November 2, 2016

Main Neanderthal admixture episode was c. 100,000 years ago.


This is really nice to read, considering that the archaeological data strongly favors a single out-of-Africa migration around that date (c. 125 Ka to Arabia and Palestine, see this, this and this among others, c. 100 Ka to South and East Asia) and that my own genetic modeling on mitochondrial DNA also fits that chronology (unlike most "molecular clock" scholastic rantings that are sold as "scientific truth" with no substantive backing whatsoever).

Admittedly the paper is not new (was published in February) but you know I have been missing important stuff with my information-overload stress crisis, so I'm making up now. Thanks to Ryan for bringing this up.

Martin Kuhlwilm et al., Ancient gene flow from early modern humans into Eastern Neanderthals. Nature 2016. Pay per viewLINK [doi:10.1038/nature16544]

Abstract

It has been shown that Neanderthals contributed genetically to modern humans outside Africa 47,000–65,000 years ago. Here we analyse the genomes of a Neanderthal and a Denisovan from the Altai Mountains in Siberia together with the sequences of chromosome 21 of two Neanderthals from Spain and Croatia. We find that a population that diverged early from other modern humans in Africa contributed genetically to the ancestors of Neanderthals from the Altai Mountains roughly 100,000 years ago. By contrast, we do not detect such a genetic contribution in the Denisovan or the two European Neanderthals. We conclude that in addition to later interbreeding events, the ancestors of Neanderthals from the Altai Mountains and early modern humans met and interbred, possibly in the Near East, many thousands of years earlier than previously thought.

As the title and the abstract say, Homo sapiens migrants out of Africa (i.e. into Asia and only later into its periphery) genetically influenced the branch of Neanderthals represented by the Altai specimen, what we can consider "Asian Neanderthals" but not the branch represented by El Sidrón and Vindija ("European Neanderthals"). This happened some 100 Ka ago, coincident with the archaeologically demonstrated dates for the out-of-Africa migration for our species and also likely Neanderthal-Sapiens hybrid fossils like Skhul-5 (right -- notice its lack of chin, a key and universal trait of H. sapiens, which allows us to give up with heavy browridges and other facial armature and still retain a strong bite, among other Neanderthaloid features, however it has a rounded and elevated skullcap with a high, almost vertical, forehead, a clear Sapiens trait). 

The flow was in both directions. I do not have access to the paper itself but it is clear in the supplemental material (EDF-1). This hybridization event was distributed quite evenly among all "Greater Asians" or "non-Africans" in our species. The slightly lower score in French is surely caused by Neolithic admixture later on, bringing African-like genetics to Europe, which are absent in most of Asia, as well as in aboriginal Australasia and America. 

Another apparent highlight in this paper (EDF-7) is that a second Neanderthal population, belonging to what I called above "European Neanderthals" (but related only to El Sidrón and not to Vindija) seems to have starred a second hybridization event affecting mostly Eastern populations (Han Chinese and Papuans in this paper's dataset). This, if confirmed, is quite unexpected and would require some explanation of the kind: there was a "European Neanderthal" population somewhere in Asia in the early times of Homo sapiens colonization and they got again admixed but this time affecting the derived populations in an irregular way. These irregularities would eventually be "flattened", I guess, at regional levels but for then the West Eurasian founders were out of the way. 

A somewhat related recent paper, also mentioned by Ryan, is S. Sankararaman et al., The Combined Landscape of Denisovan and Neanderthal Ancestry in Present-Day Humans (Current Biology 2016), also pay-per-view, so judging on supplemental materials only. I must say I don't like this one that much but at least table S2 offers a summary of the state of the art of estimates of Neanderthal and "Denisovan" (H. heidelbergensis) admixture in a lot of populations (and not just three). It is apparent there that there is more Neanderthal admixture towards the East of Asia or "Greater Asia", what is very much counterintuitive and demands that a Neanderthal population (a "European Neanderthal" one per Kuhlwilm's data) existed somewhere towards the East of Asia, enabling for this secondary Neanderthal admixture event. 

Perplexing maybe but that's what the data says. I wish we could find and sequence some of those Eastern Neanderthals which are so far just a genetic ghost, with the only possible known paleontological evidence being the Narmada skullcap, which is admittedly very much Neanderthal-like but is not associated in any way to Neanderthals' typical industry: the Mousterian, which has never been found east of Iran nor south of Mongolia. 

Some have argued that the Zhirendong jaw, one of the key evidences for c. 100 Ka H. sapiens settlement of much of Asia, is a hybrid one, with clear H. sapiens traits (among others it has a chin) but also maybe "archaic" traits (among others its chin is rather small). If so, then we may be at the same time in this case before evidence of both the early migration of our species, Homo sapiens, to East Asia (or SE Asia, as it's quite to the south of China) and the second admixture event with Neanderthals, with those ghostly Oriental Neanderthals, related to El Sidrón ones in the Far West, quite paradoxically, that we have yet to properly identify.

76 comments:

  1. I hope it's alright if I switch the discussion to the more recent post.

    "I prefer to consider the whole region loosely as a single entity, regardless of continent, because in the end: is there any difference if CF'DE evolved in Sudan, Eritrea, Yemen or Ethiopia?"

    I do agree with your scenario (or at least think it`s the most plausible one we have with the limited information available), but if CT (CF`DE) coalesced in Yemen, and DE coalesced in Africa, that`s by definition a back-to-Africa migration, with the expansion of D by definition a distinct Out-of-Africa migration.

    "No, I'm not really mistaking anything (even if I did not consider Tibetans specifically) because you're obviously cherry-picking the data points to fit your conjecture."

    I think you may still be. Correct me if I'm wrong here - but your objections solely re: haplogroup M, and not the admixture side, right?

    "1. M-dominated: South Asia (only!!!)"

    The only reason I neglect South Asia is because it is clearly of recently mixed (intrucive) West Eurasian and East Eurasian ancestry, with the M clearly coming from the East Eurasian side.

    Your categorization:

    . South Asia (M very dominant)
    · SE Asia (N/R very dominant)
    · Australia (N dominant, just some M)
    · Papua/Melanesia (N/R & M)
    · Middle East Asia (China, Korea & Japan) (N/R & M)
    · NE Asia (Mongolia, Siberia) and Native America (N/R & M)
    · West Eurasia & North/NE Africa (N/R dominant but also some M: M1, CZ, etc., not considering L(xM,N) here)

    I generally agree with your characterization, but my categories are this:

    1) Populations maternally descended from a population that included both haplogroup M and N. Obviously drift has altered the overall frequencies, but all of these populations have deep and diverse mtDNA from both haplogroups N and M. I'd include everyone except West Eurasia / North Africa in this group.

    2) Populations maternally descended from a population that was exclusively N. Any mtDNA of haplogroup M seems to have expanded relatively recently into the region and intrusive. This would be West Eurasians / N Africans.

    Still digesting some of your critique though - a lot strikes me as valid. Give me a bit to consider it.

    ReplyDelete
    Replies
    1. For the casual reader, the discussion comes from:

      → http://forwhattheywereweare.blogspot.com.es/2016/10/the-300-improved-genomes-project.html

      I asked Ryan to move it here, as this entry was directly motivated by it.

      Delete
    2. I still don't see a clear correlation between greater Neanderthal ancestry and mtDNA M. However I understand your reasoning better now, particularly after your later comments on Fu 2016.

      However it is much more suggestive IMO of some sort of founder effect in Western Eurasia. Maybe it did not affect everyone in the region at first but later became more dominant. In any case, we should expect to see at least some of the "East macro-Asian" secondary (El Sidrón related) Neanderthal admixture if it was just a matter of "selection" or "founder effect" and we do not (or almost not), so it seems more likely to me that it was a founder effect but with an implication: the group moving West was almost absolutely unaffected by that secondary Neanderthal admixture, what implies that it was, back then, only affecting some Eastern macro-Asians and not all of them.

      This could fit your hypothesis but it does not work well in the end because of South Asia, which has clearly less overall Neanderthal admixture than East Asia or Papua and is however much more M-dominated than these Oriental populations are.

      You argue that: The only reason I neglect South Asia is because it is clearly of recently mixed (intrucive) West Eurasian and East Eurasian ancestry, with the M clearly coming from the East Eurasian side.

      Not at all. The M in South Asia is South Asian, with very few exceptions. East Asian ancestry in the subcontinent is very constrained to some populations, almost all linguistically rooted in SE Asia: very strongly in the case of Tibeto-Burmans (Himalaya area in essence) and much more weakly (but still apparent) in the case of Austroasiatic populations (mostly around Orissa). Otherwise South Asians are made up of paleo-Indians (ASI) and Neolithic West Asian inflows of Caucasus affinity and almost certain Zagros origins (ANI) (add maybe a small bit of Indoeuropean ancestry from the steppes, but not really meaningful for our debate, more so as even the most ancient proto-IEs also have a strong Caucasus-Baloch, i.e. ANI, component).

      You could still argue that the ANI or Zagros Neolithic component has reduced the Neanderthal fraction in South Asia but then what about tribals like the Irula, who seem quite unadmixed?

      In any case it's not an East Asian + West Eurasian admixture what we see in South Asia but an old South Asian + West Eurasian (West Asian to be precise) admixture instead. This admixture is well known, clinal (graded) and we can explore if and how well it correlates with Neanderthal ancestry. So far my impression is that it doesn't or at least not significantly, that even the most ASI populations do not have the East Asian Neanderthal ancestry fraction but much smaller ones.

      There could still be a negative correlation between mtDNA N or R and Neanderthal ancestry in East Asia but I don't see it: Papuans (high R) and Australian Aborigines (high N) are similar to Han (mix of N/R and M) and AFAIK also to SE Asians (high R). So it looks like something else is going on here and not such a simple clear-cut lineage distinction (it doesn't work with the related male lineage K2 either).

      Delete
    3. One possibility could be that some (but probably not all) of the East Asian branches of mtDNA M, very early on in the colonization of Asia and Australasia, were the ones involved in that secondary admixture (which later became diluted, after the Westward migration of populations not involved in it). But we are already getting into population-specific stuff and something very hard to explain is that Papuans are by haploid lineages as related (mtDNA R) or even more related (Y-DNA K2b) to West Eurasians than to East Asians (Han). Then why they have almost the same secondary (Sidrón-like) Neanderthal fraction as the latter? Is it a coincidence? In any case it seems impossible to explain on haploid lineage grounds, really.

      Delete
    4. Maju,

      What is the evidential basis for your distinction between South Asians and East Asians and association of mtDNA haplogroup M more strongly with the former? The evidence I have at my disposal indicates that mtDNA haplogroup M is no more common in South Asia (roughly 50% to 60%, higher in some demographically miniscule tribes) than it is in East Asia (roughly the same, though tendentially higher in Japan and in some minor populations like the Tibetans). Kivisild et al. 1999 had 57.5% M in their rather southerner- & tribal-heavy pool of 550 Indians (250 Telugu, 122 samples with mixed-caste status but predominantly Indo-Aryan language speakers from Uttar Pradesh and Kashmir and with Uttar Pradesh samples being borrowed from a separate Gypsy study, 86 Lambadi, 62 Lobana, 18 Buksa, and 12 Tharu). A sample of non-Tharu Hindus collected in Chitwan District of the central Terai, Nepal (n=24) was 54.2% M, and a sample of Hindus from New Delhi (n=48) was 50.0% M. I think you may be overemphasizing a tendency toward high frequency of mtDNA belonging to subclades of haplogroup M that has been observed among certain tribal minorities.

      There is a general cline in the frequency of haplogroup M in East Asia, increasing from south to north. Of all the nation states, it appears that Japan probably has the highest frequency of mtDNA haplogroup M. However, if you lump all Southeast Asian populations together with East Asians and North Asians (and perhaps you would like to add Americans for good measure), you might find that mtDNA haplogroup M is more common among South Asians than among that sprawling group.

      Delete
    5. Check the autosomal structure here (for instance): http://forwhattheywereweare.blogspot.com/2016/03/south-asian-autosomal-structure.html

      Check the mtDNA frequencies here (for instance): http://www.scs.illinois.edu/~mcdonald/WorldHaplogroupsMaps.pdf

      I tend to consider you a knowledgeable person on these matters, Ebizur, so I am utterly surprised I need to explain such basic stuff to you of all people, really. Maybe you are way too narrowly focused on Y-DNA?

      "The evidence I have at my disposal indicates that mtDNA haplogroup M is no more common in South Asia (roughly 50% to 60%, higher in some demographically miniscule tribes) than it is in East Asia (roughly the same, though tendentially higher in Japan and in some minor populations like the Tibetans)".

      Could you point me to some "evidence" that is dated more recently than 1999? I thought that my own reference, the McDonald 2005 maps was already quite old (but handy anyhow and reasonably good too). It has some references to its own sources if you want to push the matter further into 21st century publications (even if they are 2001-2004 or so).

      For other papers on mtDNA, including for example the seminal work of Metspalu 2004, which included frequency maps of some of the most common lineages identified back in the day, check also here: http://ourorigins.wikia.com/wiki/MtDNA_haplogroup_M

      "There is a general cline in the frequency of haplogroup M in East Asia, increasing from south to north."

      Yeah, but if you go southwards towards Melanesia it also increases and the same happens, even more dramatically, if you head westward into South Asia. This is probably because M is older and SE Asia was the origin of the secondary expansion of a population (IMO after the Toba catastrophic episode) defined by the dominance of mtDNA N/R (rather R but no clear distinction can be made, as most N subclades were quite apparently involved too) and Y-DNA K2 ("MNOPS"). Only in Australia we see a partial exception which is mtDNA N(xR) along Y-DNA C4.

      In case you need references, it'd be Karafet 2014 for the K2 issue (see here) and my own work for the overall analysis (with mtDNA as main reference), as explained in this dedicated page.

      My understanding re. M is that it is not at all "parallel" to N but an older expansion, in which pre-N was unsuccessful, remaining as small or "private" lineage until it could expand in this secondary episode which I tend to associate with the post-Toba scenario, i.e. approx. 74-50 Ka BP for the full N→R→U/R0 coalescence process (using Western refs. because the first UP, the beginning of the end of Neanderthal dominance, is an unmistakable chronological reference here, while in the East most things are less precise). It's not just N which expands: many M subclades do too, but none with the striking success of N or R and their (almost perfectly) "coupled" Y-DNA K2.

