Sunday, January 17, 2021

That old chestnut: Northeast vs Northwest Euros


In the last comment thread reader Greg put forth this question:

David, when are you going to explain the genetic discrepancy between Northeastern and Northwestern Europeans? You know, the one that people believe is due to Baltic Hunter-Gatherer admixture, whereas you believe it is due to genetic drift? You ought to make a post about this issue at some point, because a lot of people are wondering what's causing the differences.

Well, Greg, this issue has been discussed to the proverbial death here and elsewhere. In fact, there were two posts and rather lengthy comment threads on the same topic at this blog just a few months ago. See here and here.

Nevertheless, it seems that a fair number of people are still befuddled, so I'm going to try to explain this one last time, as briefly as a I can using just a handful of f4-stats.

Admittedly, Northeast Europeans generally do pack higher levels of indigenous European hunter-gatherer ancestry than Northwest Europeans. This is especially true of Balts, who show more of this type of ancestry than even Scandinavians in practically every type of analysis.

The f4-stats below back this up unambiguously. Note the significantly positive (>3) Z scores, which suggest that Latvians and Lithuanians harbor more Baltic hunter-gatherer-related ancestry than Norwegians and Swedes.

Chimp Baltic_HG Norwegian Latvian 0.001301 7.114
Chimp Baltic_HG Swedish Latvian 0.001017 4.205
Chimp Baltic_HG Norwegian Lithuanian 0.001023 7.341
Chimp Baltic_HG Swedish Lithuanian 0.000763 3.408

Greg, I know what you're thinking: the naysayers are right! But wait, because there's a twist to this tale. Check out these f4-stats:

Chimp Baltic_HG Norwegian Belarusian 0.000265 1.934
Chimp Baltic_HG Swedish Belarusian 0.000152 0.7
Chimp Baltic_HG Norwegian Polish 6.4E-05 0.519
Chimp Baltic_HG Swedish Polish -0.000235 -1.074

Please note, Greg, that none of the Z scores reach significance, which means that these Northwest Europeans and Slavs are symmetrically related to Baltic_HG. They're also symmetrically related to other relevant ancient groups such as the Yamnaya steppe herders. This, of course, suggests that they harbor very similar levels of basically the same ancient genetic components.

Chimp Karelia_HG Norwegian Belarusian 0.000136 0.844
Chimp Karelia_HG Swedish Belarusian 7.9E-05 0.32
Chimp Karelia_HG Norwegian Polish -4.7E-05 -0.304
Chimp Karelia_HG Swedish Polish -0.000134 -0.54

Chimp Yamnaya_Samara Norwegian Belarusian -0.000134 -1.085
Chimp Yamnaya_Samara Swedish Belarusian -6.6E-05 -0.34
Chimp Yamnaya_Samara Norwegian Polish -0.000225 -1.995
Chimp Yamnaya_Samara Swedish Polish -0.000311 -1.574

Chimp Barcin_N Norwegian Belarusian -0.000335 -2.809
Chimp Barcin_N Swedish Belarusian -0.000284 -1.491
Chimp Barcin_N Norwegian Polish -0.000222 -2.057
Chimp Barcin_N Swedish Polish -0.000318 -1.662

Chimp Baikal_N Norwegian Belarusian 0.000186 1.3
Chimp Baikal_N Swedish Belarusian -7E-05 -0.33
Chimp Baikal_N Norwegian Polish -4.6E-05 -0.351
Chimp Baikal_N Swedish Polish -0.000477 -2.277

Interestingly, pairing up Ukrainians with English samples from Cornwall and Kent produces similar outcomes. But that's because most ancient ancestry proportions in Europe show a closer correlation with latitude than longitude.

Chimp Baltic_HG English_Cornwall Ukrainian 0.000282 2.242
Chimp Baltic_HG English_Kent Ukrainian 0.000225 1.748

Chimp Karelia_HG English_Cornwall Ukrainian 0.000323 2.175
Chimp Karelia_HG English_Kent Ukrainian 0.000239 1.634

Chimp Yamnaya_Samara English_Cornwall Ukrainian -6.6E-05 -0.569
Chimp Yamnaya_Samara English_Kent Ukrainian -0.000112 -0.977

Chimp Barcin_N English_Cornwall Ukrainian -0.000519 -4.641
Chimp Barcin_N English_Kent Ukrainian -0.000598 -5.232

Chimp Baikal_N English_Cornwall Ukrainian 0.000385 2.874
Chimp Baikal_N English_Kent Ukrainian 0.00036 2.836

Now, Greg, if at least in terms of genetic ancestry, Latvians, Lithuanians, Belarusians, Poles and Ukrainians all qualify as Northeast Europeans, then what makes them different, as a group, from Northwest Europeans? Do you believe that the key factor is admixture from Baltic hunter-gatherers? Or is it genetic drift?

Of course, considering all of the f4-stats above, logic dictates that it must be relatively recent genetic drift.

Keep in mind, however, that this only applies to Balto-Slavic speaking Northeast Europeans without significant Uralian ancestry. Overall, Uralic speakers have a more complex population history, and indeed genetic differences between them and Northwest Europeans are in large part due to somewhat different ancestry proportions and also Siberian admixture.

See also...

So who's the most (indigenous) European of us all?

90 comments:

  1. Northwest Euros have more farmer ancestry.

    Farmers are cool though.

    ReplyDelete
  2. Rate model. The Baltic drift is merged into the Baltic HG component.
    https://i.imgur.com/oLB6C2S.png

    ReplyDelete
  3. No, it's not.

    Baltic_HG doesn't look like an extreme Balto-Slavic group in fine scale analyses. It just looks like WHG with a bit of EHG.

    On the other hand, Baltic_BA does show a lot of Balto-Slavic drift.

    ReplyDelete
  4. @Davidski

    Yes, of course. What I mean is, the reason why the fits are much better in the model above is because the Baltic_HG (a ghost that is similar to LTU_Narva, which is derived from Baltic_BA) carries the Balto-Slavic drift.

    ReplyDelete
  5. Fun little experiment. I subtracted the before mentioned Baltic HG component from the Polish average.

    Result is they now have a closer distance to Northwest Europeans (and Sintashta) than to East Europeans.

    Distance to: Polish_-12.5Baltic_HG
    0.03572305 Dutch
    0.04474442 Norwegian
    0.04680316 Swedish
    0.06299259 RUS_Sintashta_MLBA
    0.06877267 Russian_Orel
    0.07073025 Russian_Voronez
    0.07091532 Russian_Kursk
    0.07619776 Russian_Smolensk
    0.07753800 Russian_Tver
    0.08316285 Russian_Kostroma
    0.08341560 Finnish
    0.10466936 Latvian
    0.11024025 Russian_Pinega

    Distance to: Polish
    0.01717779 Russian_Voronez
    0.01832130 Russian_Orel
    0.02090743 Russian_Smolensk
    0.02260559 Russian_Kursk
    0.03131693 Russian_Tver
    0.04503245 Swedish
    0.04871370 Latvian
    0.05043032 Russian_Kostroma
    0.05396479 Finnish
    0.05676400 Dutch
    0.05891787 Norwegian
    0.07307935 RUS_Sintashta_MLBA
    0.08062288 Russian_Pinega

    ReplyDelete
  6. The models like crm posts seem broadly plausible to me and OK. Albeit it is slightly an abstraction as can't really say when drift happened.

    Caveat I am unsure if the population of GAC is really so important for East Europe today as the only contributing populations or it is more like Chalcolithic Hungarian populations are also / more important, and then correspondingly extra HG is slightly higher by 3/4% to compensate. There could be more substructure to regional farmer populations contributing.

