and Western Mediterranean populations

    Ron Pinhasi, Daniel Fernandes, David Reich

    Steppe-related ancestry is known to have reached central Europe ca. 3000 BCE, while Iran-related ancestry reached Greece by 1500 BCE. However, the time course and extent of their spread into the central/western Mediterranean remains a mystery. We analysed 48 Neolithic and Bronze Age individuals from Sicily, Sardinia and the Balearic Islands aiming to investigate when and how continental European and Aegean influences affected these insular populations. Results show that the first Balearic settlers had substantial Steppe-related ancestry which was subsequently diluted by increasing proportions of farmer-related ancestry. In Sardinia, we identified the appearance of Iran-related ancestry from the Aegean as early as the Middle Bronze Age, with no genetic influences seen from populations carrying Steppe-related ancestry despite cultural or commercial exchanges with Bell Beaker populations. In Sicily, during the Early Bronze Age, and possibly earlier, we found evidence for admixture with groups carrying both these ancestries. These results suggest that Steppe-related migrants had a crucial role in the settlement of the Balearic Islands and their ancestry reached as far south as Sicily, and that the population movements that brought Iran-related ancestry to the Aegean also impacted the Western Mediterranean around the same time the first civilizations started to develop.

    ReplyDelete
  73. "Iranian"-related? All that means is it can come from any where in West Asia. It isn't nearly a comparably accurate term as Steppe.

    ReplyDelete
  74. @Arza
    “Steppe-related ancestry is known to have reached central Europe ca. 3000 BCE, while Iran-related ancestry reached Greece by 1500 BCE.”

    Iran-related ancestry reaching Greece by 1500 BCE surely wasn’t IE as IE just started to arrive in India and Iran at that time. Interesting how that not-IE component influenced Greek language which is 50% not-IE.

    ReplyDelete
  75. @Matt

    Re: Neanderthal in Africa, it reminds me of the original (wrong) Mota results, where many different African populations had equal amounts of West Eurasian ancestry despite that not making any sense in terms of geography/known history/uniparentals - hard to believe that San, Mbuti, Biaka, Dinka, and Yoruba would all have the same amount of back-migration if it was recent at all. Which pointed to the actual baseline being zero inflated by a constant amount. Here the standard errors are so huge we can't tell if it's the same but the central values are suspiciously similar.

    Early modern human admixture into Vindija more than Altai Neanderthal would shift the D stats the same way wouldn't it? It doesn't seem implausible for that to have happened.

    If we do have a Back-to-Africa gene flow with Neanderthal is there really any good reason it shouldn't be closer to Crown than to Basal Eurasian? That actually might fit better in some ways.

    ReplyDelete
  76. @ Slumbery: Good point. Sintasha_o3 is indeed most probably a case of it's own and as such not like any modern Uralic population in the Ural area.

    ReplyDelete
  77. @ Davidski

    Sikora et al. 2018:
    In summary, the 87 Sr/ 86 Sr data suggest, that individual JK1963/DA238 was probably a local relying on terrestrial foodstuffs, while individuals JK1968/DA234 and JK2067/DA237 might well have been locals incorporating a component of Baltic sea resources in their diet. Individual JA2065/DA236 could have been of foreign origin, or a local using almost exclusively Baltic Sea dietary items.

    Levanluhta_IA_o:DA236 and Levanluhta_IA_o:JK2065 is one person. Same with Levanluhta_IA:JK1968 and Levanluhta_IA:DA234. Additionally - are Levanluhta_IA:DA238 and Levanluhta_IA:JK1970 from different datasets?

    ReplyDelete
  78. @ EastPole
    Iran-related ancestry reaching Greece by 1500 BCE surely wasn’t IE

    That's clear even from this abstract. On Sardinia we had Iran-related ancestry, but not the steppe one and in effect Sardinia was non-IE practically until historical times.