      Delete
    6. One thing that distinguishes South Asia is the depth and diversity of mtDNA haplogroup M there, so I do agree with you (Maju) about South Asia being very significant to haplogroup M and its expansion. I think the diversity is the more important data than the actual haplogroup frequency too. South Asian haplogroup M is *old*.

      "Not at all. The M in South Asia is South Asian, with very few exceptions."

      ASI forms a clade with East Eurasians though. When I say "East Eurasian" here I mean in the very broadest sense.

      And to be clear - I mean South Asians are admixed on the autosome, not mtDNA.

      "This could fit your hypothesis but it does not work well in the end because of South Asia, which has clearly less overall Neanderthal admixture than East Asia or Papua and is however much more M-dominated than these Oriental populations are."

      My point re: South Asia is that the mtDNA haplogroups frequencies in South Asia are no longer representative of the autosomal DNA as a whole due to male-biased migration of ANI. I'm viewing South Asia as mostly a union of ASI women with ANI men - with ASI forming a clade with East Eurasians and ANI forming a clade with West Eurasians.

      I think you're missing the point here a bit still too. As I said before, I'm suggesting that the additional admixture happened in a population that had both N and M. Pointing to Neanderthal admixture peaking outside of South Asia has nothing to do with my claim.

      Archaic ancestry does increase noticeably in ASI though. Go to the Sankararaman paper again. There isn't a great selection of ASI-rich populations, but there are two that I see - the Khonda Dora (a Dravidian speaking hill tribe from SE India) and the Onge. Both show inflated Neanderthal ancestry. So I think it's reasonable to infer that ASI was probably similar to other macro-East-Eurasians in its Neanderthal ancestry levels.

      And of course the Irula you mention. Their Neanderthal ancestry is slightly lower, but still above any West Eurasian except the Saami, so that seems to support my point, no?

      On an aside - there's something I don't get about the Irula - they're very ASI, but don't they have some of the highest mtDNA haplogroup U frequencies of any tribal groups? And IIRC haplogroup U peaks in southern tribals, and is lowest in eastern tribals, with northern tribal Indians intermediate. I don't think this relates to the current discussion in a huge way, but it's somewhat counterintuitive.

      Delete
    7. "But we are already getting into population-specific stuff and something very hard to explain is that Papuans are by haploid lineages as related (mtDNA R) or even more related (Y-DNA K2b) to West Eurasians than to East Asians (Han)."

      K2b is most certainly *not* a West Eurasian lineage, and its descendants in Europe are mostly from recent expansions between 14,000 and 2,000 years ago. I think it's pretty clear that K2b and K2 are originally either South Asia or South East Asian, and arrived in Europe via the steppe in multiple waves. And those early waves often had noticeable East Eurasian admixture/affinity (Villabruna did).

      "There could still be a negative correlation between mtDNA N or R and Neanderthal ancestry in East Asia but I don't see it"

      I'm not claiming such a correspondence at all.

      Take this Treemix run by Davidski as an example: https://drive.google.com/file/d/0B9o3EYTdM8lQaVlGbzBvX3FVdHM/view?usp=sharing

      Everything above Gujarat I'm referring to as West Eurasian, and everything from Gujarat down I'm referring to as East Eurasian.

      What I'm claiming is that at that population that was the original East Eurasian branch had:

      - a mix of mtDNA haplogroups M and N
      - mostly Y-chromosome haplogroups K and C
      - more Neanderthal ancestry than the West Eurasian branch

      Whereas the original West Eurasian branch had:

      - exclusively mtDNA haplogroup N
      - exclusively Y-DNA haplogroup F(xK)
      - decreased Neanderthal ancestry

      And I'm further claim that any blurring of those lines is due to later migrations. And I feel in particular the haplogroup diversity in each group is indicative of this (moreso than the frequencies).

      Now as for the reasons for this... you've made me seriously consider why.


      On the weird Paleoafrican to Oceania admixture signals, Razib has a couple posts on it:

      http://www.unz.com/gnxp/ghost-populations-in-our-genetic-closet/
      http://www.unz.com/gnxp/the-ancient-scramble-for-eurasia/
      http://www.unz.com/gnxp/the-flying-wing-into-a-graph/

      The short story seems to be that these weird signals are probably telling us something real, but we have no idea what that thing is.

      Delete
    8. "ASI forms a clade with East Eurasians"...

      Why do you say that. I have never seen anything like that: the South Asian "ASI" component is very specific of that subcontinent, distinctive, and only when a dualistic polarization is generated, quite artificially (or in modelings involving the controversial "basal eurasian" thing), such a tendency appears. But it is a mirage, not anything that can be accepted as realistic, just a product of too shallow improductive analysis.

      It is unavoidable that, as ANI is West Eurasian, ASI will appear as "not West Eurasian" but that's about it. If you make a dual polarization West Eurasia vs Africa, with some Indians in it, I'm sure that ASI will appear as "more African" but only because it is "not West Eurasian" and Africans are also "not West Eurasian". Bipolar modeling is junk.

      ... "the mtDNA haplogroups frequencies in South Asia are no longer representative of the autosomal DNA as a whole due to male-biased migration of ANI."

      Probably true but there's still not significant East Asian ancestry, ASI is local.

      "with ASI forming a clade with East Eurasians and ANI forming a clade with West Eurasians."

      The only "clade" that ASI and East Asians (or Papuans, Onge, etc.) form is one called "not West Eurasian". Otherwise they have no relation whatsoever.

      As West Eurasians are a branch of "not West Eurasians" (plus a small fraction of that African-like thing called "Basal Eurasian") this "not West Eurasian" clade is not any actual clade, unless you're dealing very specifically with the issue of Basal-Eurasianness, in which case (and ONLY in that case) it would be indeed true for West Eurasians and the derived ANI component in South Asia.

      "there's something I don't get about the Irula - they're very ASI, but don't they have some of the highest mtDNA haplogroup U frequencies of any tribal groups?"

      Good question. I would have to look up the details but out of the top of my head, I'd say that they carry U sublineages that have been in South Asia since the beginning of U scatter, probably U2 and U7 sublineages. This input should be tracked to the very original split between ASI and WEA (West Eurasians) c. 50-40 Ka ago (or maybe up to 30 Ka, as blade tech arrived to India c. 38 Ka ago if I recall correctly), so it shows up as ASI ("old Indian"), while only the autosomal component arrived since Neolithic is distinctly WEA-like (ANI). I would think that there should be some hidden structure to ASI (South Asia is big enough to have some such structure, just as we do find a structure in ancient West Eurasia) but AFAIK nobody has researched it yet (aDNA would help of course but not strictly necessary: ADMIXTURE should be able to detect such deeper structure as well in modern datasets).

      "And IIRC haplogroup U peaks in southern tribals, and is lowest in eastern tribals, with northern tribal Indians intermediate. I don't think this relates to the current discussion in a huge way, but it's somewhat counterintuitive."

      Not sure why you say that, Metspalu 2004 is maybe one of the key references and U sublineages (U2 mostly) is rather scattered: U2i and U2b are very important in Uttar Pradesh, U2c in Bengal, while around Andrah Pradesh (SW) there is also concentration of U2i and U2a and U2i is also very high in Kashmir and Sindh. There's some lower density area maybe in and around Gujarat and of course in the SE but that's about it.

      Delete
    9. "K2b is most certainly *not* a West Eurasian lineage, and its descendants in Europe are mostly from recent expansions between 14,000 and 2,000 years ago."

      K2b, specifically P1 sublineages Q and R1 are indeed West Eurasian lineages shared with South Asia (and in the case of Q also with Native Americans and some NE Asians, but only because it migrated from West Eurasia via Altai). Both Q and R1 must have coalesced in West Asia.

      That Vilabruna is the first KNOWN European R1b1 only underlines this, because it implies that its precursors MUST have been in Europe since at least Gravettian times (no West Asian immigration to Europe between Gravettian and Neolithic). R1a must have expanded more recently but some branches of R1b have been lurking in Europe for millennia necessarily, just as happened with mtDNA H, so often its "couple". Another issue is if it re-expanded in the Neolithic but, even if it was restricted to Italy or wherever, R1b must have been in Europe for at least 32.000 years.

      "I think it's pretty clear that K2b and K2 are originally either South Asia or South East Asian"...

      Very specifically South East Asian, K2b almost certainly coalesced in Sundaland while K2a/NO probably did further North. See: http://forwhattheywereweare.blogspot.com/2014/06/y-dna-macro-haplogroup-k-m526.html

      While NO migrated northwards (along mtDNA N/R clades), K2b1 did to Papua (also with mt R subclades) and P1 migrated Westwards, leaving a track: P* at Bengal, R2 further West in India, R1 might even have coalesced in Pakistan (although West Asia is also possible). Q almost certainly coalesced in Iran, while R1b and R1a did in West Asia.

      So there's a whole story to Y-DNA K2 and mtDNA R (also N but very especially R) expanding from some chronological point (surely after the Toba catastrophe) from SE Asia in various directions. The western branch (P1) left a track in India but impacted much more clearly West Eurasia. However on the way they picked other Y-DNA lineages: IJ, G and LT (at least T). MtDNA M1 however must have come with the group from SE Asia however (its relatives are all East Asian), as probably did N1, N2 and X. N1 and N2 left some track in South Asia but not X, however X2 migrated eastward via Altai with Y-DNA Q1, confusing the mind of Dr. Standford and his followers for a long time (not anymore it seems, I'm talking of the "Solutrean hypothesis" of course).

      ...

      Delete
    10. ...

      "... and arrived in Europe via the steppe in multiple waves."

      Nope. Totally in disagreement: the steppes don't seem to have played any major role in the settling of Europe and even the Indoeuropean expansion in the Chalcolithic is tracked only to the West Eurasian (i.e. European) steppe, not further East. The East Asian lineages we see in NE Europe, mostly among Uralic peoples (Y-DNA N1 and mtDNA C notably), correspond to a migration that began around Mongolia after the LGM, had arrived to Eastern Europe probably around the Epipaleolithic. That is the only sizable pre-Altaic East Asian genetic influence in Europe and cannot be considered associated to the steppes (grasslands), rather the taiga (forests) instead. They were the first known users of pottery in West Eurasia (it's much older in East Asia) although it is unclear if they inspired its development in the Eastern Mediterranean Neolithic or this happened independently. They were almost certainly the precursors of modern Uralic peoples.

      This E→W migration is like the reverse version of the W→E migration that would produce Native Americans and bring "mode 4" (UP or blade tech) to East Asia via Siberia, Mongolia and North China, dated to c. 30 Ka BP and with almost certain roots in Altai, which in turn was settled (H. sapiens) from further south (Uzbekistan, Iran, Afghanistan/Pakistan maybe) at the very beginning of the Upper Paleolithic (47 Ka ago for Altai, maybe 50 Ka for Uzbekistan sites).

      The story of our migrations is tropical and subtropical until c. 50 Ka ago, even in China and Japan it's probable that they didn't settle in mildly cold areas until around that date (maybe a bit earlier?) Around that date, maybe because of increased density in the "Tropical and Subtropical Asian" regions inhabited until then, new and more challenging areas are settled, notably the Western "Neanderlands" but there is also a clear push towards the North, to places never before inhabited by our species like Altai, Japan, North China, etc. It is also probably then when the different "races" begin to form as distinct "looks", thanks to the consolidated subcontinental "boundaries" that demographic pressure itself enforced (with less dense areas like NE India or Afghanistan-Balochistan as buffer).

      Delete
    11. "Take this Treemix run by Davidski"

      Not realistic: it does not even represent the "Basal Eurasian" migration edge. And that is the only reason why West Eurasians appear sligtly more upstream in the tree. If you can remove the "Basal Eurasian" or in general the African-like genetic component, then West Eurasians should look as a simple branch of "macro-Asians", one among several. West Eurasians are the only Macro-Eurasian pop. with extra African (or similar) admixture, that is something we just cannot ignore.

      And that's why autosomal DNA is not always the best guide in our understanding (secondary admixtures, sample sizes, etc. can distort the results a lot, assuming the interpretative algorithm is fully correct, not necessarily the case). IMO mtDNA is much better: much simpler and straightforward. It's like a pilot wave we should always try to follow, Y-DNA is also useful but the tendency of secondary admixtures being male-biased may distort things too much, autosomal DNA is truly complicated to interpret, especially for deep chronologies. After all autosomal DNA interpretation is a matter of statistics and you know what they say about statistics, right: "there are lies, damn lies and statistics". Less dramatically, I can accept statistical interpretations with a grain of salt but ALWAYS with that grain of salt, else I may begin believing too many unlikely things for my own sanity.

      Delete
    12. PS- Another possible cause of distortion can be the archaic admixture frequencies: it's not big but it creates an anti-African vector that pushes West Eurasians artificially towards Africans, in addition to the previous. In your tree, there is the Denisovan admix. axis but not the Neanderthal one, which we know is not equal everywhere. So it's quite probable that the question you are trying to answer is itself distorting the apparent answer of that tree.

      Anyhow too many West Eurasians not to form a distinct clade in any algorithm vs all other macro-Asians. Sampling strategy matters too and, in doubt, various sampling strategies should be contrasted to see if they change the result and how.

      Delete
    13. PS2- Out of the top of my head: is "Basal Eurasian" (in part, probably not in full, because we can't ignore E1b) a negative of the extra El Sidrón Neanderthal admixture now detected in Eastern macro-Asians (Han and Papuans at least)? I've never seen such a refined testing, so I would not discard it.

      Delete
    14. Maju wrote,

      "Check the autosomal structure here (for instance): http://forwhattheywereweare.blogspot.com/2016/03/south-asian-autosomal-structure.html"

      Autosomal analyses are irrelevant. I asked you to explain your grounds for singling out South Asians (as opposed to East Asians) as exhibiting particularly high frequencies of mtDNA haplogroup M.

      "Check the mtDNA frequencies here (for instance): http://www.scs.illinois.edu/~mcdonald/WorldHaplogroupsMaps.pdf"

      Maps and pie charts are not so good for this sort of comparison as a table of frequencies or, best of all IMHO, an ordered list with frequencies, sample sizes, and citations.