    (In more words, I think assuming a single admixing farmer population is probably a simplification for purposes of tractable models and more farmers from across Europe probably survived and contributed to subsequent populations. Their substructive probably contributes at least slightly to differentiation today. But this is relatively minor. Also would not affect his model's relative contributions of the Steppe_EBA component either).

    Re; drift, I do think that I would add

    1) it doesn't seem like the East European samples in general are overall further from particular outgroups relative to the prediction from proportions, more so than West Europeans. So it's looks more like a slightly separated genetic drift path than a *longer* genetic drift path or "extra" genetic drift

    (and accordingly if a West European and East European sample with broadly similar ancestry proportions seem to have different distances in G25 from an ancient population, like an MN Farmer population or whatever, I think that's probably not attributable to different levels of overall genetic drift. Save for Lithuanians / Latvians.)

    2) it looks like the major bit of genetic drift in East Europe which seems to define difference in models peaks in Latvia_BA, which is also a peak of HG admixture, then seems to decline overall in subsequent populations after that. whereas genetic drift in Ireland/Britains, seems to accumulate more over time.

    so that suggests to me that a large drift happened either in the HG population admixing into the immediate ancestor of Latvia BA, or in that immediate ancestor itself, rather than in fits and starts of steadily over time. then diluted as this population mixed back into East European mainstream. That wouldn't be surprising as small HG populations are exactly where you'd expect to find large founder effects specific to them. (if something like the Baltic_HG model in crm's model didn't exist, then at the moment the reality is in modelling terms seems indistinguishable from if it did.)

    ReplyDelete
  7. So to recapitulate then, why are Balts in particular asymmetrically related to Baltic_HGs compared to NW Euros but Slavs not? Say indeed the Balto-Slavic component originated with basically the same components as NW Euros but then experienced significant drift, are Balts then something like drifted Balto-Slavic base + additional Baltic HG?

    ReplyDelete
  8. Did you use transversions only for this?

    ReplyDelete
  9. Interesting to see (what I interpret as) signs of Baltic ancestry in Swedes (at least over Norwegians) in the 1st, 2nd, and 4th set of F-stats here.

    I think this is also hinted at if you carefully inspect the Vikings paper, especially the supplementary.

    You also see it pretty clearly if you're lucky enough to have access to genotypes for lots of present-day Scandinavians and Baltic people.

    I think that pretty recent, Common Era population movements have impacted the genetic structure of many places in Europe to a greater extent than has been appreciated to date. The Vikings, Spain, and Rome papers all point to this.

    ReplyDelete
  10. One caveat to my previous comment; I mentioned a lot of stuff about drift in my previous comment and how it's relates to patterns in by G25, one word of caution on this is that PCA can tend to latch hold to the strongest patterns in a dataset, so depends on the composition of the dataset, and although this is generally going to be close to real differentiation, it could have some differences which show up in fine structure.

    For an illustration, I did some cross-plotting of G25 distance squared against Fst, to look at how closely they relate to fine structure in West Eurasia:

    https://imgur.com/a/bCn0sSv

    (Squared scaled PCA distances generally correlate closely to Fst. There is a recent paper by Florian Prive that mentions this).

    The correlation is really pretty good, R^2 about 0.8, which is pretty high.

    But at the same time, there is a difference where some pairs are above the line, showing more distance in G25 than expected from Fst, and some are below the line, being closer in G25 than Fst.

    Thre are some patterns to this that are interesting, but hard to summarize.

    This is not to say that the Fst estimator distance is necessarily a best measure of distance and anything is wrong with G25 (still the best single PCA for global variation), just for caution about interpreting way too strictly and there is still some difficulties with capturing very fine structure.

    ReplyDelete
  11. Yeah, I just realized there is excess WHG in Northwest Europe too. Making the difference in ancient ancestry between East/West Slavs and Norse small.

    Scandinavians get excess WHG from late Neolithic farmers in Northwest Europe similar to Blatterhole_MN.

    Bell Beaker Britain also had this excess WHG but it is watered down in modern British Isles due to Celtic ancestry.

    British Isles have definite difference in ancient ancestry from East/West Slavs. More farmer and less hunter gatherer.

    ReplyDelete
  12. Geographically...Northern France, Northern Germany, low countries are also in Northwest Europe.

    But North France, West/South German, and Low countries have significant difference in ancient ancestry from Northeast Europe.

    They are comparable to Hungary, Slovenia, Croatia which are much more south than them geographically.

    ReplyDelete
  13. @Greg,

    If you think about it, there are two main reasons for the North to South difference in Europe.

    1: Anatolian farmers in Early Neolithic did not settle East Europe much farther north than Hungary. This is ultimately why there Northeast Europe has less farmer ancestry.

    2: Steppe migrations mainly came from Corded Ware. So they mainly went from north to south. This is why Northern Europe has higher Steppe ancestry.

    Modern East/West Slavs don't have a lot of hunter gatherer compared to what existed in the region in Bronze and Iron age. There used to be more excess hunter gatherer populations in Poland, Hungary, Ukraine. But it was not universal, seems to have varied a lot.

    Welzin BA in Germany, Bronze and Iron age samples from Hungary have lots of hunter gatherer. Their DNA really demonstrates that they impart descended from the last hunter gatherers of mainland Europe.

    It wasn't just in East Baltic that late hunter gatherer ancestry survived.

    There were hunter gatherer populations in East Baltic, Poland/Ukraine who "survived." The ones in Northern Russia who were wiped out by Fatyanovo, the ones in Scandinavia mainly disappeared but contributed a little to Saami.

    ReplyDelete
  14. One thing seems certain: the currently observed gradient of the Balto-Slavic drift has nothing to do with medieval Slavic migrations (or has only little to do).

    ReplyDelete
  15. @Davidski

    Thank you for the explanation! This has been long overdue, and I'm going to be linking to this post whenever someone brings this issue up in the future. When do you estimate this recent genetic drift to have occurred?


    @Genos Historia

    I took a look at your West Eurasia index2, and Dutch people don't seem to differ from other Northwest European ethnicities in terms of admixture proportions. According to your list, Englishmen seem to have more Steppe admixture than I imagined at 43.7% "Yamnaya", although I don't know what you're using as a proxy for the WSH ancestry. I do agree with you in regards to Northern Frenchmen, Belgians and South Germans though. Where do you draw the line between Northern and Southern Europeans?

    I am aware of the genetic history of Northern Europe in general; the only "hangup" that I've been having of late is in regards to Northeastern Europeans not clustering as tightly with Northwestern Europeans as they should given their practically identical autosomal proportions, although now I obviously understand that this topic is considerably more nuanced than I had previously thought. As far as I'm concerned, there are two primary "clusters" within the general European macrocluster: one firmly Northern European and one firmly Southern European, glued together by autosomal intermediaries such as the French and the former Yugoslavs to form a gradient of sorts.

    ReplyDelete
  16. I think this process started within a specific Corded Ware population in Eastern Europe, and was very rapid, so that it already hit its peak in the Bronze Age.

    Since then it's been diluted considerably, even in the Baltic states.

    Early Slavic samples - and I mean real Slavs and not samples like the Medieval Germans that we analyzed here recently - will actually look very Baltic and more eastern than Poles, Belorussians, Ukrainians and even most Russians.

    So it'll be interesting to see when the West and East Slavs acquired their western ancestry and from which populations exactly.