    ReplyDelete
  79. @ Anthony

    Those who think that one pot in one burial (burial X) is the decisive feature when determining the language spoken by Bolshoy people, go and see the pots yourself:

    Bolshoy pot, figures 2-4: http://tinypic.com/view.php?pic=21ep9v5&s=9#.XAJmKjgzaUk

    Ymyyakhtakh pots: https://bigenc.ru/archeology/text/4915666
    figure 1, https://cyberleninka.ru/article/v/vafelnaya-keramika-pozdnego-bronzovogo-veka-severo-zapadnogo-poberezhya-ozera-baykal

    ReplyDelete
  80. @Arza

    Yeah, they're from different datasets. But they weren't coming up as duplicates or relatives in my quality control scans, so I missed that.

    ReplyDelete
  81. @capra: "San, Mbuti, Biaka, Dinka, and Yoruba would all have the same amount of back-migration if it was recent at all"

    That would be an indication against if so, but the statistics are probably noisy enough and the amounts fine enough that this may well not be so. So at the moment I wouldn't consider that too strong evidence against.

    Early modern human admixture into Vindija more than Altai Neanderthal would shift the D stats the same way wouldn't it? It doesn't seem implausible for that to have happened.

    Yes, possible. The relevant papers are:

    "The limits of long-term selection against Neandertal introgression" - https://www.biorxiv.org/content/biorxiv/early/2018/07/04/362566.full.pdf

    Prüfer K, et al. (2017): "A high-coverage Neandertal genome from Vindija Cave in Croatia" Science:eaao1887 -

    http://science.sciencemag.org/content/sci/358/6363/655.full.pdf
    Hajdinjak M, et al. (2018): "Reconstructing the genetic history of late Neanderthals" Nature 555(7698):652–656 - https://www.eva.mpg.de/documents/Nature/Hajdinjak_Reconstructing_Nature_2018_2522088.pdf

    Only consideration I can find is: "However, when we applied an approach that uses the extended length of haplotypes expected from recent introgression into the analysed late Neanderthals, we did not find any indications of recent gene flow from early modern humans to the late Neanderthals (Supplementary Information 11)" in H et al, but obviously that doesn't talk about deeper, older flow.

    So that could be the case as well; I don't think there's anything a priori wrong with a complex split of Eurasians and Africans with some backflow though. If I remember some of the tree models (Schlebusch, Skoglund papers, probably others) with structure reflecting an expansion out of a population more closely related to the Out of Africa population into all other African populations (combined with deeper more basal structures) that could allow that.

    If we do have a Back-to-Africa gene flow with Neanderthal is there really any good reason it shouldn't be closer to Crown than to Basal Eurasian? That actually might fit better in some ways.

    It seems a bit less likely presuming Crown split off geographically further into Eurasia; though if any backflow was a clade with either Crown or Basal you'd have shifts in D-stats reflecting that (Basal Eurasian populations further, etc.) and so I suspect that any Neanderthal bearing population would have to be pre-Crown+Basal split (which would preclude a no-Neanderthal Basal).

    ReplyDelete
  82. @Azra, thanks for that. Interesting that this suggest Balearics first colonised by steppe shifted individuals. I wonder if this relates to some abstract I remember reading earlier this year about unusual persisting population structure there? If I recall correctly, down to the present.

    ReplyDelete
  83. @ Anthony

    I did more investigations. The core area of the waffle pottery is on the northwest coast of Lake Baikal and they are dated to 1700-1300 BC

    http://www.academia.edu/22134712/_2011_May_Waffle_pottery_at_the_Baikal_coast_%D0%92%D0%B0%D1%84%D0%B5%D0%BB%D1%8C%D0%BD%D0%B0%D1%8F_%D0%BA%D0%B5%D1%80%D0%B0%D0%BC%D0%B8%D0%BA%D0%B0_%D0%BF%D0%BE%D0%B1%D0%B5%D1%80%D0%B5%D0%B6%D1%8C%D1%8F_%D0%BE%D0%B7%D0%B5%D1%80%D0%B0_%D0%91%D0%B0%D0%B9%D0%BA%D0%B0%D0%BB_

    https://cyberleninka.ru/article/v/vafelnaya-keramika-pozdnego-bronzovogo-veka-severo-zapadnogo-poberezhya-ozera-baykal

    Waffle pottery on Taimyr is from Abylaakh I, age 3100 +/-60 years.