      "Could you point me to some "evidence" that is dated more recently than 1999?"

      Do you have any reason to doubt the accuracy of their results? I should not have to remind you that the authors of that 1999 study are the same as the authors of the work that you have cited (Metspalu et al. 2004).

      Now, let us consider the meaning of the word seminal, from Latin sēmen "seed." In this context, the intended sense is "highly influential in an original way; constituting or providing a basis for further development; highly original, influential, and important; highly original and influencing the development of future events; containing seeds of later development." Synonyms include the following: influential, important, ground-breaking, original, creative, productive, innovative, imaginative, formative. It is interesting that you have considered a paper published by a group of authors in 2004 to be more "seminal" than a paper published by the same group of authors on the same topic in 1999.

      Leaving that aside, the 2004 paper that you have cited has not found anything that contradicts what I have stated. The authors have written, "We found haplogroup M ubiquitous at almost 58% among the caste, and 72% among the tribal populations (Table 2), which is largely consistent with previous reports." Scheduled Tribes accounted for 8.2% of the total population of India according to the Indian census of 2001 and 8.6% of the total population according to the census of 2011. Metspalu et al. 2004 is closer temporally to the census of 2001, but even taking the higher figure from the census of 2011, their data should indicate a frequency of approximately (0.58 x 0.914) + (0.72 x 0.086) = 59.2% mtDNA haplogroup M in the total Indian population. Of course, the representativeness of their samples is another question; note that from Uttar Pradesh, by far the most populous state of India, which with its approximately 200,000,000 inhabitants accounts for 16.5% of the population of the entire country, the authors have sampled mostly tribal Tharus (n=26) and Bhoksas (n=5), who are known to be genetically distinctive, with extremely high estimates of East Asian and indigenous South Asian ancestry based on analysis of both autosomal and mitochondrial DNA.

      The ideological climate is an acute problem when studying the people of a country like India. One must consider any data that do get published carefully, but your suggestion that a study published in 1999 is questionable while a study that has arrived at the same conclusions as the previous study and that has been published by the same authors in 2004 is ground-breaking and all-important is utter nonsense.

      See also: "Of the four major matrilines identified till date L, M, N and R, about 60% of Indians trace their maternal roots in Indian specific branches of haplogroup M that is reported to have emerged from the African haplogroup L3." (Revathi Rajkumar, Jheelam Banerjee, Hima Bindu Gunturi, R Trivedi, and VK Kashyap, "Phylogeny and antiquity of M macrohaplogroup inferred from complete mtDNA sequence of Indian specific lineages." BMC Evolutionary Biology 2005 5:26.)

      Delete
    15. Maju wrote,

      "Yeah, but if you go southwards towards Melanesia it also increases and the same happens, even more dramatically, if you head westward into South Asia."

      The frequency of haplogroup M in Southeast Asia appears to be about 40% on average. Samples from Oceania that have equalled or exceeded that average are samples from the Admiralty Islands of Papua New Guinea (42%, so roughly the same as the Southeast Asian average), Mussau (6/12 = 50% Q2; the island itself and the sample size are both very small, and haplogroup M is represented in this sample by only the subclade Q2), Bougainville (52%), West (i.e. Indonesian-governed) New Guinea (66%), New Britain (80%), and the Chamorros of the Mariana Islands (92% E).

      Basically, there are two zones of high frequency of mtDNA haplogroup M in Oceania. One zone includes West New Guinea (but not mainland Papua New Guinea) and the offshore islands of Papua New Guinea (including Bougainville, but not the politically independent Solomon Islands), throughout which haplogroup Q1 is frequent (though haplogroup M28 is even more important than Q1 in New Britain). MtDNA haplogroup Q1 also has been found with lower frequency in the Moluccas and Nusa Tenggara to the west of New Guinea. The other zone is the Mariana Islands, whose natives exhibit extremely high frequencies of the typically Austronesian haplogroup E (including both its E1 and E2 subclades). Elsewhere in Oceania (at least in Australia, mainland Papua New Guinea, the Solomon Islands, Fiji, and Polynesia; data regarding Micronesian populations are sorely lacking), the frequency of mtDNA haplogroup M is not particularly high, and it is in fact lower than the frequency of haplogroup M in Southeast Asia.

      Our discussion is quite trivial; I have chosen to participate only in order to point out the dubiousness of some of the evidence you have adduced to refute Ryan's hypothesis. However, I agree with you insofar as it does seem very likely that the frequency of haplogroup M in South Asia as a whole is somewhat higher than the frequency of haplogroup M in Southeast Asia as a whole, and that distinction becomes much starker if one excludes subclades of haplogroup M that are shared with northerly East Asians. Please keep in mind that many studies have not supported some early claims of a precedence of South Asia vis-à-vis East Asia in regard to the origin of mtDNA haplogroup M as estimated by either frequency or diversity.

      Delete
    16. @Ebizur:

      "Maps and pie charts are not so good for this sort of comparison as a table of frequencies or, best of all IMHO, an ordered list with frequencies, sample sizes, and citations."

      Those maps are the best, easy to read, reference I know of on the kind of global regionalized frequency data you are pushing for. Maybe you have something better? Nah, you would have already mentioned it.

      "I should not have to remind you that the authors of that 1999 study are the same as the authors of the work that you have cited (Metspalu et al. 2004)".

      But FIVE years, five critical years of research passed between one and the other. I have been following these matters for almost two decades now and it is a fact that the data improved a lot once we got several years into the 21st century. As rule of thumb I don't like citing pre-2005 papers for that reason, with a handful of very honorable exceptions, of course.

      "Do you have any reason to doubt the accuracy of their results?"

      No. But looking at the paper itself, it is not at all what you said. i.e.: "Kivisild et al. 1999 had 57.5% M in their rather southerner- & tribal-heavy pool of 550 Indians".

      Actually there is ONLY one population from the South (Andrah Pradesh Lambadis) which is 64% M. All the rest is North Indian (51-55%), Pakistanis (38%) and Kashmiris (28%), so the average is heavily distorted by the great weight of the Northwest.

      Also the sample is 299 (not 550) of which 86 are from the South (a very specific subpopulation).

      Unless you're talking of another paper (this one is cited as Kivisild 1999b, so there's probably some 1999a paper that I could not find). In any case the (limited) data also points to a very high frequency of M in South Asia, particularly once you get into India and towards the south of it.

      Around 60% is roughly coincident with what the McDonald maps say, however these probably take into account the even greater frequencies of the lineage towards the East and Southeast of the subcontinent, as well as in the central areas south of the Ganges, the true "ASI" core. The Northwest and to some extent also the Western coast are comparatively more R-influenced (both Indian-specific R lineages and U, as well as some N, particularly W in Pakistan).

      Delete
    17. I'm not going to bother with a semantic discussion about my usage of the adjective "seminal". I meant that it was a very important paper in our understanding of South Asian (and by extension global) mtDNA patterns. Enough said for a mere adjective.

      Re. the frequencies, I take that there is diversity among populations but that the McDonald average datum of around 65% M is very real. And this is a quite higher figure than what we find elsewhere, included East Asia.

      Also, as Ryan said, it's not just frequency: basal diversity is also astonishingly high, higher than in East Asia + Oceania (although maybe NE India and Burma, both high M "border" regions may tip the scales).

      "... about 60% of Indians trace their maternal roots in Indian specific branches of haplogroup M"

      I find that figure a bit low but it is close enough to what McDonald reports (which seems to be c. 65%). Higher in any case than what we find in East Asia, barring some specific regions like Burma or Japan, which seem to retain a higher frequency of what I'd call "pre-Toba genetics". Han Chinese (the vast majority of East Asians) have significantly less M and so do SE Asians in general (with pockets however of similar high M frequencies to those of South Asia). NE Asia (Mongols, Eastern Siberia) again have high M frequencies but most of it is of the very specific subclades: CZ and D, which we can consider to some extent "Northern specialists", along with N-derived A.

      That's the picture: mainline East Asia (Han, SE Asians) is not "low" but definitely "less high" in M than other regions like South Asia that retain a greater fraction of "pre-Toba" M-dominated genetic pools. However in some cases these fractions are largely "post-Toba", for example in NE Asia (the label must be used with caution).

      "The frequency of haplogroup M in Southeast Asia appears to be about 40% on average".

      Maybe. I'm not questioning that (although it significantly decreases if we take D apart).

      "Samples from Oceania"...

      Sure some islands in Melanesia are high in M and seem to retain very old ("pre-Toba" type) genetic pools to a large extent. Even Papuans have large fractions of M-derived Q, however they also have large fractions of R-derived lineages, notably P. And this is also "coupled" with K2-derived Y-DNA, so to a large extent they were affected by this inferred secondary "post-Toba" expansion of the pair Y-DNA K2 plus mtDNA R (and some N in other cases).

      ...

      Delete
    18. ...

      "Our discussion is quite trivial; I have chosen to participate only in order to point out the dubiousness of some of the evidence you have adduced to refute Ryan's hypothesis. However, I agree with you insofar as it does seem very likely that the frequency of haplogroup M in South Asia as a whole is somewhat higher than the frequency of haplogroup M in Southeast Asia as a whole, and that distinction becomes much starker if one excludes subclades of haplogroup M that are shared with northerly East Asians".

      OK, I can agree with all that. I thought it was a more substantial discussion actually, about something important.

      "Please keep in mind that many studies have not supported some early claims of a precedence of South Asia vis-à-vis East Asia in regard to the origin of mtDNA haplogroup M as estimated by either frequency or diversity".

      It may be debatable, I really can't say with 100% accuracy. However I have worked on the matter and the basal diversity of M seems still higher towards South Asia, notably if you count those sublineages that seem to have coalesced first, those that have the shortest stems downstream of M. Of these I count:
      · 5 in Western South Asia
      · 10 in North-Eastern South Asia (incl. 4 in NE India)
      · 6 in Indochina (mainland SE Asia)
      · 1 in Papua

      So the centroid of this scatter is rather towards Bengal, even Assam, however my method tries to infer a more reasonable origin and goes a step beyond the raw centroid. What I do is to consider not just the descendants but also the ascendant, in this case L3 and so move the estimated origin 1/3 or 1/4 towards the ancestor, which is roughly in Eritrea. That's why I draw the M estimated origin in NW India. You are perfectly free to disagree with this methodology, of course, but I do find it very reasonable.

      I actually do the same with N and R and that's why N does not fall near Tonkin or Canton but rather towards Burma. Inversely R should probably fall slightly farther West towards Bihar considering only the raw centroid but I think that pulling it a bit towards N looks quite reasonable, so Bengal is it. It's just an educated guess in all cases: uncertainty is too high, and in the case of both R and N I could easily agree with a source in maybe the centric location that is central Thailand (because of what we know of the related Y-DNA K2 and to some extent also C, involved with mtDNA N(xR) in Australia at the very least).

      A bit of common sense regarding plausible routes, good areas for expansion (for example not a desert, not the Tibetan plateau, not Neanderthal occupied and very cold Central Siberia) and the little we know about the late Middle Paleolithic from archaeology should be used complementarily to the available DNA data. But sure: there is some uncertainty that we just cannot get rid of.

      Delete
    19. If you want to be convincing, you should cite studies that provide detailed data rather than posting a link to a map with pie charts.

      The team of Kivisild, Metspalu et al. has published many papers on the mtDNA of South Asian populations. All their results agree (which is not at all surprising if one considers that the sample sets for all their papers on this region overlap to a great degree) and suggest an overall frequency of haplogroup M in India (and Sri Lanka) of approximately 60%. cf. T. Kivisild et al. 2003, "The Genetic Heritage of the Earliest Settlers Persists Both in Indian Tribal and Caste Populations," American Journal of Human Genetics 72: 313–332, 2003: 642/1090 = 58.9% M Indian & Sri Lankan Castes, 206/260 = 79.2% M South & East Indian tribes. In their 2004 study about which you seem to be singularly enthusiastic ("Most of the extant mtDNA boundaries in South and Southwest Asia were likely shaped during the initial settlement of Eurasia by anatomically modern humans," BMC Genetics 2004 5:26), they have unsurprisingly found a maximum frequency of haplogroup M among their sample of Mundari Austro-Asiatics, though with a rather small sample size (78/90 = 86.7% M).

      You must know that the population of South Asia is strongly concentrated in the north, from Khyber Pakhtunkhwa and Kashmir through Punjab, Haryana, Uttar Pradesh, Bihar, Bengal, and the valley of Assam. There are a few other hotspots, of course, such as the Karachi area in Sindh, the Mumbai area in coastal Maharashtra, and Kerala (and Tamil Nadu is also more densely populated than the Indian average), but an overwhelming majority of the population of South Asia lives in the Indo-Gangetic Plain. Therefore, including 250 Telugus of Andhra Pradesh in a sample of 550 Indians is quite excessive. Most of their remaining Indian samples were very marginal peoples (e.g. Banjaras, Tharus) who do speak Indo-Aryan dialects, but who are not necessarily representative of the mainstream castes of North India.

      I have previously mentioned in passing the results of one other team of researchers, Simona Fornarino, Maria Pala, Vincenza Battaglia et al., who have found the following:

      Non-Tharu Hindus collected in Chitwan District of the central Terai, Nepal:
      13/24 = 54.2% M
      10/24 = 41.7% R
      1/24 = 4.2% N(xR) (specifically, a member of haplogroup I)

      Hindus from New Delhi, India:
      24/48 = 50.0% M
      22/48 = 45.8% R
      2/48 = 4.2% N(xR) (specifically, one N1d and one W)

      Tharu tribals from Chitwan District, south-central Nepal:

      Village #1
      43/57 = 75.4% M
      14/57 = 24.6% R

      Village #2
      67/76 = 88.2% M
      9/76 = 11.8% R

      Tharu in Chitwan District total
      110/133 = 82.7% M
      23/133 = 17.3% R

      Tharu tribals from Morang District, southeastern Nepal:
      30/40 = 75% M
      10/40 = 25% R

      Tribals from Andhra Pradesh, India:
      22/29 = 75.9% M
      7/29 = 24.1% R

      The sample sizes are not so large as those of Kivisild et al., but their results agree inasmuch as they suggest that approximately 50% to 60% of members of caste populations and a significantly higher percentage of outcastes/tribals/adivasis belong to haplogroup M.