    ReplyDelete
  17. @Davidski
    I think your story about Slavs is very confusing. We have to start from the very beginning to try to untangle it.
    We know that R1a-M417 was not very popular and dominating on the steppe, so it was not PIE. Migration of R1a to Poland where CWC originated was not PIE migration but the migration of proto-Indo-Slavs, one of many IE groups on the steppe. Then Z93 left and went east. What R1a CWC remained was proto-Balto-Slavs. Balts went north, mixed with HG, drifted. What remained was proto-Slavs who mixed with remaining EEF-HG tribes. Then there was that Baltic and Slavic melting pot from which Unetice and Trzciniec cultures emerged. Western Unetice was influenced more by BB so their language is uncertain, but from Nitra, Eastern Unetice and Trzciniec all sorts of Slavic and Baltic tribes originated, diverged, converged etc.
    You cannot say anything specific about the origin of Slavs because as linguists say : “we can hardly date the ‘emergence’ or ‘separation’ of proto-Slavonic or proto-Slavonic dialects from Indo-European dialects because of the proper uninterrupted Indo-European origin of Slavonic.”
    Now that R1b are clearly linked with Italo-Celto-Germanic languages we can separate their history from our.

    ReplyDelete
  18. @EastPole

    "Migration of R1a to Poland where CWC originated was not PIE migration but the migration of proto-Indo-Slavs, one of many IE groups on the steppe."

    It doesn't seem plausible to me that there were many different IE groups on the EMBA steppe, so to say the different later IE ethnicities in an embryonic state. It makes more sense to think that ethnic/linguistic differentiation arose as a consequence of geographic distance and separation, and therefore mostly after migration.

    ReplyDelete
  19. @EastPole

    Nitra, Eastern Unetice and Trzciniec

    And Hatvan/Otomani-Fuzesabony, Gava, Vekerzug or even La Tene:

    source - all modern averages + Yamnaya, Iron Gates and Barcin to capture things that are way outside of modern diversity

    Target: HUN_Vekerzug_IA:I20746
    Distance: 2.4279% / 0.02427860 | R3P
    65.2 Slovakian
    21.8 Basque_Spanish
    13.0 TUR_Barcin_N

    Target: HUN_Vekerzug_IA:I20745 (Y-DNA I1)
    Distance: 2.5237% / 0.02523745 | R3P
    88.4 Slovakian
    8.2 SRB_Iron_Gates_HG
    3.4 Yamnaya_RUS_Samara

    Target: HUN_Vekerzug_IA:I20743
    Distance: 2.3913% / 0.02391327 | R3P
    62.4 Bosnian
    21.2 Basque_Spanish
    16.4 SRB_Iron_Gates_HG

    Target: HUN_LaTene_IA:I20774
    Distance: 2.0560% / 0.02055971 | R3P
    65.0 German_East
    20.4 Basque_French
    14.6 Lithuanian_PZ

    Target: HUN_LaTene_IA:I20774 (without German_East)
    Distance: 2.0866% / 0.02086649 | R3P
    44.4 Russian_Smolensk
    30.8 Orcadian
    24.8 Basque_French

    Target: HUN_Hatvan-Fuzesabony_BA:I20772 (Y-DNA R1a)
    Distance: 1.8706% / 0.01870620 | R3P
    53.4 French_Pas-de-Calais
    39.4 Latvian
    7.2 Yamnaya_RUS_Samara

    Target: HUN_Gava_BA:I20771
    Distance: 3.1854% / 0.03185423 | R3P
    63.6 Slovakian
    28.2 Spanish_La_Rioja
    8.2 SRB_Iron_Gates_HG

    G25 coordinates:
    https://drive.google.com/file/d/1H8JsPW3JyOk4rBZz7mcl5w1eQ0ox1qGS/view

    ReplyDelete
  20. David, if the drift had happened in the CWC in Eastern Europe, we would have expected to see him in the Steppe MLBA.

    As I understand it, early Slavs are supposed to look like Lithuanians or Latvians, and maybe even Bronze Age Balts. Where and from what period do these samples come from?

    If we do not genetically distinguish the early Slavs from the Balts, we can only infer about the Slavic homeland on the basis of linguistic data (toponymy).

    ReplyDelete
  21. Notwithstanding above caveat, for a visualization of the differences in real present European day European distances from Steppe and MN farmer, controlled for distance to Africa (overall drift), against simulations using Europe_EN (Koros), Europe_MN (GAC_Poland, Iberia_CA), Steppe_EMBA (CWC_Baltic_Early and Yamnaya_KAZ) and Iron_Gates_HG - https://imgur.com/a/pB43NzJ

    Uses DistanceAfricaComposite-DistanceX as a control for effects of drift on G25 population distance. Basically a similarity index where higher = more similar. Blue dots are a real populations, black dots are simulations, red dots are midpoint between them.

    The closeness of the blue dot to the black dot indicates how well the model predicts this measure. If they overlap perfectly then the model is predicting that measure very well, while if the blue dot is higher, then the real population should be closer to the target population (net of drift) than predicted by the model, or if its lower then the real population should be further to the target (net of drift)than predicted by the model.

    Simulations using those populations form a linear cline, but the real populations deviate away from that by having a relatively higher distance to Steppe/GAC (as measured by distance from Africa to control for overall drift).

    The populations which are most outlying the simulation are obviously those ones with contribution from East/Central Asia or transmediterranean (which would make them deviate from the model that has no such components). But there is also some small difference for Lithuanians/Latvians.

    (Another look at proportions from the simple 6 way model above: https://imgur.com/a/Xkjmrs6. Populations are tending to draw on Yamnaya_KAZ rather than CWC_Baltic_early only when they "want" more CHG than the model is able to provide. It's clearly an inadequate model for most populations, only used it as seemed like a quick enough way to get simulations out)

    (This is assuming that this is a halfway sensible way of controlling for drift in populations; it may not be. I'm just trying it out.)

    ReplyDelete
  22. Actually scratch that last comment about using distance from Africa to control for drift. It's not right. It's actually the difference between the squared distances that controls properly...

    ReplyDelete
  23. @Greg,
    "Where do you draw the line between Northern and Southern Europeans?"

    I would include Northern Germany, Northern Dutch, British Isles, Scandinavia as in Northwest European category. Davidski used to call this 'North Sea' before ancient DNA.

    But I wouldn't include West/South Germany, South Dutch, and North French. Even though geographically they are in Northwest Europe. But genetically they pull too south.

    The foundation of Northwest Europe is Bronze age Britain/Netherlands. But really, just Northwestern Corded Ware or maybe more specifically Single Grave culture is the foundation.

    Bronze age Britain, Iron age Scandinavia, Early German tribes are all very similar to each other. France/Celts pull are the same as them but have extra farmer ancestry.

    Most modern Northwest Europeans pull south from them and this is because of French-like/Celtic admixture throughout history.

    ReplyDelete
  24. @Greg,

    Northwest Europeans come from Single Grave (also Battle Axe) variant of Corded Ware.

    Northeast Europeans (Balto-Slavs) come Eastern Corded Ware groups we don't know yet.

    Southwest Europeans are Single Grave plus lots of southern Farmer ancestry.

    Corded Ware (Indo Europeans) is the grandaddy who made the last change in Europe's population genetics, and established the foundation for modern variation.

    Northern Europe has more Steppe because Corded Ware first went into Northern Europe then moved Southwest, but Spain then eventually Central Italy were also impacted.

    ReplyDelete
  25. @Greg,

    Since Bronze age there has been a continuous cline connecting all of Western Europe from North to South. There hasn't been mutually exclusive North to South clusters. France and Southern Germany have always been intermediates.