    Charcoal from Bolshoy Oleni Ostrov is dated to c. 1500-1200 BC.

    Vardøy Ware to which the Bolshoy pot belongs was distributed in northern Sweden and Norway in the early 2nd millennium BCE.

    Scandinavian 'Waffle Ware': http://www.helsinki.fi/hum/arla/keram/it_f1.html
    Taimyr Waffle Ware: http://tinypic.com/r/2ujqwr8/9

    Scandinavian and Bolshoy waffles are diamond-shaped, while Taimyr waffle is rectangular, and the Scandinavian 'diamonds' are older than Taimyr 'squares'.

    ReplyDelete
  84. @Kristiina

    If not from Siberia where do you suppose BOO got its "Nganassan/BHG" component from?

    Even if BOO made 100% local Vardøy ceramics I am saying there is evidence for Circum polar people in the area (as supported by the paper) around 1500BC which could help explain the eastern acestry in BOO.

    Otherwise we may not have to look any further for PU aDNA.

    ReplyDelete
  85. Same kind of thing I did with Neanderthal scores as using the G25 using Basal Rich K7*: https://imgur.com/a/pAxhjiL

    Though drawback of this is I had to use an extensive set of proxies for pops out of range of the K7 datasheet, and may result in shaky values for Southeast Asian and Oceanian component particularly. Values for components which are well represented (BR-rich, Villabruna, ANE) should be more solid.

    ReplyDelete
  86. On a separate note (and yes, I know this is mostly irrelevant), huge news with the upcoming Bronze Age conference.

    Looks like they found Steppe ancestry during the Chalcolithic in the West Mediterranean. I will repeat - Steppe-ancestry, Chalcolithic, Western Mediterranean.

    Does my (largely though not completely stolen from Tomenable) theory seem so unlikely now Davidski? L51 in Western Europe originated from a migration of L23 (pre-L51) metallurgical elites across the Mediterranean, spreading copper smelting and other cultural traits (like those present in Los Millares, just check its Wikipedia).

    Oh, and what's this - another paper in that same upcoming conference describing how the Caucasus harboured different localities of people of heterogenous ancestry which acts to "challenge many of the too simple models developed for [the hypothesis] of the Eurasian steppe zone [migration] directed towards Europe"?

    If I'm wrong in this comment, I'm going to look like a huge idiot, but it's worth the risk.

    Also, sorry for interrupting this thread, but to be honest this conference is just that important.

    ReplyDelete
  87. https://www.orea.oeaw.ac.at/fileadmin/Institute/OREA/Events/2018/Genes/GenesIsotopesArtefacts_Programme.pdf

    ReplyDelete
  88. @Arza
    “ "EastPole
    Iran-related ancestry reaching Greece by 1500 BCE surely wasn’t IE"

    That's clear even from this abstract. On Sardinia we had Iran-related ancestry, but not the steppe one and in effect Sardinia was non-IE practically until historical times.”


    They should check Graeco-Armeno-Aryan theory then and consider a possibility that some common features in Greek and Indo-Iranian were spread by areal influence across a then-contiguous Graeco-Aryan-speaking area due to the language of Iran-related not-IE population and was not inherited from Proto-Indo-European.

    ReplyDelete
  89. They seem to suggest this introduction of Steppe admix into the West Med. came from continental Europe, however if this was during the Chalcolithic as they say, this is archaeologically impossible. We do know of post-Megalithic Chalcolithic migrations spreading copper smelting technology across the Mediterranean though - which I've said is what I think was the route of L51 into Europe.