      Delete
    20. "If you want to be convincing, you should cite studies that provide detailed data rather than posting a link to a map with pie charts"....

      ... and references for the origins of that data, that you are free to browse.

      As you can imagine, to do all that work on the putative history of Eurasian mtDNA, I had to look at many different studies and not just one or two, in order to best understand each haplogroup's geography (frequency was not that important to me), however I did that years ago and why should I bother revising all the many studies (there are dozens and dozens) when the matter of contention is so generic, unspecific, that no single one of them will provide an answer? The pie charts' McDonald map is the best available to try to settle the issue under discussion here, unless you want to go specifically to this or that population or much more derived haplogroup, what is not the case.

      I'm not even sure which is your point anymore: you're producing data for non-tribal Indians only for the Northwest, the area most affected by West Eurasian affinity. What does it mean? Nothing at all.

      Most studies don't even consider M as such anymore but its sublineages distinctively (what makes sense unless what you want is a pie chart of generic frequencies). You go for example to Chandrasekar 2009 and you don't get absolute frequencies for all haplogroups but only for those he was studying. The same happens often enough for other regions. A lot of patience and digging is needed to generate a chart that gathers all frequencies, so I choose to let McDonald, who did the job in 2005, do it for me and for you as well.


      "... they suggest that approximately 50% to 60% of members of caste populations"...

      In Northwestern India and nearby Nepal. That is not data we can expect to be representative of all India!!!

      Delete
    21. Why do you think that East Asians and Oceanians are not closer to each other than they are to West Eurasians? this can't be just an effect of "Basal Eurasian" since this differential relationship still holds when the West Eurasian is a Palaeolithic European or Siberian, which lack this "Basal Eurasian" affinity as far as i know.

      Delete
    22. Well, the only way we can track a history of those populations at such paleohistorical depths is not in autosomal DNA but in haploid DNA, and that one says that the histories not just overlap but that there is a secondary common origin for everything within mtDNA N (or more importantly R) and Y-DNA K2, and that affects all Eurasians but more importantly East Asians, Papuans and West Eurasians (in this order if we follow K2 branching history).

      In autosomal DNA there can be way too many confounding factors, for example it may be true what you claim (don't know the data, so unsure) but we have to consider that, besides the "Basal Eurasian" or African-like admixture thingy, there are other admixtures that surely affected the original West Eurasians, very particularly genetics taken at the passage through South Asia, most apparent in Y-DNA IJ (but also other lineages), which should imply a sizable "primeval Indian" genetic impact on the founder effect of the West Eurasian "first population" of the earliest Upper Paleolithic, an element that should be totally absent in Australasians and only present in East Asians by West Eurasian mediation (Siberian Y-DNA Q1 migrants leading to Native Americans but also influencing NE Asia).

      It is clear that West Eurasians are closer to East Asians than they are to Papuans, while East Asians appear similarly distant from West Eurasians and Papuans instead (example of Fst distances' table). IMO that's largely because West Eurasians are not just "more African" but also "more Indian" in a primeval sense of the terms.

      Another factor we can't ignore is that there are differences in archaic admixture, which are no doubt small, but should have a multiplying effect in overall genetic distances, as any fragment of Neanderthal or "Denisovan" admixture is much more drifted away, accumulates much greater mutational density, from the common "Sapiens" root than the Sapiens-specific mutations here and there that scatter the rest of the genome. So, when compared with a relatively "low" archaic admixture in West Eurasians, Han and Papuans will tend to cluster no doubt. I wonder how it'd work with Onge, who are also an isolated basal Eurasian branch on their own right and are not known for having extra archaic admixture.

      So I would be extremely wary of attempting to build history out of autosomal DNA: it's too large, too revolted, too affected by lesser changes, be them by admixture, founder effect or mere drift by isolation (endogamy), to draw any conclusions, more so when the figures are so similar, not statistically significant.

      Delete
    23. On the Fst table you link, Papuan seems more distant to modern Africans than Han and French are, this is possibly an effect of drift or their Denisovan admixture. Taking this into account, Papuan does indeed looks closer to Han than anything else, and viceversa.

      Delete
    24. Actually I just realized it's not an Fst table, even if similar. I just looked for something like that online in a hurry and was not precise enough, sorry.

      Wikipedia has some genetic distance values in its relevant entry but no Papuans nor other isolated branches of the SE (Australian, Andamanese, etc.) And I can't find any good reference table of the kind. :(

      But notice that Fst and such distances are always average of the overall genome and that, if we could do instead an organized analysis segment by segment, we would see (I think) that in different sets of segments the affinity between these three populations would vary, just because the average produces a slight tendency, it may mean nothing, as it can be caused by random factors such as founder effects or secondary admixtures. I definitely would not build a whole theory of the deep paleohistory on slight differences of the average genetic distance. On the other hand, with haploid genetics you get a much better feeling of what was going on, not a purely numerical one but a (phylo-)logical and even "visual" one (if you put it in graph or maps form). I for one, I'm more logical and visual than mathematical and statistical -- I don't mean to deny value to these tools, but only when the figures become statistically significant, i.e. large enough to matter. It's not just an issue of approach it is also an issue of naturally occurring "noise", of the Chaos present in every single phenomenon that Benoit Mandelbrot and others discovered in the midst of the 20th century and that is not still sufficiently mainstream, in spite of being necessary for all kind of science and technology, from weather prediction to electronics. Random fluctuations do happen and they have unpredictable long term effects in what is called sensibility to initial conditions, alias "the butterfly effect". That's why weathermen can't tell us what weather will be a month from now: our ability to predict stuff in chaotic environments is null and ultimately all environments are chaotic in the long run; we can't even predict the behavior of planets in the very long run. Newton was wrong or at the very least not right enough.

      Delete
    25. ... so, what I meant to say, is that we cannot infer the past (just as we can't predict the future) in a mechanicist way with extreme precision: noise, randomness, hidden factors, do happen and did happen, affecting the outcome in ways we can't really expect to fully understand. Maybe in the future with a better understanding of the genome and the use of extremely powerful computers to calculate not just averages but organized, structured, comparisons, we might gain a better understanding.

      Delete
    26. Hrm. This is hard to discuss if we don't agree on the basics here.

      "Why do you say that. I have never seen anything like that: the South Asian "ASI" component is very specific of that subcontinent, distinctive, and only when a dualistic polarization is generated, quite artificially (or in modelings involving the controversial "basal eurasian" thing), such a tendency appears. But it is a mirage, not anything that can be accepted as realistic, just a product of too shallow improductive analysis."

      Would you agree that the Onge are the best living representative today of ASI? If so, I'd challenge you to find any tree or formal test that shows them clustering with West Eurasians over East Eurasians. As far as I can tell there aren't any. In fact, the #1 hit on Google Image Search for "treemix onge" is your blog btw. There's a reason the Onge are used as a proxy for "Eastern Non-African" by Lazaradis and others. I'd rather not continue down this line until we can agree on the general shape of human diversity. QpWave and D statistics would work if you don't like Treemix (though it's a very rigorous program).

      "The only "clade" that ASI and East Asians (or Papuans, Onge, etc.) form is one called "not West Eurasian". Otherwise they have no relation whatsoever. "

      Eastern non-African is the conventional term, but you're wrong in saying they share nothing in common. They share additional drift with one another. Any formal test shows that. I'll switch to this term to avoid confusion, but make no mistake - this is a clade.

      "K2b, specifically P1 sublineages Q and R1 are indeed West Eurasian lineages shared with South Asia (and in the case of Q also with Native Americans and some NE Asians, but only because it migrated from West Eurasia via Altai). Both Q and R1 must have coalesced in West Asia."

      I didn't say P1. I said K2b, and I meant just that - K2b. And yah, the post you linked of yours is what convinced me of this as well.

      K2b

      "
      That Vilabruna is the first KNOWN European R1b1 only underlines this, because it implies that its precursors MUST have been in Europe since at least Gravettian times (no West Asian immigration to Europe between Gravettian and Neolithic)."

      We can't say that unfortunately because Gravettians are not a valid genetic group, as mentioned above. If there were migrations of different peoples all using Gravettian tools we're pretty blind to that without genetics. And as mentioned before, Villabruna had clear East Asian (Han-related) admixture! So clearly some sort of East-to-West backmigration happened.

      Delete
    27. "So there's a whole story to Y-DNA K2 and mtDNA R (also N but very especially R) expanding from some chronological point (surely after the Toba catastrophe) from SE Asia in various directions. The western branch (P1) left a track in India but impacted much more clearly West Eurasia. However on the way they picked other Y-DNA lineages: IJ, G and LT (at least T)."

      Agreed. But IJ, G, LT (likely H as well) - this is the list I made of "West Eurasian haplogroups." It's the ones that were "along the way" and NOT in SE Asia pre-Toba. That's my point. These males with IJ, G and LT had some mtDNA as well. What haplogroup was it? And why weren't those haplogroups swept up too?

      Anyways - this dichotomy between K2 and F(xK) is what I'm get at here.

      I think whatever the story of K2 is, it's a different one from R and N. The distribution of them are very, very different. And N reached Europe tens of thousands of year before K2's descendants (~17,000 years based on that paper?). Obviously we've only seen so many samples, but so far I don't think there's much correspondence between K2b and any mitochondrial DNA.

      I'd suggest considering the possibility btw that haplogroup N and possibly haplogroup M were present on both sides of the Toba ash cloud.

      "Nope. Totally in disagreement: the steppes don't seem to have played any major role in the settling of Europe and even the Indoeuropean expansion in the Chalcolithic is tracked only to the West Eurasian (i.e. European) steppe, not further East."

      No, the Fu paper disproves this. Villabruna and some (not all) of his relatives had Han-related admixture to varying degrees. And I think it's hard to suggest how it could have got to Europe but not the Near East if not by a northern route.

      "The story of our migrations is tropical and subtropical until c. 50 Ka ago, even in China and Japan it's probable that they didn't settle in mildly cold areas until around that date (maybe a bit earlier?"

      Agreed completely.

      "but there is also a clear push towards the North, to places never before inhabited by our species like Altai, Japan, North China, etc. "

      Not just by us too. Denisovans and Neanderthals were headed north as well it seems.


      Delete
    28. "Not realistic: it does not even represent the "Basal Eurasian" migration edge."

      Take any other you prefer then. The Onge will always form a clade with Han, to the exclusion of West Eurasia.

      How many edges it shows is just a function of how many you tell it too though. That doesn't make it wrong.

      The relative position of the branch between West Eurasians and Eastern Non-Africans isn't relevant. Just that they form separate clades.

      "or in general the African-like genetic component, then West Eurasians should look as a simple branch of "macro-Asians", one among several. West Eurasians are the only Macro-Eurasian pop. with extra African (or similar) admixture, that is something we just cannot ignore."

      With respect, you're assuming an outcome here Maju. And Razib's tree that I posted earlier addresses your concerns and still shows what I'm talking about: http://www.unz.com/gnxp/the-ancient-scramble-for-eurasia/

      Recall too that this holds for samples that lacked Basal Eurasian too, like WHG.

      "And that's why autosomal DNA is not always the best guide in our understanding"

      Sample sizes distort the picture with uniparental markers a lot more. aDNA, mtDNA and Y-DNA don't lie. They just tell us different things. I'm a member of R1b and H clades that are very, very specific to Britain and Ireland. Looking at those two pieces of data alone misses the Slavic and West Asian ancestry I have. Neither is lying - they're just telling different stories. But aDNA is much, much less subject to drift.

      "After all autosomal DNA interpretation is a matter of statistics and you know what they say about statistics, right: "there are lies, damn lies and statistics". "

      People who say that really don't understand statistics IMHO.

      "nother possible cause of distortion can be the archaic admixture frequencies: it's not big but it creates an anti-African vector that pushes West Eurasians artificially towards Africans, in addition to the previous. In your tree, there is the Denisovan admix. axis but not the Neanderthal one, which we know is not equal everywhere. So it's quite probable that the question you are trying to answer is itself distorting the apparent answer of that tree. "

      I think the fact that the structure of the tree doesn't change when we add or subtract edges from Neanderthals and Denisovans tells us that the structure is mostly correct. It holds for West Eurasians with higher Neanderthal ancestry too like Bichon and other WHG. This doesn't fly.

      "Anyhow too many West Eurasians not to form a distinct clade in any algorithm vs all other macro-Asians. Sampling strategy matters too and, in doubt, various sampling strategies should be contrasted to see if they change the result and how."

      Huh?

      "S2- Out of the top of my head: is "Basal Eurasian" (in part, probably not in full, because we can't ignore E1b) a negative of the extra El Sidrón Neanderthal admixture now detected in Eastern macro-Asians (Han and Papuans at least)? I've never seen such a refined testing, so I would not discard it."

      Basal Eurasian is too big a component to be explained by archaics. We're talking about ~30% of the DNA of some samples. ~0.5-1% additional admixture can't rewrite another 29% of the genome. It has a bit of its own drift too.

      What I think the really interesting question is whether or not El Sidron admixture could be found in the Paleolithic European samples too like Goyet-Q116 and El Miron if we looked hard enough.


      Delete
    29. "MtDNA M1 however must have come with the group from SE Asia however (its relatives are all East Asian"

      Interestingly enough though, M1 seems to have entered Africa from the NW though. Possibly Iberia. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1945034/

      Agreed with respect to its deeper origins, but I wouldn't be at all surprised if we later found out that the Aurignacian M samples belong to M1.

      Delete
    30. Apologies for the walls of text btw.

      "
      "And IIRC haplogroup U peaks in southern tribals, and is lowest in eastern tribals, with northern tribal Indians intermediate. I don't think this relates to the current discussion in a huge way, but it's somewhat counterintuitive."

      Not sure why you say that, Metspalu 2004 is maybe one of the key references and U sublineages (U2 mostly) is rather scattered:"

      Sorry, I should have been clearer. I was refering to tribal groups only, not the general populace. Eastern tribals have the least U. Southern have the most. It might speak to a south Indian origin for U maybe?