    But in Northeast Europe, I think there may have been huge distinction between North and South. With no cline.

    Italian/Croatia BA like populations south of Poland. And Welzin BA and Slavic like populations north of Hungary.

    This distinction would be due to the combination of Corded Ware never moving into Southeast Europe and Anatolian farmers having a weak presence in Northeast Europe.

    ReplyDelete
  26. @Davidski

    "Early Slavic samples - and I mean real Slavs"
    What about Bohemian slavs?
    What time period and what lands will satisfy you?

    ReplyDelete
  27. @Simon_W

    “It doesn't seem plausible to me that there were many different IE groups on the EMBA steppe, so to say the different later IE ethnicities in an embryonic state. It makes more sense to think that ethnic/linguistic differentiation arose as a consequence of geographic distance and separation, and therefore mostly after migration.”

    Steppe is huge. Geographic distances are large. Steppe was very poorly populated before Yamnaya. Small groups living close to rivers. More probable that PIE split into different IE languages on the steppe than in Poland.
    R1a and R1b migrated separately from the steppe. They didn’t mix because there are no R1a and R1b mixed tribes. Most likely genetic barrier was due to cultural differences: different languages, religions etc, and that barrier was already present on the steppe before 3000 BC. No R1a in Yamnaya.

    ReplyDelete
  28. @Draft Dozen

    One of the Bohemian Slavs is obviously mixed. The other one isn't, but the genome is of poor quality.

    So we need a new sequence of that one.

    ReplyDelete
  29. @EastPole

    We know that R1a-M417 was not very popular and dominating on the steppe, so it was not PIE.

    This statement doesn't make any sense.

    Just one R1a-M417 may have been present in the PIE community on the steppe.

    And there was at least one, otherwise we wouldn't see the massive founder effect in R1a-M417 that obviously started in the Proto-CWC population.

    ReplyDelete
  30. @ambron

    David, if the drift had happened in the CWC in Eastern Europe, we would have expected to see him in the Steppe MLBA.

    But Steppe MLBA derives from Fatyanovo, and Fatyanovo existed already around 2,600 BCE.

    ReplyDelete
  31. @EastPole

    " Steppe was very poorly populated before Yamnaya. "

    It is nonsense.

    " More probable that PIE split into different IE languages on the steppe "

    It is nonsense.

    Davidski said...
    "But Steppe MLBA derives from Fatyanovo"

    It is not fact.

    ReplyDelete
  32. @Davidski

    “Just one R1a-M417 may have been present in the PIE community on the steppe.

    And there was at least one, otherwise we wouldn't see the massive founder effect in R1a-M417 that obviously started in the Proto-CWC population.”

    I used to believe that we will find a R1a-M417 dominated powerful advanced culture on the steppe the expansion of which will explain the expansion of IE languages and cultures. You told us that it is not the case. So I am puzzled. It is hard to believe that it all started from the small group of poor steppe refugees.

    ReplyDelete
  33. It's a fact that Steppe MLBA derives from Fatyanovo.

    And there's Z94 in Fatyanovo, as well as Z2124.

    ReplyDelete
  34. "They didn’t mix because there are no R1a and R1b mixed tribes."

    I always thought its consensus meanwhile that Indoeuropeans had been a concept of patrilinial identity.(All male tribal members originate in the founder of that tribe. While females are taken from outside the tribe, even from completely different cultures). Wich seems not exclusively IE, but maybe nomadic (The Old Testament of the bible describes that exact behavior for semitic tribes)

    If that would have been the case, then of course, two IE tribes wouldnt exchange males. Only females.

    ReplyDelete
  35. @Archi
    Before wheel and wagon steppe was not densly populated. It is a fact.

    ReplyDelete
  36. @EastPole

    I guess you missed the part where I said that Usatovo was R1a.

    ReplyDelete
  37. One of the Bohemian Slavs is obviously mixed. The other one isn't

    Yeah, while RISE568 is heavily Baltic-shifted, RISE569 plots where one would expect the Proto-Slavs:

    https://i.postimg.cc/3x9k4vmR/bohemian-Proto-Slav.png

    The South Slavic cline doesn't lie ¯\_(ツ)_/¯

    ReplyDelete
  38. @Davidski
    "It's a fact that Steppe MLBA derives from Fatyanovo."

    All samples known to this day had Z94-. "Poltavka Outlier" has Z94 +, but in fact Volsk-Lbische of the same time.

    So Fatyanovo could only be the Sintashta brothers. Literally everything indicates this, Sintashta is more connected with the Czech Republic than with Fatyanovo.

    EastPole said...
    "Before wheel and wagon steppe was not densly populated. It is a fact."

    This is a lie. These are your inventions of an ignorant person and this is a fact.

    ReplyDelete
  39. @Arza

    South Slavs also have western ancestry. This is especially true of Slovenians and Croatians.

    ReplyDelete
  40. @Archi

    What do you want to bet the there's Z94 and Z2124 in Fatyanovo?

    ReplyDelete
  41. David, I see you've succumbed to Michał's argument that RISE569 is the daughter of RISE568 with some German.

    Arza, thanks for the models and PCA. Cool thing! Is the La Tene sample already the L1029 or not this one yet?

    ReplyDelete
  42. So how do you explain the difference between the two samples?

    One of them has to be mixed.

    I'm guessing you think RISE568 is the daughter of some Balt? Or maybe two Balts?

    ReplyDelete
  43. So, you guys not argue. She is my daughter, alright?

    —-
    Speaking of Balto-Slavic ancestry. Do we have any genes from Rzucewo culture? According to wiki (yeah I know) they were a mix of Narva, GAC and early CW. Which to me sounds like same components at least in big lines that Balts are made of.

    ReplyDelete
  44. Further to my above posts, following might be interesting to anyone looking at PCA methods and the questions of what genetic drift does to euclidean PCA distances and methods of finding ancestors by minimizing the Euclidean distance. (Vahaduo and other folks)

    One of the things that comes up in this topic is whether extra genetic drift in some branch drives populations not to choose a real ancestor population. So, you know, a population X is actually descended from population Y + population Z, but extra genetic drift in X somehow leads method's like Vahaduo admixture to infer the wrong ancestor.

    I don't this is the case. For one reason why, one way to test this idea, is to simulate the process of extra genetic drift. I'll do this by adding an extra fake dimension to Global 25, which represents and introduces extra drift into copies of real populations, then see if this has any effect on the inferred ancestors Vahaduo assigns from the same source, relative to the real populations.

    E.g. see here: https://imgur.com/a/cZpQmIZ

    First I added on an extra fake dimension (called PC26-FAKE) with values of 0.01 for only 3x population averages and 0 for every other population, simulating an imagined process where those 3x had all gone through an extra bottleneck that provided further genetic drift (these copes have FAKE after their name).

    Then I run a Vahaduo admixture model on the resulting European dataset.

    You can see that the extra drift dimension does not change the inferred source ancestor proportions. There is some very minor difference in inference, but really it just adds to the Euclidean distance of the populations with an extra fake drift dimension and produces exactly the same fit.

    This makes a lot of sense, in that we would not expect this wholly orthogonal extra drift which only happens in the target to favour *any* of the sources more or less. To actually infer different ancestors, we would have to be following a *different* drift path as captured by G25, not just have extra drift. (This is probably also the case where you have some extra drift from a population that is not included in the sources that is a pure outgroup to the sources you have included - it doesn't favour any of them.)

    Another question might be what this extra genetic drift dimension does to the relative distance to outgroup. E.g. what it does to the (Y:Africa)-(Y:X).