    ReplyDelete
  90. @an idiot

    Where do they say that the Balearic islands samples date to the Chalcolithic? Pretty sure they were first settled during the bronze age, steppe ancestry wouldn't be a revelation.

    ReplyDelete
  91. @Bob Floy They don't, but they say that the Steppe samples in Sicily date to possibly (possibly basically meaning they likely do but the carbon dating range is broad) to before even the Early Bronze Age - and surely the source of Steppe admixture in Sicily is the same as in the Balearic islands. Which basically means that this migration of Steppe-admixed people across the Mediterranean was before the Bronze Age.

    However they seem to be going down the path of a continental origin that then made its way island-hopping - again, if before the Bronze Age, this isn't really archaologically possible.

    ReplyDelete
  92. @An Idiot

    I can't see anything in these talk abstracts that supports your theories. You seem to be reading too much into them.

    ReplyDelete
  93. @an idiot

    "Which basically means that this migration of Steppe-admixed people across the Mediterranean was before the Bronze Age."

    In all seriousness, I don't see how what was said in that abstract equals this^.

    ReplyDelete
  94. @an idiot

    Also, even if they were suggesting that, there's still a substantial body of evidence to the contrary, which would have to be overturned somehow before the whole narrative could be flipped.

    ReplyDelete
  95. Unless they are being very lenient with their dating in saying the Sicilian Steppe-admix dates to before the Early Bronze age, a Copper Age Steppe sample from that region basically completely rules out anything to do with Yamnaya.

    And as for the abstract about the Caucasus - that is really hard to make sense of to be honest, but I presume it means that they found Steppe-admixed groups separate from the CHG groups in the Caucasus proper and (this part I'm probably reading in too much admittedly) that this Steppe signal was not DIRECTLY from the Steppe itself, but from an older migration. To be honest there's no way they came to that conclusion, but I like to think it - not least because of that Chalcolithic Armenian.

    ReplyDelete
  96. @Bob Floy By across the Mediterranean, I mean limited to the West Med only of course, as that's all that the sampling entails.

    I think they mean a migration from somewhere like France to these West Med islands, but I disagree with that - the original Copper Age Bell Beakers being Steppe-admixed in the first place fits much nicer with the technology being non-Bronze (as Yamnaya had Bronze metallurgy).

    The only other alternative besides mine, in my opinion (ASSUMING that the Sicilian example is pretty clearly from the Copper Age and they're not just being ultra-speculative), is a very early pre-Yamnaya, pre-Corded Ware migration into Europe from the Steppe that has somehow gone unnoticed (which I doubt).

    HOWEVER, the thing about aDNA is that only one sample is needed to destroy a hypothesis. So it doesn't matter how many samples you have.

    ReplyDelete
  97. "during the Early Bronze Age, and possibly earlier"

    We've all seen enough literature at this point to know that this kind of language shouldn't be interpreted too strongly, especial in an abstract. Whatever hard data they have would need to be closely analyzed first.
    And again, it wouldn't jive with the mountain of data suggesting otherwise, which should tell you that something is amiss. We've seen these research teams playing fast and loose with their results lately, for example by referring to any kind of CHG ancestry as "Iran-related".

    ReplyDelete
  98. The mountain of data means nothing if a good sample contradicts it. The Caucasus paper, I feel, will be the more important one though. If they manage to get a wholesome coverage of the entire Caucasus from the 5th millennium BC as they are suggesting they have, we should get a lot of answers.

    ReplyDelete
  99. By the way, I have no dog in this fight as such. At one time(many years ago) I supported the Renfrew theory. But the substantial data that we have at this point really does suggest that the Yamnaya/steppe theory is essentially correct, and for that to be overturned or even heavily modified there would have to be some pretty serious new discoveries which would explain away everything we have now. It's unlikely.