      Delete
    31. Don't worry too much about the "word wall" because we do need to talk in some depth in order to understand each others (that's where Twitter totally fails). I also tend to that way too often: it's OK, unless it becomes a rant or an endless battle of "I'm right, you're wrong" and we are not there at all yet.

      "Would you agree that the Onge are the best living representative today of ASI?"

      No, not at all. The Andamansese are a distinct unique branch of the macro-Asian subset of Humankind, one that might have lesser admixture with South Asian or East Asian genetics but is only very thin and extremely old. Also there's nothing particularly "Indian" as in contrast to "East Asian" or "Papuan" or whatever else in them: they are absolutely unique, as can be all those other groups.

      ASI is probably best studied without the Andamanese, unless they are used as an outgroup and never as reference. In this aspect they are almost indistinct to using East Asians, Papuans, etc., even if each population has its own issues (some East Asian inflow into India, Denisovan admixture in Papuans and maybe some extremely ancient South Asian inflow into Andaman, almost indistinct from generic macro-Asian roots).

      The Andamans were never part of the continent but were instead a rather distant island and haploid genetics suggests links with both India and East Asia but all them extremely old in any case.

      "If so, I'd challenge you to find any tree or formal test that shows them clustering with West Eurasians over East Eurasians."

      That would not make sense. It's almost like asking the same for Dai, which have been used as control in some tests instead of Onge, when considering the issue of the "Basal Eurasian" (African-like) ghost component in West Eurasians. They produce the same results. Onge can be used for macro-Asian or whatever the name you want to call the mainline component of Eurasians and associated peoples of the Pacific but for Indian specifics it is irrelevant (and misleading if people are using them).

      I doubt there is any unmixed ANI populations alive but some tribals like the Irula are pretty close. Otherwise we would need ancient DNA. Also I doubt that ANI is monolithic: I would expect it to split in various groups instead, for example one for the South and another for the North of the subcontinent, maybe more (bear in mind that the Northern Deccan Plateau acted as semi-arid barrier, so the connection between the northern and southern parts of South Asia was almost restricted to the coasts in the Pleistocene, particularly the Western one). I think that South Asia deserves a much more detailed attention because the level of the research, both genetic and archaeological, is rather poor and yet it is one of the most crucial regions in the paleohistory of Eurasia.

      ...

      Delete
    32. ...

      "Eastern non-African is the conventional term"...

      I coin my own terms when I don't like the ones served to me, when I feel they are tendentious or otherwise poor: "macro-Asian" (or "macro-Eurasian" or just "mainline Eurasian" if you wish) is fine. "Basal Eurasian" still looks African and it is in fact probably African in part if not all, and is not in any case the main component in West Eurasians, just a minor one. Notice that there are many Africans (Northern and Horner ones) that are largely or even mainly "Eastern non-African" in your terminology, so not the best term. It would be like calling "Western non-Indian" to ANI, what is a nonsense when it is one of the main components of Indian genetics.

      Also "Eastern Eurasian" implies some sort of exclusive cladistic between (1) ASI (aboriginal or pre-Neolithic South Asian), Andamanese, the main East Asian component and the various Australasian ones, vs. (2) West Eurasians. This is false and only may appear as such because West Eurasians have some African-like admixture that the rest lack. Once we distill that secondary admixture component out, they are one and the same.

      "... you're wrong in saying they share nothing in common. They share additional drift with one another. Any formal test shows that."

      That's not what I know or think I know. What I have read in studies like the one by Lazaridis, which raised the issue, show that the main component by far in West Eurasians is the mainline macro-Asian one (what you call East Eurasian) and that, in addition to that they have some admixture that is African-like but not "L0" (Khoisan) nor "L1" (Pygmies), nor seemingly "L2" (Yoruba), so it is necessarily within the populations dominated by "L3(xM,N)", which are either strictly African (Dinka-like) or something now extinct as such distinct population that lurked in non-Neanderthal West Asia and has left few (but not zero) remnants. IMO it is a bit of both, there are indications supporting the NE African element (notably Y-DNA E1b) but also some suggesting an OoA remnant in parts of West Asia, now hard to detect but maybe apparent in some L3 and L0 lineages that seem old in Arabia (notably Yemen) and some autosomal component that I think I detected in an ADMIXTURE exercise, in wait of independent confirmation. It needs more research in any case.

      Note: I reckon that using mtDNA labels is simplistic but also helpful because they illustrate the successive branching, always approximate and subject to remix, of Homo sapiens. I know better than that but space and time are limited, hope I'm convening the correct notions.

      ...

      Delete
    33. ... ... "this is a clade".

      It is not and cannot be a clade vs. West Eurasians, because the vast majority of West Eurasian ancestry lays in "Eastern Eurasian" (in your terminology) and not in "Basal Eurasian", which is just a lesser extra. You are confusing admixture with clade, it's like saying that African Americans are a clade vs mainline Africans just because they carry some Europan admixture: they are not, they are an admixed subset (so just a subset once admixture is removed). In this case we know which is the origin of that admixture because it is very recent and the source population still exists, but in the case of West Eurasians the issue is blurrier, particularly because some insist on not comparing with Dinka (on false claims that they might be significantly admixed with the Eurasian backflow) or do not search the depths of the West Asian genetics in search of that true Basal Eurasian component, which IMO can be found in some Arabs, just as some "Aterian" component (which would also appear as "true Basal Eurasian" even if it never left Africa and is probably different from other co-factors behind this "ghost") appears to be lurking in South Moroccans at frequencies of c. 14%.

      Delete
    34. "Gravettians are not a valid genetic group"

      Gravettian is an intrusive culture and population in its core area of Moravia and surroundings (probably from West Asia ultimately see this). From there it spread to West and East, being, along with the Aurignacian "proper" that preceded it by some 10,000 years, the only pan-European culture of the Upper Paleolithic, Epipaleolithic and even later on. It's possible and I'd say even probable that it admixed in its expansion with the precursors populations (both the "Aurignacian proper" layer and the various "Aurignacoid" cultures that preceded this one) so we have three layers:

      (1) earliest UP or "Aurignacoid" (various cultures, maybe diverse genetics but in general derived from West and Central Asia, and ultimately from India and beyond).

      (2) Aurignacian proper: one of those cultures, located most clearly in Istallosko (NE Hungary) since c. 48 Ka ago, expanded through all Europe around 40 Ka ago.

      (3) Gravettian: a new culture, intrusive initially in Moravia, with plausible West Asian origins, since c. 32 Ka ago, also expanded through all Europe and even into Siberia (Mal'ta is culturally Gravettian or at least Gravettian-influenced) in the subsequent many millennia (in Iberia is just a few millennia older than Solutrean, so it was no "blitz").

      Then came the LGM and all new cultures are quite clearly European at the origin, as the Franco-Cantabrian region became a major cultural engine on its own, producing first the Solutrean (expansion into Iberia south of the Ebro, North Africa if we accept that Iberomaurusian is derived and also later into Hungary, producing an Epipaleolithic offshoot in Poland) and then the Magdalenian (expansion in all West and Central Europe, reaching again to Hungary). The Franco-Cantabrian engine would still produce the Magdalenian-derived Azilian in the early Epipaleolithic (expansion to Iberia mainly, where it's known as "laminar microlithism") but then the epicenter seems to shift to North France and the Rhine area, with the (also Magdalenian-derived) Tardenoisian culture expanding in most of the previously Magdalenian area (in Iberia known as "geometric microlithism"). There are other groups that we should not ignore in North Germany/Denmark and various Epigravettian groups from Italy to Russia but space and time are limited. In any case we see first arrivals from West Asia before the LGM (two waves surely) and then secondary expansions from within Europe... until the Neolithic changes things again with a new West Asian immigrant wave. So no wonder that Paleoeuropeans are so different from West Asians and the main EEF component, at least 24,000 years of separate drift mark the difference; while there was no Bosporus back then, the Balcans and West Anatolia were so sparsely populated that acted as a quite effective barrier (or buffer, or filter, if you wish) as long as no major migration took place.

      We don't know yet how much Gravettians were distinct from the previous Aurignacian/Aurignacoid layer, and how they mixed with each other in each region, but everything points that these were two distinct immigration waves that must have brought with them distinct genetics within the wider West Eurasian set. So when I say "at least Gravettian" I mean that I don't know if R1b (and other possibly related genetics like mtDNA H) arrived to Europe with Gravettian or the previous Aurignacoid wave but that it is definitely not a "Solutrean" or "Magdalenian" arrival, because these cultures were strictly home-grown in Europe itself (in fact Magdalenian can be described as a "refined Aurignacian"). There are no possible immigrations between the LGM and the Neolithic.

      Delete
    35. "Villabruna had clear East Asian (Han-related) admixture!"

      I'm not aware of that. Can you reference your claim a bit more precisely, please?

      Delete
    36. "And Razib's tree that I posted earlier addresses your concerns and still shows what I'm talking about: http://www.unz.com/gnxp/the-ancient-scramble-for-eurasia/"

      Well, I do not trust too much TreeMix: it produces sometimes contradictory results, so I take such stuff with a grain of salt, more so when the West Eurasian sample is almost as large as the rest-of-Eurasia one: sample biases always tend to favor the distinctiveness of the over-represented group in ALL statistical analyses of autosomal DNA. So I'd suggest to try it again using only one West Eurasian sample instead and making sure that East African sample are are effectively unadmixed "L3" Dinkas, for example.

      That tree looks extremely odd in fact, because it shows arrows from East Africa to nearly non-West Eurasian macro-Asians. That is necessarily an artifact and almost certainly makes up for missing stuff in the main "black" tree. I would consider that "evidence" rubbish ("junk in, junk out") but it's true that I should bother countering it with my own produced counter evidence, yet I have not the technical skills nor the willpower to bother anymore, so I will politely disagree and leave it at that.

      Incidentally I don't follow Razib any more because he had a too long period (not sure if it continues) in which his entries were like 80% extreme right-wing politics, which is too much of a toll for me, sincerely.

      "Agreed. But IJ, G, LT (likely H as well) - this is the list I made of "West Eurasian haplogroups." It's the ones that were "along the way" and NOT in SE Asia pre-Toba."

      With doubts about G, all them are of South Asian roots. You have to follow the phylogenetic tree in full and F, as well as its descendants LT and IJK, all look like originating in South Asia (to the NW maybe but not further West in any case). You should also consider the mtDNA tree and its "geostructure" for a complementary understanding of the ancient migrantions, I find it even more clear. Anyhow, focusing on Y-DNA, CF might have coalesced along the OoA pseudo-coastal (tropical/subtropical) migration route, maybe in South Asia already, F did as well as most of its descendant branches, although a few seem to have reached SE Asia, notably K2 (of course). C and D probably coalesced in SE Asia. The K2-derived P1 and also the intriguing but now ultra-rare Western C1 lineage imply a backward migration to the West, via North India, culminating in the "conquest of the Neanderlands" (or Upper Paleolithic settlement of West and Central Asia, as well as Europe and NE Africa). This is also traceable in the mtDNA, although there are some differences, notably the incorporation of Y-DNA IJ and T (and possibly "pre-G" as well). This may have been caused by either matrilocality of the migrating "nation" or, alternatively by the formation of a new, admixed, "nation" in the NW edges of South Asia prior to the push into "the Neanderlands". We just don't know enough to discern the how and we may never know.

      ...

      Delete
    37. ...

      "I think whatever the story of K2 is, it's a different one from R and N. The distribution of them are very, very different".

      Actually they are extremely parallel, notably mtDNA R (but also carrying along some other N lineages): we see it in Papua (K2a + Q), in East Asia (NO + B, E, F, etc.), in South Asia (P1*, R2 + various R) and in West Eurasia (R1, Q + R0, U, JT, etc.) The overall genetics are more complex (people mix and remix every generation) but still the pattern is very clear and overlaps way too well.

      "And N reached Europe tens of thousands of year before K2's descendants (~17,000 years based on that paper?)"

      I don't think that molecular-clock-o-logy helps at all, for me it's rather a pseudoscience, a burning nail with no strong support, it's almost like saying "the Bible says..." and has the same kind of scholastic reasoning behind: appeal to authority and tradition rather than to hard science. At the very lest it's worthless when it contradicts other data, particularly the archaeological record, which I discussed in a previous comment for the case of Europe.

      But in this case it also contradicts another key evidence: the important lineage Q, which MUST have reached Altai in the early Upper Paleolithic (probably c. 47 Ka ago) in order to have then migrated eastward (as Q1) with this techno-culture into Mongolia and North China, reaching Beringia maybe in the LGM and America proper at least 17 Ka ago. Q clearly originates in Iran or some nearby area, so there is no way to disdain, to negate that P1-derived lineages were in West Asia (and possibly also in Europe) at the very beginning of the Upper Paleolithic. You may want to argue when did it reach Europe as such (I already did before in two comments, and IMO it must be much older than a mere 17 Ka) but West Asia must have hosted both Q and R1 since the very early Upper Paleolithic NECESSARILY.

      Delete
    38. "Basal Eurasian is too big a component to be explained by archaics. We're talking about ~30% of the DNA of some samples".

      I don't think so. I wonder what samples you have in mind, maybe Levant-Mesolithic? These should be quite African-admixed but may also carry some significant local "true BA", nowadays more diluted. I really wish there was more research on Palestinian genetics, because they are without doubt a key population.

      Or are you thinking in the CHG anomaly? Those were quite heavily BA apparently... but only per one study, so I'd like to know more, with more contrast, with more quality "second opinions".

      In any case, I never meant that the archaic extra admixture is all the causes of this issue, not at all: it's a complex one with many facets. What I meant is that a, say, extra 1 percentual point of Neanderthal admixture in Han may cause a (just a hunch) 4% appearance of divergence, because Neanderthals are so much more drifted apart from us that their segments would look much more different. But there are other sources of this effect: actual African admixture and probably also "true BA" admixture.

      We weren't anyhow talking of CHG but of modern West Eurasians, Europeans specifically, whose percentage of "BA" (African or true BA) is much much smaller than your "30%". Not sure but maybe 5-10%?