    Here's a visualisation of that: https://imgur.com/a/xUBjjxT

    Basically the raw Euclidean distance to Africa is not as affected by extra drift as the distance to close populations, and this means that extra drift makes (Y:Africa)-(Y:X) fall! Thus the earlier comment where I said that (Y:Africa)-(Y:X) probably provides a measure of distance net of drift probably fails.

    But (Y:Africa)^2-(Y:X)^2 seems to be totally unaffected by the extra drift at all! It basically works like an outgroup f3 stat.

    Anyway tl;dr:

    1) Adding extra dimensions of genetic drift probably does not cause Vahaduo admixture to select the wrong populations as source/ancestors (or at any rate different ancestors).

    2) Difference between squared distances probably approximates f3 stats, and when using an outgroup appears to be unaffected by extra genetic drift.

    The squared distance difference might be useful method to have in any calculator, as it apparently approximates f3 stat.

    ReplyDelete
  45. How do you explain the difference between WEZ56 and WEZ40?

    I don't take RISE568 seriously because it has little coverage, no radiocarbon dating, and molecular dating separates her 2,000 years from RISE569.

    I don't think RISE569 is a mix, as we already have some similar genomes from pre-medieval Central Europe.

    ReplyDelete
  46. The Tollense Valley is near the Baltic, and obviously there were some foreign soldiers there.

    RISE568 has enough data to be run in many analyses and is indeed very Baltic-like.

    And this sample is also very similar to early Slavs, as you shall see.

    ReplyDelete
  47. @Genos Historia

    "But I wouldn't include West/South Germany, South Dutch, and North French. Even though geographically they are in Northwest Europe. But genetically they pull too south."

    And where do you draw the line between North and South in Eastern Europe?

    "Most modern Northwest Europeans pull south from them and this is because of French-like/Celtic admixture throughout history."

    Do you mean Celtic as in post-Roman Gaul, or that the continental Celts themselves were southern-shifted compared to the nearby Germanic tribes?

    ReplyDelete
  48. @Arza
    “Nitra, Eastern Unetice and Trzciniec

    And Hatvan/Otomani-Fuzesabony, Gava, Vekerzug or even La Tene”

    I don’t know whether this ‘drift’ is real or only noise. In higher Principal Components dimensions which explain low percentage of variation, as is the case here, all sorts of strange effects happen. If it is real then probably Balto-Slavic CWC, which was originally the same as Indo-Slavic CWC (red circle), mixed with some HG around Baltic, in Poland and Hungary because there are many ‘drifted’ samples there:


    https://postimg.cc/WFL0GrPc

    As you know in Poland and Hungary there was the boundary between EEF and HG’s and all sorts of EEF-HG mixes existed.

    ReplyDelete
  49. I think, put simplistically, slavs are Balto-Slavs + a carpian/ Dacian component, then additional regional substrates which were distinctive (eg Germani -Lombards in Slovakia; Romano-Illyrians in central Balkans, etc)

    The C-D component is relatively easy to understand
    The B-Sl component is more arguable.lately I’ve been wondering if that population came from northeast Poland. Eg warriors from Oltzyn etc

    ReplyDelete
  50. @Greg,

    Yes, Celts were significantly more "southern" than early Germans. We can already see this in DNA from Lombards, Saxons compared to Iron age France.

    But, Celts were closer to Early Germans than to Bronze age Spain and Iron age Italy. So they were more 'northern' than 'southern', but really could not fit in any category, and can only be said to be an intermediate.

    Celtic admixture made Britain and Ireland more "southern" than they were in the Bronze age. If not for Celtic admixture, British isles would cluster with Scandinavia.

    Lots of Celtic, and also French Medieval, admixture makes most Germans more southern than British Isles and Scandinavia.

    But there isn't really distinct categories of Southwest and Northwest European. There's really just clines. It has been this way since the Early Bronze age.

    Sweden>Norway>British Isles/North Germany>South Germany>North France>Southwest France>Spain

    ReplyDelete
  51. @Greg,

    In East Europe, in bronze age, there don't seem to have been cline connecting north and south. So, the distinction between north and south would be more real than in West Europe. This is due to different patterns of Corded Ware and Anatolian farmer migrations.

    North of Hungary would be Northern, South of Poland would be Southern.

    Or just look at a map of Corded Ware to find the line, anything south of Corded Ware in Eastern Europe would be Southern.

    But the Slavic migrations I think blurred this line and created a cline connecting north of south.

    However, in east east Europe the line between north and south is still significant. As there is no real cline connecting Ukraine and Romania. This distinction although watered down ultimately goes back to Corded Ware, Anatolian farmers which is interesting.

    ReplyDelete
  52. @Rob,

    In PCA, East/West Slavs are intermediate between South Slavs and Balts. It make sense this means early Slavs had Balkan admixture. That would be one way to lower hunter gatherer ancestry.

    ReplyDelete
  53. @ Genos

    It depends what we mean by Balkan, of course. Lets remember that the pre-Slavic northern Balkans had Germanic foederatii. Will Balkan Roman towns by like Rome (large Aegean immigration), or will a very different pattern of immigration be present.?

    Even though the basic genomic pattern seems evident (i.e. a cline from southern Greece to modern E Baltic), that is the ''net result'', so to speak. But under that belies a series of criss-crossing movements and social networks which perspectives like this might not do justice.

    ReplyDelete
  54. David:

    "I have a strong feeling that the earliest Slavs, and indeed Proto-Slavs, will cluster with those western-most (right-shifted on the plot) Baltic BA samples".

    "And I think the Proto-Slavic cluster will be in the part of the plot where Poles overlap with Balts".

    The same is apparent in the paper Busby - the Slavic homeland on Suvalkija. However, I see two problems here:
    1. Such migration should spread the N1c which is lacking among Western and Southern Slavs.
    2. The toponymy of these areas is Baltic.
    Therefore, I believe that the Baltic genomes of the early West Slavs are a local artifact after populations represented by individuals such as WEZ56, N47 and RISE568 (molecular dating moves this sample to the Bronze Age).

    ReplyDelete
  55. David, I don't think WEZ56 has come from very far. See how some Kashubians are getting closer to WEZ56:

    http://www.historycy.org/index.php?showtopic=164942&st=720

    ReplyDelete
  56. @ambron

    "The same is apparent in the paper Busby - the Slavic homeland on Suvalkija. However, I see two problems here:
    1. Such migration should spread the N1c which is lacking among Western and Southern Slavs.
    2. The toponymy of these areas is Baltic."

    Regarding the N1c, is it possible that it only arrived when the Eastern Balts/Lithuanians fairly recently migrated westwards, and that it didn't exist among the Western Balts and possible Slavs who inhabited the region earlier?

    That of course doesn't solve the issue with the toponymy, but how old are the toponyms, i.e. what kind of Baltic?

    ReplyDelete
  57. Angantyr, of course, we are talking about the oldest layer of toponymy.
    I even know the ideas that N1c in Pribaltica is a medieval founder effect. However, we must follow the principle of the economy of thinking.

    ReplyDelete
  58. @ambron
    "2. The toponymy of these areas is Baltic."
    Hydronyms also not slavic
    http://900igr.net/up/datai/130273/0027-019-.jpg

    ReplyDelete
  59. Draft, yeah, I know, thanks! Toponymy is the naming for places in general, so it also includes hydronymy.