    ReplyDelete
  100. Shulaveri scares me though, if Olympus Mons was somehow right all along the confusion will probably put me off archaogenetics forever aha. The odds are small, but way way higher than they were only a few months ago. Here's hoping for Leyla-Tepe.



    ReplyDelete
  101. Here's a rule everyone should follow: whatever Olympus Mons says, the opposite is likely to be true.

    R1b-M269 in Shulaveri? Haha.

    ReplyDelete
  102. @Bob Floy I have no major biases either, but the Steppe hypothesis doesn't sit right in my head for various reasons (swastikas in the chalcolithic near east, modern Y DNA phylogeny distributions suggesting a homeland of L23 where Reich and Max Planck see the IE homeland, the fact that Yamnaya is clearly Z2103 and Corded Ware R1a (so where did L51 come from?), the spread of copper smelting technology, the list goes on for quite a while). A year ago, I was dead-set on Yamnaya though.

    I do not support Renfrew by the way. I support Tomenable - pre-PIE spread with copper smelting technology at first and then with the great migrational period out of the Steppe, not with agriculture.

    ReplyDelete
  103. @ An Idiot

    It looks like Sicilian Early Bronze Age postdates Central European BBC expansion.

    https://www.ancient.eu/article/1190/bronze-age-sicily/

    The Bronze Age in Sicily, considered one of the most important periods of the island's prehistory, witnessed the establishment of a unitary and in some ways artistically vibrant culture. The three main phases of the period take their name from the most important centres at the time in question: Castelluccio (Early Bronze Age)

    and

    Early Bronze Age Sicily

    In Sicily the oldest phases of prehistory were overcome at the end of the 3rd millennium BCE, when it received a new cultural wave, probably from the Middle East, today labelled with the name of the Castelluccio culture, from the homonymous prehistoric site near the city of Noto. This cultural facies (segmentation), rather unusual compared to those of the Copper Age, is verified in the south-east and south of the island, up to the provinces of Agrigento and Caltanissetta (in the west and in the middle of the island), and constitutes the “starting line” of the Sicilian bronze age. It is certainly dated to 2169±120 BCE (calibrated value) thanks to radiometric dating performed on 18 coal samples which proved to be the oldest of this culture and which were found at the archaeological site of "Muculufa", a few kilometres north-east of Licata town.


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

    Castelluccio culture is an archaeological feature dating to Ancient Bronze Age (2000 a.C. approximately) of the prehistoric civilization of Sicily

    Castelluccio culture is dated to a period between 2200 BC and 1800 BC,[2] although some believe it to be contemporary to Middle-Late Helladic period (1800/1400 BC).



    @ EastPole

    Very good point!

    ReplyDelete
  104. @Davidski The fact that they lacked copper smelting technology makes me think they were typically Caucasian in Y DNA (as you know I see M269+ as spreading copper smelting tech during its earlier phases), so I'd agree, but I wouldn't be willing to bet against Shulaveri M269. Leyla-Tepe dates to the formational period of L23 however, and does possess copper smelting tech, as well as kurgans of some form, so I think they're more likely.

    I think it's good to point out though that any specific hypothesis is likely to be incorrect. It's super unlikely that any of us have truly nailed it all, given how little data we have from the Chalcolithic period in general and especially in West Asia.

    ReplyDelete
  105. @Arza Previously, a Sicilian dated to 2500-1900 was described as Bronze Age, so I'd wager this is from before then.

    But it doesn't really matter, as you would have thought that a Bronze Age invasion of Central Europe would spread Bronze smelting tech with them during their migrations. If this Sicilian Steppe sample is Copper Age, it isn't likely (in fact, very unlikely) that their immediate ancestors had Bronze technology.

    ReplyDelete
  106. @Arza Basically, no Bronze = surely not Yamnaya in origin.

    ReplyDelete
  107. @An Idiot

    I can already tell you that the chances of L23/M269 being from West Asia, or any part of Asia, are zero.