      And in any case it is an admixture, not the main component: we Europeans are 90% or more Han-like or Onge-like, forming a clade with then just as African Americans do with Yoruba. The only difference is that these issues are much older and we lack clear sources of that BA admixture to compare with (they are only inferred by hard systematic work like that of Lazaridis and others). It's a bit like inferring the percentage of archaic admixture (not Neanderthal nor Denisovan but African archaics) in Africans: we have some hints in their DNA but we cannot say much more until more data and more research is available.

      Delete
    39. "M1 seems to have entered Africa from the NW"

      I did not know of that paper. I have to read it in full yet but looks quite intriguing, really, especially on light of the latest findings re. U6.

      However I'm admittedly reluctant to admit that, considering its clearly East African spread, somewhat parallel to that of Y-DNA T. The paper is rather old (2007), so I wonder if there has been newer research that confirms or contradicts that.


      "Sorry, I should have been clearer. I was refering to tribal groups only, not the general populace. Eastern tribals have the least U. Southern have the most. It might speak to a south Indian origin for U maybe?"

      Diversity, basal or phylogenetically sorted diversity, is what matters, not frequency, which is often misleading. And AFAIK, most mtDNA U basal subclades are absent or very rare in South Asia: U1, U5 and U6 specifically (add U* if you find it). Only sublineages of U2'3'4'7'8'9, are present in India at significant and likely old dates, notably U2 and U7. So almost certainly they represent an old backflow from West Asia, surely at the early Upper Paleolithic, which in India is more recent than in West/Central Asia, Europe and NE Africa (LSA). We know that U2 has been present in Europe since at least Gravettian (Kostenki), while U7 has a more "coy" distribution concentrated in the NW.

      Incidentally, when I wrote back in the day about that U2 discovery in Kostenki, illustrating it with a quality reconstruction of the head, someone from South India thought he looked rather like people he knew. Although it's not like Indians and Europeans are that different in looks once we take out the pigmentation; I was illustrated of that by a picture of two Tamil cousins, one of them albino, who looked almost totally like a very pale Viking (he even sported a beard, something rather unusual in Indians, a light yellow beard of course).

      This also applies to some extent to Australian Aborigines in fact, with many cases of albino Aborigines being taken from their parents in times more racist but not so distant from ours, because the authorities believed they were kidnapped white children or something. If we take off pigmentation, we can really think of an Australo-European phenotype continuum, which forms "a clade" (in looks only) vs East Asians, particularly those towards the North. I don't know exactly why does this happens but it's something that intrigues me since long ago.

      Delete
    40. "I don't think that molecular-clock-o-logy helps at all, for me it's rather a pseudoscience, a burning nail with no strong support, it's almost like saying "the Bible says..." and has the same kind of scholastic reasoning behind: appeal to authority and tradition rather than to hard science. At the very lest it's worthless when it contradicts other data, particularly the archaeological record, which I discussed in a previous comment for the case of Europe."

      I'm not talking about the molecular clock numbers here. I mean actual samples we have. There's no K2b in Europe until Villabruna (at least not West of Russia). N is there in our earliest samples.

      Delete
    41. "It is not and cannot be a clade vs. West Eurasians, because the vast majority of West Eurasian ancestry lays in "Eastern Eurasian" (in your terminology) and not in "Basal Eurasian", which is just a lesser extra."

      You're the one confused here I'm afraid. There's West Eurasian that does not have anything to do with Basal Eurasian. Modern Europeans are West Eurasian + Basal Eurasian. This is commonly accepted and obvious from looking at any tree. Basal Eurasian shows up as admixture. The West Eurasian clade is no more or less basal than East Eurasian.

      Delete
    42. "There's no K2b in Europe until Villabruna (at least not West of Russia)".

      So what? Have you even considered the geographic and sample-size limitations of the research on ancient Y-DNA, particularly for the Paleolithic? There is only one other Y-DNA sequence from Italy, from Paglicci in South Italy, only two sequences from Western Europe (both from Belgium, which in that period belongs mostly to the Central European or Rhine-Danube province) and absolutely no sequences from the region most densely populated by far since the LGM (SW Europe). So we have huge gaps in our understanding: what we know is interesting but we need to know more, much more before we can jump to conclusions as you try to do.

      The key issues here are:

      1. LACK OF EVIDENCE IS NOT EVIDENCE OF LACK: that we haven't yet found something does not prove it was not there, just that we need to keep searching.

      2. The Villabruna date of 14 Ka BP is a TERMINUS ANTE QUEM (i.e. a minimally old date) for the presence of R1b1 in Europe (and anywhere on Earth as far as we know). A datum or factoid that must be integrated with the rest of our knowledge: the "geostructure" of R1b (inferred from modern hierarchical diversity) and the sequence of our Prehistory reconstructed on archaeological grounds. So, unless you can prove otherwise, R1b MUST have been in parts of Europe since the Gravettian expansion or earlier, because there's no logical way it came from outside Europe after the Gravettian intrusion. This also implies that R1b should have first scattered (not just to Europe but also to Africa, Central Asia, etc.) in the pre-LGM period. This fits well with my theory of P1 (the Western subclade of K2) migrating Westward with the first UP peoples, which I have already exposed in detail and is not just linked to the endeavors of R1b but also to Q, R2, etc.

      Of course, more data is needed to confirm this theory but with what we have now, I cannot even conceive how it could be demonstrated false, really. I have been proven wrong sometimes, so I reckon I must be a bit cautious here, but I have also been proven right many more times, so you should be cautious too.

      You seem to be spousing the Indoeuropeanist hypothesis about the origins of European R1b, which just don't fit the available data in any way. It's basically a junk-in, junk-out "theory" based only on the fact that Mal'ta boy carried R* (why wouldn't he and why would that "prove" anything) and of Yamna peoples carrying a local variant of R1b, not particularly related to those of Western Europe (again proof of nothing). On the opposite side we have a very strong piece of evidence in the "geostructure" of R1b and particularly Western European one, which is very clearly centered in South France (S116) and somewhere around the Netherlands (U106), and not at all towards Eastern or even Central Europe. Plus the major issue of massive R1b density (and rather high diversity) among Basques, a non-IE people. Ancient DNA evidence, still very limited, suggests so far expansion associated to Bell Beaker, which is a cultural "phenomenon" originated in SW Europe in fact, and not to Corded Ware, which is the intrusive Indoeuropean/Kurgan culture. Of course too many datapoints from Central-North Europe and too few from Western-South Europe don't help to solve the riddle but probably just cause increased confusion, however wishful thinking is also a cause of confusion here: "I want to be Indoeuropean by patrilineage, so I will fight for it no matter what". I guess it is a motivation but in what regards to the scientific method, it's nothing but trash.

      Delete
    43. PS- Notice how that wishful thinking "reasoning" can also be read as "I want to belong to the caste of the winners and not to that of the historical losers". I don't need to explain why this is pathetic, do I?

      Delete
    44. "You're the one confused here I'm afraid. There's West Eurasian that does not have anything to do with Basal Eurasian. Modern Europeans are West Eurasian + Basal Eurasian".

      I understand that well but when we compare modern peoples, we are bound to work with the already consolidated mixture. So "ancient pure West Eurasian" is not the same at all as "modern admixed West Eurasian", say French, Sardinian or CEU, the typically used samples. We would need to first "distill" the AWE ("ancient West Eurasian") component, what is quite complicated, because there are a number of confounding factors at play, not just BA.

      "The West Eurasian clade is no more or less basal than East Eurasian".

      Modern West Eurasians are all much more "East Eurasian" (macro-Asian) than BA, even Palestinians and peninsular Arabs, even EEF were that. This may be a key reason why we cannot agree: you seem to imagine a much greater impact of the BA ghostly component than research actually allows for.

      Delete
    45. You say modern West Eurasians are "East Eurasian" but that´s not what the researchers found. Instead they found Palaeoeuropeans and what they call "ANE" to be closer to each other than to other Eurasians, and said other Eurasians (what they called "East Eurasians") to be closer to each other than to any of Palaeoeuropeans or ANE.

      Delete
    46. ANE is closest to Native Americans and only secondarily to SOME Paleoeuropeans (EHG but not WHG) or modern NE Europeans. This makes total sense if we see Mal'ta for what it is: a Gravettian Central Asian, i.e. almost certainly a mix of (a) the first Central Asian settlers of Altai (which did not originate in Europe but in Southern Central Asia, with blurry inclusion of Pakistan and Iran, etc., which is surely the main origin of Paleoeuropeans and all other Paleo-Westerners) and (b) secondary arrivals from Gravettian Eastern Europe. So maybe it's not so much that there is ANE in EHG but that there is pre-EHG in ANE, plus the blurry issue of shared West-Central Asian ultimate origins of both).

      In any case, AFAIK, ANE does not show any significant "Basal Eurasian", so they'd be within what you call "East Eurasian" and I dub "macro-Asian" or "mainline Asian", right? Just like the Onge.

      Let's be clear on how I envision this scatter:

      1. Macro-Asian root (main Indian and SE Asian macro-population).

      1.1. West Eurasian root (early UP settlers arriving via India and Pakistan, probably via Southern Central Asia with a key site in Uzbekistan) and scattering into two primary branches:

      1.1.1. First Central Asian or First Siberian, leading partly to Native Americans via UP spread into Mongolia and surroundings c. 30 Ka BP (and obvious sex-biased massive admixture with other "macro-Asians" that had already expanded northwards in China, etc.) This branch is proto-ANE but not yet ANE itself as defined by Ma1, which is a more recent fossil.

      1.1.2. Mainline West Eurasians expanding into West Asia and Europe (via Turkey and the Balcans, with key secondary core in Central Europe at first). We can differentiate these into:

      1.1.2.1. Ancient West Asians, some of which obviously admixed with something we call "Basal Eurasian", which is African-like genetics but not quite Yoruba or Pigmy, rather something Dinka-like, or maybe just remnants of the OoA persisting in the Persian Gulf marshy "oasis" (??).

      1.1.2.2. Paleoeuropeans, of which we distinguish WHG, EHG and some unclassifiable individuals. They don't seem to carry "BA". They may have arrived from West Asia in two waves: first the "proto-Aurignacian" one c. 50 Ka BP and later the Gravettian one c. 32 Ka BP. As Gravettian expanded, rather slowly but persistently, to all Europe it also reached Central Siberia, influencing in the Mal'ta cultural group in obvious archaeological terms and quite possibly also genetic ones, producing what we call "ANE", which is almost certainly a misleading category, as it is surely NOT directly ancestral to anything but Central Asians and West Siberians. It's quite possible that what we see in "ANE" is largely just proto-EHG and/or other "root West Eurasian" stuff, for which we lack better proxies as of now (but will hopefully be sequenced in the future).

      Delete
    47. PS- I said that the early colonization of Europe by H. sapiens happened via Turkey and the Balcans. This is correct for the greatest part but we cannot ignore that there was probably a minor route into Eastern Europe via the Caucasus, which might explain some of the WHG/EHG differences better than ANE. Sequencies from Mezmaiskaya, Uzbekistan, Iran, Pakistan... with enough time depth would help a lot here but we must understand that they are hard to produce (for reasons of preservation maybe but also of availability and interest, Eurocentrism and in general local-centrism is still too strong in the Academy).

      Delete
  2. I'd add that part of my N/M hunch has to do with Fu et al 2016. If you haven't read it (or rather the supplementary info that's freely available, which is all I've seen) - you will absolutely love it. The oldest WHG sample is Y-DNA R1b in Italy and there's inflated Upper Paleolithic continuity in the Basque country going back to the Aurignacian.

    http://www.nature.com/nature/journal/v534/n7606/full/nature17993.html

    With David's take:

    http://eurogenes.blogspot.ca/2016/05/the-genetic-history-of-ice-age-europe.html
    http://eurogenes.blogspot.ca/2016/05/on-modern-genetic-affinities-of-ice-age.html
    http://eurogenes.blogspot.ca/2016/05/villabruna-cluster-near-eastern-migrants.html

    The paper shows Neanderthal ancestry decreasing over time, and suggests this is solely due to natural selection. At the same time though, we see a population that is initially Y-DNA C1 and a mix of mtDNA M and N. Then we see successive waves gradually transform Europeans into a population mostly dominated by just Y-DNA F and mtDNA N and with increased autosomal affinity to modern Near Eastern populations.

    So it strikes me that in addition to selection, this population turnover could be partly responsible for decreasing Neanderthal ancestry too. The implication of that would be the existence of an ancient population that was rich in both Y-DNA F and mtDNA N, and poor in Neanderthal ancestry. It would also imply that at the LGM, what we consider "west Eurasian" today probably had a very limited geographic range (maybe from Pakistan to the Balkans, if that).

    I'd also point out that their trend line for decreasing Neanderthal ancestry is a rather poor fit:

    http://www.nature.com/nature/journal/v534/n7606/fig_tab/nature17993_SF1.html

    It seems more like two distinct steps downward - one with the appearance of the Villabruna/WHG group, and another with the spread of agriculture to Europe (bringing with it "basal Eurasian").

    It's a weak fit I know. Just a hunch.

    ReplyDelete
    Replies
    1. I did mention "the Vilabruna paper", as you called it, here: http://forwhattheywereweare.blogspot.com/2016/05/large-paleoeuropean-dna-survey.html

      I must advance that there are many caveats to make to your interpretation, partly because we cannot consider all Paleolithic Europeans to be a single homogeneous group and also because there is absolutely no reason to imagine an LGM immigration from West Asia. Actually the immigrations from West Asia happened before the LGM and then nothing of the like happened again until Neolithic, at least as far as we know from the Archaeological record, by which, in Europe we have the following cultural sequence:

      1. An earliest UP period beginning c. 49 Ka BP with diverse techs that are usually attributed to Homo sapiens recently arrived with that "mode 4" or "blade" tech from West Asia and beyond (the exact origin of the "UP revolution" is unknown, David Sánchez proposed Uzbekistan as most likely within the data we have).

      2. Aurignacian expansion c. 41 Ka BP (origin in Hungary, where Aurignacian proper dates to c. 48 Ka BP), roughly coincident with the HE4 cold spell and the Campanian ignimbrite supervolcano.