    ReplyDelete
  60. Awesome analysis of the La Tene sample from Hungary made by ph2ter:

    https://anthrogenica.com/showthread.php?22778-New-ancient-samples-G25&p=739316#post739316

    This is how it looks like on a PCA:

    https://i.postimg.cc/WNnmb74D/HUN-La-Tene-IA.png

    ReplyDelete
  61. @ Arza

    What’s the upshot meant to be ? Sure , the La Tene sample seems mixed

    Unfortunately the follow-up comments by Cold Mountains & Copper-Axe are incorrect, but serve to illustrate the challenges in tessalating DNA, archaeology & history. There were no 'Scythians’ in Siberia or Mongolia, so they could not 'come from there', nor are the people at Glinoe Daco-Getans. They're as Scythian as Apple-Pie

    ReplyDelete
  62. @Rob,

    Do you think Scythians on this map, actually represent distinct groups who all spoke different Iranian languages?
    https://en.wikipedia.org/wiki/Scythians#/media/File:Scythia-Parthia_100_BC.png

    Are you saying actual Scythians only lived in East Europe? While, what historians call Scythians are just other Iranians?

    What do people base the concept of Scythian on? Lingustics proving common language, archeaology proving common culture, claims made by ancient Greeks on the extent of where Scythians live?

    It would be annoying if Scythian is a modern miss conception which never existing.

    ReplyDelete
  63. @ Genos

    Yes it seems like there’s a dizzying array of different groups; sometimes with vastly distinctive ancestry . But that’s because they come from different times and places ; and can in fact be rationalised into a narrative
    But yes- historical sources point to the early Scythians in the west. No source ever calls out Scythians in Siberia. Those inner Asian groups have so much inner Asian admixture that they might have been non-IE, or para-Tocharian
    The real Scythians are Europeans with post -Catacomb, post -Srubnaja & Hallstatt ancestry

    ReplyDelete
  64. @ Rob
    The upshot is that people with genomes virtually identical to modern Slavic ones lived in Central Europe well before the Middle Ages (200 BCE).

    You can repeat this analysis with MX265 (half-Polish, 600 BCE) or BR2 (2/3 Sorbian, 1200 BCE) and the result will be nearly the same. Not to mention I20745 which is almost fully Slavic-like.

    ReplyDelete
  65. Clearly people like Scy305 are either non-Iranic speaking locals (which Scythia was full off) or people assimilated into Scythian culture, and quite recently so by the looks of it since other samples from that dataset do seem to have quite a bit of ancestry from iron age steppes nomads who have a clear genetic and cultural link to the chaps who decided the lands north of the Altai were a fun place to catch a break, make some sick deer art and break horses.

    Kinda funny that samples from Pazyryk or Uyuk with 35-40% Siberian ancestry are so admixed that they might as well be non-IE but a "Scythian" who derives the vast majority of their ancestry from southeastern Europe is as Scythian as apple pie...

    Either way they are not genetically representative of Scythians. It doesnt matter if you define Scythian as the name for a cultural horizon of presumeably Iranic speaking steppe nomads or just those specific to the Pontic region during the classical period, so my point still stands. Its silly to characterize such ancestry as simply being Scythian in my opinion. Especially if we are purely talking from a genetic standpoint.

    Also Herodotus states that the Scythians came out of Asia following the Cimmerians, plus they are first mentioned in Assyrian sources after having come down from the Caucasus. That isnt the west in my book.

    Scythian is also used by historians such as Strabo to refer to the Central Asian steppe nomads as well as the steppe invaders of Greco-Bactria who in all likelihood were the Yuezhi and historically they were some of the most eastern nomads we know of.

    In similar fashion, the classical scythians are referred to as Saka paradraya by the Persians.

    ReplyDelete
  66. @ Copper Axe


    ''Kinda funny that samples from Pazyryk or Uyuk with 35-40% Siberian ancestry are so admixed that they might as well be non-IE but a "Scythian" who derives the vast majority of their ancestry from southeastern Europe is as Scythian as apple pie...''

    As I suggested, the problem is a problematic back-projection using 'mixed-argumentation'. Nothing unique to your perspective, it is deeply embedded in the zeitgeist of modern archaeologists who have been misled by Herodotus.
    Please highlight any contemporary sources which confer a Scythian identity on Pazyryk people. In fact, Pazyryk are late ''Scythians'' , even verging on Slab Grave ancestry.


    ''Either way they are not genetically representative of Scythians''

    They are entirey representative of Scythians, because that is where Scythians were. We have contemporary sources, coins issued in their name, inscriptions - these are the Scythians; not the ones from the Altai, who were never called 'Scythians' except by modern archaeologists.
    Moreover, your claim that Glinoe are Dacians is archaeologically unfounded - they have catacomb burials under elaborate barrows + horse gear. Getans and Thracians have a completely distinct material culture, even though they share ancestry with the late Scythians. So these societies were open to interacting, but clearly communicated their distinction by the material culture they used.


    ''Also Herodotus states that the Scythians came out of Asia following the Cimmerians, plus they are first mentioned in Assyrian''

    Herodotus is wrong. It is the Cimmerians who are Inner Asian, if we define 'Cimmerians'' as the earliest Iron Age steppe nomads in the west

    The earliest historical references to Scythians, and Cimmerians, comes from Near Eastern sources, not Herodotus. They clearly illustrate that Sc & Cimm. raided northwestern Iran & eastern Anatolia. This locates them in the western steppe.


    ''from the Caucasus. That isnt the west in my book.''

    LOL don't be silly. Its exactly the western steppe


    '' Strabo '' wrote at the turn of the Common Era.

    The crux of the issue is, early Scythians were influenced culturally by Inner Asian groups, and will show Inner Asian ancestry. But the Scythians formed in the western steppe, and that's where they always were, as late as 250 AD. We have no idea what those Altai-Siberian nonads were. Without any clear evidence for them to have ever been called Scythians, such claims are purely imaginary

    ReplyDelete
  67. Arza, fact, these decomposition analyzes are very interesting. Have you seen the Krakauer Berg analyzes?

    And it is obvious - the West Slavic genome, similar to RISE569, was in Central Europe long before the Middle Ages. It proves a high degree of biological continuity in the region.

    MX265 - 50% Pole in autosomes, 100% Pole in uniparentals.

    ReplyDelete
  68. Rob is writing sheer nonsense as always. The Scythians lived like a solid in Central Asia, Siberia, Europe, and they were the same Scythians. Herodotus called the Central Asian Saks Scythians. And the Persians called the Black Sea Scythians Sacae. All these are just random names established in their traditions, because the Black Sea Scythians did not call themselves Scythians, they called themselves Skolots. The fact that the Scythians from the east of Central Asia and it cannot have two opinions here, it is just that we have a Greek tradition to call them Scythians, if the Persian tradition would call all the Sacae, and if they borrow the name from the Black Sea Scythians, then we would call them Skolots.

    Glinnoe is not even a purely archaeological Scythian burial in a categorical sense, everyone who is buried there is not Scythians. This is just a scientific fact. About the Glinnoe burial ground, we can categorically assert that this is the Getae, and we can also categorically assert that it is not Scythian, it does not have anything in general with one Scythian burial ground, this is a scientific fact, to assert the opposite is to engage in anti-science.

    Herodotus is not wrong, only Rob is wrong. We know for sure that the Scythians lived in southern Siberia beyond any scientific doubt.