    ReplyDelete
  108. @Davidski I guess we just have to wait and see. If that Caucasus paper has enough samples from across the entire Caucasus range during the Copper and Bronze Ages (as it seems to suggest) and it shows no sign of M269/L23, I'll almost certainly concede but be very confused (see my point above about why Steppe origin doesn't sit right in my head).

    It's only two weeks until we find out

    ReplyDelete
  109. @ An Idiot

    In Central Europe Bronze Age starts with Unetice Culture, well after BBC expansions:

    https://en.wikipedia.org/wiki/Bronze_Age#Central_Europe
    In Central Europe, the early Bronze Age Unetice culture (1800–1600 BC)

    ReplyDelete
  110. @Arza That's not true, the Bell Beaker culture in Central Europe was clearly part of the Bronze Age.

    ReplyDelete
  111. @ Anthony "If not from Siberia where do you suppose BOO got its "Nganassan/BHG" component from?"

    Of course. it comes from the east. However, we have so little relevant ancient DNA that the arrival of ENA to northeast Europe is not at all clear. We do not know when it happened, and we do not know if there are several different migrations via different routes.

    ReplyDelete
  112. I have found a very good fit for Sintashta_O3 in nMonte.

    Sintashta_MLBA_o3:Average
    Fit: 1.4939
    EHG 37.5%
    Shamanka N 0%
    Sintashta MLBA 10.83%
    West Siberia N 51.67%

    Modern VUR populations also take 5-10% EHG with the same setup, but that is still significantly less and the fits are much worse. Sintashta_MLBA_O3 looks nothing like anything else.
    Other notes:
    - Modern FU populations not only have Baikal-like ancestry that is missing in all the Sintashta outliers, but actually they all have more BHG than WSHG. (The fits are really bad though.)
    - Among the Sintashta outliers, only O3 reacts well to direct EHG ancestry. For O2 and O1 the above model does not work this well. Especially not for O1, but the fit of O2 also improves when EHG is swapped to Poltavka.

    ReplyDelete
  113. @Slumbery

    These G25/nMonte fits shouldn't be taken too literally, because they'll vary from one population to another.

    The reason for this is that some populations dominate some of the dimensions due to recent drift. But you'll still be able to get sensible models for them, especially if there are ancient samples available that are directly relevant to their ancestry.

    ReplyDelete
  114. @ Slumbery: some features of a new Finnish Iron Age sample (Tavastian, 1100-1200 AD) were just published in a local newspaper. N1c1-Z1927, "looks like Narva Culture Balt or other Latvian or Swedish HG's or neolithic Ukrainian". "Not like modern Saami", i.e. no BHG, I'd guess, but don't know about WSHG.

    There is a load of new samples on the way. Interesting times, really.

    ReplyDelete
  115. @Huck Finn
    I make that sentence more accurate: Modern FU populations _that have significant Siberian ancestry_ not only have Baikal-like ancestry that is missing in all the Sintashta outliers, but actually they all have more BHG than WSHG.

    Of course populations that have no or only low level Siberian ancestry do not count. I cannot say for example that Estonians have more BHG than WSHG, because they have pretty much noise level Siberian genome-wide ancestry (at least when we count only what is more recent than ANE into EHG) that does not even show up with my nMonte setups.

    @Davidsky
    Thanks for the input. Do I understand well, you are saying that (some) modern FU populations have a lot of relatively recent drift and that makes a mess of the fits when modelled with ancients?

    ReplyDelete
  116. @Slumbery

    Yes, some Uralic populations pull into their own spaces in a few dimensions, which worsens their fits no matter what you do.

    This is a problem only if you're unaware of it, and it can be advantage when looking for very specific signals of Uralic admixture.

    ReplyDelete
  117. JK1968/DA234 from Levänluhta appears on the G25 spreadsheets under both labels. If you could you run DA237?

    ReplyDelete

Read the rules before posting.

Comments by people with the nick "Unknown" are no longer allowed.

See also...


New rules for comments

Banned commentators list