      3. Gravettian expansion since c. 32 Ka BP (in Central Europe, much later in some peripheral areas like Iberia). This is a clearly intrusive group in its core area of Moravia with not fully clear origins but most likely from West Asia (technically close to Levant techs considered likely precursors). In Eastern Europe and Italy it remained put, evolving locally, until Neolithic, being known as "Epi-Gravettian".

      4. Solutrean (since c. 21 Ka ago): out-of-nowhere new tech, developed in the LGM in Dordogne (Aquitaine) as far as we can tell and expanding primarily into Iberia (too cold to go north, although it seems to influenced Hungary later and Poland already in the Epipaleolithic).

      5. Magdalenian (since c. 17 Ka ago): plausibly evolved from Aurignacian remnants, maybe refuged in the Low Countries (?), it also expanded from the hyper-dynamic Dordogne or Perigord district of Aquitaine, the "metropolis" of the Ice Age (sported a population density, judging on findings, many many times anywhere else for some reason, even as much as four times denser than other dense regions in the quite dense Franco-Cantabrian province).

      So if we are looking for the source of demographic or genetic changes in UP Europe since the LGM, the logical place to look at is not West Asia (isolated from Europe or almost so in that period) but Aquitaine. In fact we can almost perfectly track a Solutrean expansion from Aquitaine to Iberia and from Iberia to North Africa (Iberomaurusian), we can also track a Magdalenian expansion later on from Aquitaine into much of Europe.

      ...

      Delete
    2. ...

      And it is since Solutrean where we see a SUDDEN decrease of the Neanderthal fraction per Fu 2016: in the relevant graph, we see how at c. 20 Ka BP two samples (both from Siberia) are still within the 3% Neanderthal ancestry range but that the sample from Europe has already transitioned to the 2.5% fraction that dominates thereafter (until Neolithic reduced it a bit more). This key sample is El Mirón (Cantabria) and is Solutrean (not the same sample as Hervella's Magdalenian one from the same cave).

      In Fu's paper "El Mirón" is used as label for the "green" group, however that sample actually is equally related to the "green" (older, maybe Solutrean at the root, although it also includes Central European Magdalenians) and the "red" or "Vilabruna" (more recent and common, plausibly more tightly Magdalenian-related, although it also includes Italian Epigravettian samples).

      The smoothing of the decrease in Neanderthal is mostly caused by the (quite confusing) inclusion of Siberian samples without properly labeling them, as if they were European. It's possible that some European samples also fit that (I'd have to find the data sample by sample in order to be fully sure) but it's in any case clear that the Siberian samples are distorting the graph way too much, creating an appearance of smooth transition when it's a sudden step at the LGM instead, surely caused by the change from Central European and Gravettian centrality (before the LGM) to a SW European and Solutrean/Magdalenian centrality (after it). The less-Neanderthal admixture thing is therefore probably something that is rooted in the patchy pre-Aurignacian settlement of Europe by our kin or, alternatively, in the Aurignacian proper itself (although Goyet seems to be evidence against this last interpretation, it is a single data point).

      "It seems more like two distinct steps downward"...

      Yes absolutely, I see that you noticed that. And once you exclude the "Siberian traps" (I mean the very questionably included Siberian samples), it becomes much more clear: it's a LGM or Solutrean cause what is acting here. But, I insist, there's absolutely no reason to think it comes from West Asia, all the opposite (unless you or someone can prove all our archaeological knowledge wrong somehow, of course).

      Delete
    3. Agreed re: El Miron. The low number of SNPs makes it a bit noisy, but I think you're right that the decline started there. And as you point out - El Miron was almost as closely related to the Vilabruna cluster as to the Aurignacian samples. I think it would be reasonable to think that that Vilabruna-like side is where the decrease in Neanderthal ancestry came from.

      Delete
    4. I wrote on it in the day: "El Mirón is actually green-red admixed", i.e. it is "Mirón-Vilabruna" per the labels they use. I don't recall well why I said that but it is probably because of the heatmap, where it's apparent that El Mirón has exactly the same shade of orange to other "El Mirón" or "green" group (a Rhine-Seine area group) than to most individuals in the "Vilabruna" or "red" group. This is true also in the PCA (also in the link above) They should have chosen some other label for the "green" cluster really.

      It is not related to the "Aurignacian" samples (of which we only know one: Goyet, all the rest of the "blue" cluster is actually Gravettian). Colors-coding per fig 1.

      Back in the day I carefully added the dates of each sample to the heatmap and it's apparent that the "Vilabruna"/"red" cluster is not even Magdalenian but basically Epipaleolithic with some older precursors: Vilabruna itself (Italian epigravettian), Rochedan (French epipaleolithic, of Magdalenian roots in principle) and El Mirón itself (Iberian Magdalenian). Most Magdalenian samples are in fact in the El Mirón (green) cluster, so one of the unexpected things here is some striking and unexpected disruption in population genetics between Magdalenian and the Epipaleolithic, most of which is considered rooted in Magdalenian.

      What happened? I can't say but the distinctive clustering is very apparent. There were several Epipaleolithic cultures and, in the case of SW Europe there are two time fractions: first the Azilian or "microlinear" one and later the Tardenoisian or "geometric", which originated in the north of France and Belgium.

      I'm hunching that a good label for the "red" group other than "Vilabruna" could be "Tardenoisian", as most samples seem for that group and time. On the other hand a good label for the "Mirón" could be "Magdalenian" (this would be a very good label indeed).

      So one thing we learn here is that the Tardenoisian peoples of the late Epipaleolithic were not Epi-Magdalenian but something else, at least genetically. El Mirón is intermediate between both groups.

      Delete
    5. I would caution against going by the archaeological cultures as one finding of this paper is that Gravettians don't form their own clade.

      If you go down to the trees in the paper that's what shows El Miron as mixed with something on Villabruna's branch. Whatever that is I think is the population that dragged down Neanderthal ancestry. IMHO a Balkan population with Steppe and East Asian admixture.

      Delete
    6. I accept the caution but the question is how to analyze things: archaeological cultures (with the best ideal level of nuance if you wish) are the material expression of cultural kin, not necessarily genetic kin, as people do change cultures and even kinship, via adoption. And by adoption I don't just mean adoption of children but mainly adoption of culture by alliance (or in later times also servitude).

      We do see a Gravettian cluster (blue) just that some details are unexpected, notably that Goyet Q53-1 is included by the authors in it. However in the heat map the distance is very large. We actually see the following serious clusters:

      1. Vestonice 13, 15 & 16, Krems as "core Gravettian" and then Ostuni 1, Vestonice 43 and Pavlov 1 as relatives. This is a very clear Gravettian cluster, it's not all the "blue" however. It also has a distant relation to most other Paleolithic Europeans (it seems ancestral in greater or lower fraction).

      2. Karelia, Afontova Gora, Mal'ta - this is an "ANE" or Paleosiberian cluster, which is also Gravettian incidentally but distinct nevertheless.

      3. Goyet Q2, Rignet 1, Burkhardsthohle and Hohle Fels as "core Rhine-Meuse Magdalenian" or "green cluster". El Mirón fits best here but is also related more weakly to cluster 1 ("core Gravettian") and cluster 4 ("red"). Goyet Q116-1 fits best here, at the distance expected for an Aurignacian, although the authors labeled it "blue" (Gravettian-like) for no reason I can find: its smallest distance is to Burkhardtshohle. This seems to indicate an Aurignacian-Magdalenian genetic link that is stronger than the Gravettian-Magdalenian one.

      4. The "Vilabruna cluster" where Vilabruna is not really the core are Rochedane and Ranchot 88, very especially, with Lochsbour, Bichon, Chaudardes and Berry-au-Bac also strongly near the core. Vilabruna is in but already somewhat in the periphery, suggesting admixture with Paleo-Italians of Epigravettian culture. El Mirón is also at the periphery of this cluster, a bit more distant than Vilabruna but along other "red" samples like La Braña 1, Hungarian KO1, Falkenstein and Les Closeaux. Based on dates, I would tentatively call this cluster "Tardenoisian" and it seems to relate best to the Magdalenian cluster #3.

      A clear missing link in all this is Aquitaine, a big blank blob in the sample but absolutely central in European Paleolithic since the LGM (and also important before). As always there's too much focus on Central Europe, in this case including North France, what does not help too much.

      In any case we can track a tentative pattern:

      1. Aurignacian (Goyet) and Core Gravettian are not relatives (in the "red" zone of the heatmap). Siberian Gravettian ("ANE") is not relative to either as well, but it is strongly related to Epipaleolithic Karelia.

      2. Magdalenians were distantly related to Aurignacian Goyet and only distantly (if at all) to "core Gravettians".

      3. Tardenoisians are probably derived from Magdalenians.

      All this fits well with the archaeological record we know of. We just need to be careful enough with heatmap measures while keeping an open mind.

      "shows El Miron as mixed with something on Villabruna's branch."

      It shows it as almost equidistant (although closer to Rhine-Meuse Magdalenian) not as "mixed". What the heatmap measures is genetic distance, not admixture (not directly). The problem is that south of the Loire (the most important middle and late UP region) there are only two samples: El Mirón and La Braña 1, three if you include one from the Rhône area. We're getting a good idea of the Central European sequence but not really of most other areas.

      Delete
    7. Erratum: "2. Magdalenians were distantly related to Aurignacian Goyet and only distantly (if at all) to "core Gravettians"".

      ... should probably read: "2. Magdalenians were distantly related to Aurignacian Goyet and only very very distantly (if at all) to "core Gravettians"". Else it's not clear that Aurignacian Goyet is closer to Magdalenians than any Gravettian sample.

      Delete
    8. "It shows it as almost equidistant (although closer to Rhine-Meuse Magdalenian) not as "mixed"."

      See figure 6.5 on page 51. I think it mostly agrees with what you're saying, but makes it far clearer to look at. Node nf1f1 is where I think the reduction in Neanderthal ancestry occurred.

      Delete
    9. I'll take your word for it, as I don't feel like double checking right now. My question would be: is it possible that the reduction of Neanderthal ancestry is not caused by "selection" but rather by differential origins, particularly thinking in Aquitaine, where all the post-LGM Paleolithic is centered in the case of Europe? Or in other words: I find easy to imagine that the "Franco-Cantabrian" populations had since the beginning less Neanderthal admixture than those of Central Europe and that became the rule later on because the latter went almost extinct, while the former expanded instead. Of course we would need samples from the Franco-Cantabrian region, particularly for pre-LGM dates, to be sure.

      Delete
    10. I don't think your scenario can be excluded, but given that Villabruna is the first sample we have with this lower (~2.5%) Neanderthal ancestry, I think it's more plausible that El-Miron's less drastic loss of archaic ancestry comes from the Villabruna side. Not exactly a lot of datapoints to work with here though.

      Delete
    11. "I don't think your scenario can be excluded"...

      I can't ask for more. We really need more data to be able to be more precise. It's a bit like trying to understand dark matter! XD

      Delete
  3. On the Kuhlwilm paper though, note that their best guess as to which population contributed to Altai was one that diverged from other AMH at or before the time that San people diverged from other AMH, with strong evidence that it was at least before the split between Eurasians and Africans.

    So it wouldn't be consistent for this to be part of the main Out-of-Africa migration.

    ReplyDelete
    Replies
    1. I don't read it that way: they find an influence more related to Africans (both San and Yoruba) than to the macro-Asians (French, Han, Papuan), so it is something like "Basal Eurasian". San and Yoruba represent imperfectly but well enough the pre-OoA Humankind, crucially without Neanderthal admixture yet. The African Humankind is (roughly) as diverse as all Humankind minus Neanderthal/Denisovan admixture and a few novel mutations maybe. You could of course use more African samples to better represent that primeval Humankind but they do the job well enough.

      Delete
    2. No, if it was from Basal Eurasian then it would be more closely related to Eurasians than to Yoruba and San. This shows a separate OoA event.

      "The ‘ancestral’ simulation produced estimates that were the most similar to the ones obtained in our data analysis, however, the differences between estimates obtained in the three data sets are too subtle to indicate a clear better fit for any of the models. We thus conclude that the source population diverged from present-day Africans at around the time of the San divergence, but this could be either before this divergence event or slightly after it."

      So not just Basal Eurasian, but Basal Human.

      Delete
    3. Hmmm...

      Not convinced: it looks like an illusion. An "L3" paleo-population, as the one that is at the origin of "non-Africans" would be almost similarly diverged from Yoruba and San. I don't think they can detect the subtle autosomal issues here that would allow them to be so precise.

      For a reconstruction of the early expansion of Homo sapiens in Africa see: http://forwhattheywereweare.blogspot.com/2013/04/synthesis-of-spanish-language-series-on.html

      And, complementarily, for the archaeo-cultural data (MSA and such) see: http://forwhattheywereweare.blogspot.com/2013/03/african-msa.html

      Yourbans would be primarily an "L2" population, which overlaps somewhat with the "L3" one but is still distinct and must have diverged prior to the OoA. The San are an "L0d" population primarily and diverged almost certainly earlier, but we do see also some "L0" and "L3" inteaction in Eastern Africa, which very probably pre-dates the "final" conformation of the Khoisan population.

      So IMO that datum can only say with certainty that it is a population whose genetics pre-date the macro-Eurasian split in various subcontinental or "racial" populations, what is coincident with an OoA or "Basal Eurasian" primitive population. It would have been interesting maybe if they would have compared with other populations like Han, Papuans, high-ASI Indians or even something like the Dinka (closer to the OoA population than Yoruba and San), but no luck.

      Delete
  4. The first admixture (100kya) is exactly what I expected to be found, based on old archaeology books of mine that talk about "transitional" Neanderthal remains from 120kya or so in the Fertile Crescent region. I believe the admixtures around 50kya were AMHS mixing with already hybridized populations. (hence the difference between the two was somewhat ameliorated) Also, as for the 100kya hybrids appearing in Altai, I suspect that the y-dna P's-to-Q's were following them northeastward as a strategy for finding resources, and even imitating some of their lithic traditions. (i.e. the "Mousteroid" characteristics of enigmatic artifacts in Siberia and America) Now as for the other admixture to the Han and Papuans, that one is weird, but perhaps it was a Levantine Neanderthal whose environmental preferences took them both westward to Spain and eastward through Iran, the Kachi plain of Baluchistan, and the Siwalik Hills of northern South Asia, to an admixture rendezvous near the Ganges?