    ReplyDelete
  69. This is a Getan burial complex, - in the lower Danube region, where they were located

    https://imgur.com/undefined


    ''The Scythians lived like a solid in Central Asia, Siberia, Europe, and they were the same Scythians. Herodotus called the Central Asian Saks Scythians. And the Persians called the Black Sea Scythians Sacae. All these are just random names established in their tradition''

    They lived like a solid ? Are are they illegally circimcising Etruscans ? You dont make sense
    There's no randomness, there is a progression of historiographic tradition from west to east. The Sakae appear in 4-500s BCE. The Saka kurgans in Turan appear ~ 450 BC.

    ReplyDelete
  70. @Rob,

    Asian ancestry in European Iranians is good evidence Iranian languages originated in Asia. I do not know about the history but this makes sense to me.

    All Scythians have Asian admix. Except for ones who are 100% from local East-Central European such as Polish like one in Ukraine and Italian like one in Hungary. There are no Scythians who are complete continuation of Srubanaya.

    Their Asian admix includes BMAC ancestry. Makes me suspect Iranian languages originated near Tajikstan.

    ReplyDelete
  71. @ Genos

    Iranian languages probaby originate in Asia, although I don't have a strong opinion on that because I haven't afforded too much detailed thought to it. But that's beside the point Im making - which is about identity & ethnogenesis. This is a very different issue.

    There are key places like Karasuk culture which were formative in the Scythian culture, the deer stones, so forth. But they are no earlier than the Scythian cultures in the west (Chernogorovka culture, which has been dated to 10000 BC). In any case, all the horse culture and bronze technology ultimately derives from the Western steppe.
    Of course, I am also aware that Scythians (even some western ones) have Inner Asian +/- BMAC ancestry. But this was neither a mass migration nor elite conquest - it was more a back-diffusion.
    The presence of such admixture does not translate into Scythians 'came from Inner Asia ' or Iran. This is a false construct due to simplistic reasoning. Rather, Scythians formed in the western steppe. That is where they are first attested, this is where the elaborate early kurgans with Near Eastern plunder are located, and that's where they continued to be attested attested for ~ 1000 years. It is only during Darius' reign that the Saka-Scythian ethonym was transferred to Central Asian tribes. What was Darius' doing at the time ? Campaigning against Greece, Macedonia and the Black Sea (e.g. real) Scythians. That is where he learned about things Scythian (although his predecessors - the Medes also were aware of them, during their campaigns in the northwest)

    ReplyDelete
  72. @Rob
    "But they are no earlier than the Scythian cultures in the west (Chernogorovka culture, which has been dated to 10000 BC). In any case, all the horse culture and bronze technology ultimately derives from the Western steppe."
    ....

    Every word that you write is not true, it is not 10000BC, it is from the 9th century BC, and no one associates it with the Scythians, at best with the Cimmerians. Equestrian culture there is clearly from the east, there are no two opinions.
    You wrote a complete deception about Tuva, are you still inventing something about Darius, can you leave at least one message without deception?
    Genetics has proven everything, what you write is strictly anti-scientific.

    ReplyDelete
  73. @arza & ambron, following might be interesting to you on the subject you discuss of when earliest samples showing tendencies of "Balto-Slavic" genetic drift show up in the record.

    Now that I have a dates on most of the samples in G25 from the latest anno file, can look more into that question. (Also added some dates on for the Tollense, Krakauer Berg samples that might be relevant).

    One way to try and summarize the "Balto-Slavic" drift is to look at distance difference between the real LVA_Baltic_BA (dated at around 550 BCE, although the similar EST_BA go up to 1250 BCE), which forms the pole of this drift, against a model. The principle being that relatively higher preference for the real population over the model indicates some participation in the shared drift.

    Here are some graphis on that: https://imgur.com/a/twLnr2b

    The first plots show that contrasting euclidean and squared euclidean distance to the real LVA_Baltic_BA average, where you can see that present day Baltic and to a lesser extent Slavic populations have lower distance to the real population, while most other populations have similar distance to both, or slightly lower to the model.

    Then other plots compare difference in these measures over time. Low can see the first samples with a clear preference for LVA_Baltic_BA over model, and really those that most prefer the model too, begin to emerge between 2000-1000 BCE.

    Note that this model is built on the assumption that farmer ancestry in LVA_Baltic_BA comes from POL_GAC, so this may inflate the difference level if farmer ancestry in LVA_Baltic_BA did not actually come from GAC.

    Same plots for Europe samples only: https://imgur.com/a/orJCK1v

    Note that actually prefering LVA_Baltic_BA over model is probably a high hurdle for detection of "Balto-Slavic" drift; some samples like CZE_Bell_Beaker_o I5025 (last dated at 2250 BCE, although removed from latest Human Origins anno file!) or HUN_Mako_EBA who tend get picked up as having Balto-Slavic drift, still aren't closer to real LVA_Baltic_BA than the model. Likewise some populations like UKR_Chernyakhiv_Legedzine or even RISE CZE_Early_Slav sample. And early populations who to me to look plausibly a bit incipiently shifted on Balto-Slavic cline, like HUN_ALP_MN_C.

    When using this measure then, it may be best to use normalize your expectations relative to present day populations (e.g. CZE_Early_Slav shows about the same level of preference for LVA_Baltic_BA over the model as present day Czechs, so this is interesting even though both populations aren't closer to LVA_Baltic_BA compared to the model. Other examples: HUN_Sopot_LN:I4184 and HUN_MN:I2375 prefrence for LVA_Baltic_BA with the range of present day Croats, Bosnians and Czechs). But it's one way to look at things.

    As one final measure, distance difference as % sum distance, a Vahaduo inspired measure that scales the distance difference relative to the distance from both: https://imgur.com/a/JN00zvI

    Here's a pastebin file with the dates in case you want to merge into your data: Ancients: https://pastebin.com/stDFZQtA, Moderns: https://pastebin.com/Vkzb4S4C

    (Obviously dates for moderns are all present day, but having in same format might be useful for merge).

    ReplyDelete
  74. @ Arza
    Im not hugely excited about an Y-hg I1 La Tene being directly relevant Balto-Slavic
    Even if such ancestry was permeating around (from wherever it was), it would not exclude later reformulations. I don't think we can at present go around the notion that Slavs were expanding in the middle of the first millenium AD (even after Ambron's claim that 5 ga-jillion people were living in Poland during the RIA).

    ReplyDelete
  75. Rob, let's be precise, please... I was talking about 1 million inhabitants of Poland and 5 million (men and women) of Slavic (by today's standards) R1a. My point was that no small Proto-Slavic homeland would not hold such a large population. Therefore, the Slavic R1a co-created other ethnoses as well - for example, the Baltic and Celtic. The sample of La Tene is important for this reason that the first typical Slavic clades M458 are to appear in La Tene.

    Matt, great job! I'm just analyzing it.

    ReplyDelete
  76. @Ambron, cool, see what you find! Btw, re-ran what I did above with a slightly more basal model (Koros_N, IronGatesHG, CordedWareEarly) and got similar results: https://imgur.com/a/nT0wEIc

    Another indirect index of when Balto-Slavic shift tends to appear might be looking at samples that most prefer more recent historical Western European samples over models: https://imgur.com/a/0mEZRF3

    Would say on a tangent that it looks like models using a mix of most early plausible ancient samples (e.g. Anatolia, WHG, Steppe_En) tends to be slightly closer to outgroups and really ancient samples (Ust Ishim, Yana, Tianyuan etc) relative to real recent samples. So probably good thing that Vahaduo modeling method doesn't try to indirectly infer ancestry by matching distances to outgroups, or anything like this, and just minimizes PCA distance directly.

    (Graphically here: https://imgur.com/a/opzjZqO

    You can see that comparing distance to model vs real ancient for both LVA_BA and GBR_Scotland_LBA, about the same time, the blue dots (European) are dispersed around the plot of preferring either the model or "real", while the other coloured dots (other world populations from outside Europe) have a slight but visible preference for the model).

    ReplyDelete
  77. Matt, how do you propose to interpret these results? Is it so that the samples above the 0.00 coordinate passed through the Balto-Slavic drift or got it in admixture, and those under the 0.00 coordinate did not?

    ReplyDelete
  78. @ Ambron

    I certainly agree that proto-Slavs did not simply emerge from Polesie

    ReplyDelete
  79. @ambron, yeah that's the problem; I think I'd say that deviation from equality is an indicator that a sample has some of that drift, but doesn't detect the absence at low bounds.

    I will experiment more with data and see if any better visualisations turn up.

    ReplyDelete
  80. @EastPole
    If it is real then probably Balto-Slavic CWC, which was originally the same as Indo-Slavic CWC (red circle), mixed with some HG around Baltic, in Poland and Hungary because there are many ‘drifted’ samples there

    Bingo. With the exception of hunter-gatherers around the Baltic. They weren't the source of this drift and the origin of Baltic_BA lies somewhere to the south.

    in Poland and Hungary there was the boundary between EEF and HG’s and all sorts of EEF-HG mixes existed

    Mixes like this one:

    3512-3350 calBCE (4621±28 BP, MAMS-28614) Romania_C_oHG.SG Gura Baciului Romania

    Target: ROU_C_o:GB
    Distance: 1.7937% / 0.01793666
    55.4 SRB_Iron_Gates_HG
    31.4 ROU_N
    13.2 Baltic_EST_BA

    Zeroes:
    0.0 Baltic_LVA_HG, 0.0 Baltic_LVA_MN, 0.0 Corded_Ware_Baltic, 0.0 Corded_Ware_Baltic_early, 0.0 Corded_Ware_CHE, 0.0 Corded_Ware_CHE_o, 0.0 Corded_Ware_CZE, 0.0 Corded_Ware_DEU, 0.0 Corded_Ware_DEU_o, 0.0 Corded_Ware_POL, 0.0 Corded_Ware_POL_early, 0.0 Corded_Ware_Proto-Unetice_POL, 0.0 GEO_CHG, 0.0 ROU_C, 0.0 ROU_Iron_Gates_HG, 0.0 ROU_Iron_Gates_N, 0.0 ROU_Meso, 0.0 RUS_Karelia_HG, 0.0 RUS_Khvalynsk_En, 0.0 SRB_Iron_Gates_HG_o, 0.0 SRB_N, 0.0 SRB_Starcevo_N, 0.0 UKR_Dereivka_I_En1, 0.0 UKR_Dereivka_I_En2, 0.0 UKR_EBA, 0.0 UKR_Globular_Amphora, 0.0 UKR_Meso, 0.0 UKR_N, 0.0 UKR_N_o, 0.0 Yamnaya_RUS_Samara

    ReplyDelete
  81. Rob, Polesie is troublesome due to the lack of the old Slavic toponymy.

    Matt, can you add samples N47 and N49 from Kuyavian CWC?

    ReplyDelete
  82. @Matt

    That's rather not a good method. Imagine a 50-50 mix between real and model pop. Difference of distances would be 0 and you would conclude that the sample doesn't show any special preference for the real population despite that it would have 50% of ancestry from real pop and thus also 50% of that specific drift.

    ReplyDelete
  83. @ Matt

    I5025 is still in the dataset under different IDs (RISE567 for shotgun and I4136 for capture data):

    I4136 RISE567, F0523, A0766, gr. 8 tooth 2015 OlaldeNature2018 (known there as I5025)

    ReplyDelete
  84. @arza, well in that example no, populations would only score 0 if they had equal distance to model and target. But the exact distance doesn't matter only whether a sample is outlying or not relative to others. However I think it's not a great method for other reasons. Still going to look for something else to plot against time...

    ReplyDelete
  85. @Matt
    You can compare relative position, but only on dist vs dist plots (and only for the nearest samples you'll get a clear visual signal - see for example Yukagir_Forest, who are half-Slavic, but very distant from Slavs). In case of dist diff vs time you can't do this reliably.

    RISE569 has dist diff close to 0. Now even if it will stand out of other samples on a dist diff vs time plot, a medieval chimpanzee will stand out too in the exactly same manner as RISE569, because it'll be practically equidistant to everyone (dist diff close to 0). Thus you'll be forced to conclude that medieval chimpanzees had the same level of Balto-Slavic drift as CZE_Early_Slav.

    ReplyDelete
  86. Hi David,

    I apologise if this question/comment seems out of place but this is the only recent blog entry of yours that I felt that I could pose this question without being off-topic.

    I recently came across this comment on YouTube:

    "
    Native american genes are pretty much identical to central asian and south central siberian. .. also close to east asian. Sure there was some caucasoid folks in north that the native american ancestors mixed to,, but native american is mostly of asian/mongloid race, only part caucasoid. The native american is mostly of haplogroup Q male lineage, witch is of same origin as east asian O and N. So east asian N/O came from same male ancestry as native american Q and those then went to different directions.
    Europeans have similarity in genes to natives, because of the multiple raids and conquest the central asian horse men committed to europe. Central asian and native american are pretty closely related actually, so thats where the similarity.
    Haplogroup R that all europeans now have is from central asian mongoloid race of people closely related to other asian groups as Q and O, but it does not affect europeans look anymore and asiatic features and mutations and such from those asian male ancestors have faded away from modern europeans and now europeans just look caucasoid. In other words the early "aryans" came of central asian horse men similar to huns raiding towards europe and takin white woman, as those groups usually were men the asian look dissappeared pretty quik from those new indo europeans, because people usually get their physical appearance more from the mother.

    This same thing is reason why finns are whiter than sami people although both sami and finns have usually the same amount of east asian male lineage haplogroup N. Well the finnougrics that lived in northern parts the native way had also woman from east with them so they werent only uralic men but also some woman, while the ones that lived in south were mostly only uralic men, so they only married white woman from the european culture living south of them, witch turned the southern finnougrians whiter that assimilated to those european cultures. The khanty, mansi people and sami in north staid closer to the older asiatic look because they did not marry as much white woman like the ones that assimilated to european farmers in south and married only white woman. So the sami having also asiatic woman from siberia with them helped to keep the asiatic look, while the southerners (finns, karelians, komis, mordvins, estonians) lost it by marrying only white woman and assimilated to the neigboring european farming cultures entirely.

    So, basicly similarly if those central asian horsemen that conquered parts of central europe, had also asian woman with them, the europeans would look pretty much asian now and there really would be no such thing as white people and entire eurasian continent would be mostly mongoloid race and history would have gone very differently. But because it was only asian men, the caucasoid/europid look came mostly from white mothers and generation after generation the asiatic look disappeared, and only thing left from those raiders from those central asian male ancestors is the letter R in dna and skill of riding horse that they learned from their fathers. (yes horse raiding was not invented by white people) Tecnicly modern days european/white people came of mixed race marriage(and rapes) between central asian man and european woman. After many generations the look of the asian fathers is completely gone and their entirely caucasoid despite having the asian marker in male lineage.

    "

    ---

    In short, I was wondering what your thoughts were on the comment.

    ReplyDelete
  87. It looks like it was written by a mentally unstable person.

    Obviously, both R1a and R1b are Y-haplogroups that are indigenous to Europe.

    And horse riding was invented by the offshoots of the Corded Ware people (Andronovo, Srubnaya etc.) and Corded Ware people came from Europe.

    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