    ReplyDelete
    Replies
    1. I don't have a clear opinion but 50 Ka ago for the divergence of macro-Eurasians seems a bit too recent. A possible explanation might be that 100 Ka is slightly under-estimated and therefore also the second figure. If the older date would be corrected to 125 Ka (within archaeological reason) instead, then the later one would be approx. 63 Ka, what is more reasonable but still a bit short for me to consider the date estimates fully credible. Notice that all these estimates have huge error margins that are almost never reported, in what I consider a bad practice. If we were to use the 5-sigma standards applied to quantum physics, we would just totally ignore age estimates, because they'd be just a massive blurr from the extremely deep past to yesterday at midnight. So tricky...

      Delete
    2. This comment has been removed by the author.

      Delete
    3. The corrected date of 63 Ka does fit the dawnings of population recovery from a deep bottleneck triggered by Toba. I imagine the cold shock of the post-Toba glacial era pushing Eurasian humans southward in Egypt, Arabia, southern Iran, and western South Asia, and the dessication concentrating them near large rivers . As the populations recover, the founder effects are perhaps: ydna E and mtdna L3 dominant on Nile, ydna F and mtdna N in lower Mesopotamia, ydna C and mtdna M at mouth of Indus. Maybe DE (ancestral to D) is in Yemen and gets a head start plying the coast, and other populations catch on to the trend. The F's and N's moving eastward from Persian Gulf lead to K's and R's. Most of the C's and M's go eastward, which is the dominant flow of people along the coast, but one group goes westward.

      Delete
    4. Notice please that what I have in mind when discussing the guesstimated "dates of admixture" you put forward is also this other entry, where I decided not to give credence nor comment much about their modeled chronologies. In that case the authors "found" (fig.2) a common history for all humankind until c. 200 Ka BP and then a common history for macro-Asians until c. 50 Ka BP. In this case I could not even figure a good palliative correction, unless we are ready to push Humankind's common roots much farther than 200 Ka BP. I think it's much safer to take such "conclusions" with a good dose of skepticism: add a big, non-reported, C.I. around the inferred dates until they become a big blur, and move on.

      More specifically: why the 63 Ka is probably not valid either, even if "better", because I infer from mtDNA that there was already a dearth of diversity in Asia, with various already distinct groups formed here and there. They were no doubt interacting more intensely than they could do later but still the people in, say, Sindh and the people in, say, Tonkin would not have been in contact, excepted migration events that we can track in the haploid DNA.

      So we cannot think of the pre-Toba Asian branch of Humankind as a small homogeneous population in some reduced area but as many such groups interacting with their immediate neighbors and growingly distinct in genetic pool (founder effects + endogamy-driven drift) but also culture. They were diverse "nations" even if they shared one common root some 25,000 years before. I guess that the best comparable example would be Native Americans some five centuries ago, as they had probably been diversifying in America, from a common root, for at least 17,000 years.

      Then Toba happened: some populations were wiped out or pushed to migrate, but some eventually even manage to exploit the empty niches and become a paleo-ethnic "superpower", as seems to be the case with the (mtDNA) N/R and (Y-DNA) K2 population with likely roots precisely in SE Asia.

      That's my reading, and is not quite like the one you make: for me mtDNA pre-N ("L3n" if you wish) was just a minor lineage at first, as was Y-DNA D, initially overwhelmed by mtDNA M and Y-DNA CF, both surely rooted in South Asia, but also very diverse in East Asia (SE Asia only at first). Y-DNA D (essentially "coupled" with M matrilineages) found niches here and there in East Asia but did not expand much, however mtDNA N and especially "her daughter" R did not just found some niche in SE Asia but starred a successful major expansion, along with essentially Y-DNA K2, after the Toba catastrophe. For full details, see the dedicated page.

      Delete
  5. My next comment is in response to this earlier paragraph in comments:

    "That Vilabruna is the first KNOWN European R1b1 only underlines this, because it implies that its precursors MUST have been in Europe since at least Gravettian times (no West Asian immigration to Europe between Gravettian and Neolithic). R1a must have expanded more recently but some branches of R1b have been lurking in Europe for millennia necessarily, just as happened with mtDNA H, so often its "couple". Another issue is if it re-expanded in the Neolithic but, even if it was restricted to Italy or wherever, R1b must have been in Europe for at least 32.000 years."

    (...And my comment also relates to my own earlier comment about P/Q's following Neanderthals northeastward as a strategy.)... Obviously seems P must have fanned out from Ganges to Indus to Central Asia to Siberia, and they must have had a combination of "Mousteroid" and some blades qualifying as early Upper Paleolithic, perhaps with some Bohunicioid finishing...all the way to Yana. Next come the prismatic blades I've read have been found in Howieson's Poort and in the Levant by 45 Ka, and when they spread to Europe by 30 Ka it is called Gravettian; they also reach NW China around this time, probably spreading within P, R, and Q populations, reaching Mal'ta-Buret' and eventually Meadowcroft. Now what I'm really leading up to saying is, I believe at some point microliths spread from India to Central Asia ("Broad Spectrum Horizon"?) and the Epigravettian? My hunch associates this with R1. (R1a east of Caspian, R1b around Black Sea, with some members traveling within greater Epigravettian all the way to Vilabruna. My question: Anyone have solid input on approx. DATES for microlithic in Central Asia? (Seems like archaeological record in the region is a bit spotty.) But 32 Ka seems maybe a little early to me for R1b? I was thinking R1b and R1a split apart in population rebound just after LGM?

    ReplyDelete
    Replies
    1. No "Musteroid" stuff. Mousterian AFAIK is a strictly Neanderthal techno-complex that is never found East of Iran/Central Asia (with possible presence in Mongolia). The techno-complexes of early Asian H. sapiens are not too well studied but in East Asia seem to dominate a simple flake industry, with almost nothing that makes it distinctive, while in South Asia certainly retained African-derived toolkits of the MSA traditions (and in some instances Aterian even). At the current level of knowledge of paleoindustries in Middle Paleolithic Asia, we cannot discern a clear trail beyond the Africa→Arabia→India first flow, we just know too little (or at least I do).

      Also if the Narmada hominin was a Neanderthal, this population did not use Mousterian techniques at all. I can't recall the details right now, but there was some discussion in this blog years ago suggesting that maybe proto-Neanderthals were a very early and very small OoA-like migration, that first went to India and only later to Europe, that might explain why European and West/Central Asian Neanderthals used Mousterian but Indian ones did not.

      Also the origin of the MSA is not related to that of Mousterian nor the techniques share much at all (other than knowledge of the Levallois technique, used irregularly anyhow, and in general Middle Paleolithic basics). These are independent developments. MSA is anyhow quite a diverse category for what I understand, but it probably share a common origin and that origin may well be the same as that of H. sapiens.

      "My question: Anyone have solid input on approx. DATES for microlithic in Central Asia?"

      Not me. Probably not before Neolithic, where microlithism became "the state of the art" of stone techs all around. However these microlithic techs are typically derived from macrolithic ones in the same areas, for example in Europe, which is best researched, so we should not think of microlithism as a "culture" at all, but just a variant, a refinement, of other previous Upper Paleolithic techs, being the result a matter of size mostly, and without any clear cut "yes/no" divide (there were also microliths before the generalization of microlithism, just that at lower frequencies).

      ...

      Delete
    2. ...

      "But 32 Ka seems maybe a little early to me for R1b?"

      Not for me. I just don't believe in academic molecular-clock-o-logy but anyhow it is clear that R1b is much much older than "his brother" R1a, for example, if you scroll down in this entry to the update, you will see a discussion on plausible chronologies based on Underhill's data and R1b-M343 yields 34 Ka BP (being therefore labeled as "Gravettian"). The expansion of modern European subclades would be much more recent however: R1b-S141 would be c. 19 Ka ("Solutrean") while R1b-M412 would be c. 15 Ka ("Magdalenian"). This is based on a recalibration in which age(R1)=48 Ka BP, what IMO is reasonable, as R1 must have been part of the early UP migration to West Asia, just like IJ and (pre-)Q -- but not R2 or P(xR,Q). All this was estimated by me long before the Vilabruna datapoint was known but this new knowledge confirms my theorization, so all is well for me, as is for Einstein's predictions about gravitational lensing and what-not. Science is that way: truth, reality it's not always as scientists claim to be, but as it is, and scientists must adapt to facts and discard bad theories and models without regrets. I do and I expect others to do the same, more so when they come with academic "certification", sometimes quite laughable considering their disregard for the scientific method and abuse of scholastics and complacency on weak speculations that have just become the fad.

      "I was thinking R1b and R1a split apart in population rebound just after LGM?"

      Pre-R1a lurked in West Asia for many millennia as a "private" (too small to matter) lineage, until it expanded already in the Holocene. R1b on the other hand must have experienced a much older first expansion, although it seems clear that it did not became dominant in Europe until much later times, remaining probably restricted to certain populations or whatever.

      Delete
    3. With a father and brother both physicists, I appreciate your respect for scientific process. For me, I only have time to make this subject an art (i.e. a hobby) rather than a science as you seem to. But I do derive great enjoyment out of guessing and then seeing if new discoveries seem to support my guesses. So anyway, whether R1b existed in Gravettian or Solutrean, as you assert (and which I have no reason to oppose), its presence at Villabruna indicates it was at least in the Epigravettian. Which brings me to ask, will you elucidate exactly what is the Epigravettian? Compared with the Aurignacian, Gravettian, and Solutrean, it has always seemed underrepresented in popular archaeology, or downright elusive. What is the relationship between the Gravettian and Epigravettian? As I seem to have read, the Gravettian is characterized by long blades pressure-flaked from a prismatic core. (This also characterizes Mal'ta-Buret' and Meadowcroft Rockshelter, I seem to understand.) Does this also characterize the Epigravettian? I had thought Epigravettian was microlithic, but you seem to say that arrives later with the Neolithic. Please shed some light on this. Thanks.

      Delete
    4. "But I do derive great enjoyment out of guessing and then seeing if new discoveries seem to support my guesses."

      That's absolutely also part of the scientific method, and probably the best part of all, not just the one where you most directly put imagination or creativity to work but also what allows it to reach further, to explore new horizons, the questions that maybe science will some day answer and that will never be able to answer without the question or the hypothesis itself, that "what if...?", that "why not...?", that "I wonder if..."

      It seems that there is one only big difference between great apes and us: they do not make questions. When able to communicate, they express emotions, give indications or answer our questions but they do not ask them as far as we know. That's maybe why we have come so far, for good or bad: what if?, why?, how?, who?, when?, what?

      I remember that when I was a kid there was a saying or quartet that our elders taught to us, it was in Spanish though and no idea if it exists in English:

      "Seis servidores honestos
      me enseñaron cuanto sé:
      qué, dónde, cómo,
      quién, cuándo y por qué".

      In English it'd be: "six honest servants taught me what I know: what, where, how, who, when and why".

      So yeah, asking questions is absolutely a key part of the scientific process and of being human (more or less the same thing). It's not yet the answers but you won't probably get any answer if you don't ask first.

      Delete
    5. As for Epigravettian, it is not one thing, but a bunch of cultural groups, sharing (unless prehistorians got it all wrong) a common phylogenetic ancestor in Gravettian proper. They include but are not necessarily microlithic, in fact I understand that Epigravettian is every cultural group (mostly in Ukraine-Russia and Italy, sometimes in the Balcans maybe) which, after the LGM, retains those Gravettian roots (instead of becoming something else: Solutrean and/or Magdalenian, as happened in West and Central Europe). Not sure but I understand that, as with other cultural groups, microlithism only becomes dominant in the Epipaleolithic.

      Notice that "epi-" means peripheral or derived, nothing else. Wikitionary describes it as "above, over, on, in addition to". Just being epi-something does not make it necessarily Epipaleolithic (sometimes called "Mesolithic", although many prehistorians reserve the term "Mesolithic" strictly for proto-Neolithic cultures where proto-domestication of crops or similar is happening, in West Eurasia that'd be the Fertile Crescent and Nubia only).

      We can also talk of Epimagdalenian and I guess of Episolutrean, we can even talk of Magdalenian proper as being an "Epi-Aurignacian", just that these terms are less commonly used or not at all. I personally only use "Epimagdalenian" in order to group all Epipaleolithic cultures that derive from Magdalenian, mostly Azilian and Tardenoisian, but also other less known groups. But in the case of Epigravettian they mostly lack even conventional names, with the literature assuming that the reader understands that it's not the same some Italian Epigravettian and some Ukranian one, even if they seem to have the same ultimate roots in pre-LGM Gravettian proper.

      "This also characterizes Mal'ta-Buret' and Meadowcroft Rockshelter, I seem to understand."

      Mal'ta is definitely classified as Gravettian, not just because of the techniques but also because of its "Venus" idols, which seem to derive from European ones (not strictly Gravettian always but very common in this culture anyhow).

      As for Meadowcroft, I just heard of it for the first time (unless I did before and forgot, it happens too). It seems an interesting Paleo-American site, more so if the dates of 16-19 Ka BP can be confirmed. Personally I would not be surprised if some sort of "Gravettoid" techs made its way beyond Altai, but this clearly needs more research (Siberian, Mongolian, North Chinese and Alaskan research).

      There are many open questions on Gravettian and also on Solutrean, even more on this culture really. Gravettian is clearly intrusive but Solutrean is a total mystery in many ways. However reading on Basque caves you see how archaeologists seem to find transitional levels, where Solutrean seems to show up as "reaction" to Gravettian in a somewhat gradual manner. I don't understand well how this can be, because it seems to have no obvious relation either with Aurignacian nor Gravettian techs, but surely, if you know the fine detail, if you are an engaged professional archaeologist, you grasp these things better.

      So I have to redirect you ultimately to the library and/or the scholarly archaeologists. I can just draft the rough picture as they tell it, I'm not so much into stones, flakes and soft hammer percussion as they are. I just try to understand the general picture they draft with their ant-like work.

      Delete
  6. This comment has been removed by the author.

    ReplyDelete

Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).