November 8, 2010

A note on West Eurasian haplogroups

Some aspects of the recent discussion in the previous post, got me thinking about West Eurasian founder effects, or major haplogroups and the possibility that they may reflect a number of coalescence areas or populations. 

So I got to consider the founder lineages of West Eurasians, let's see:

mtDNA:
  • Major: R0 (incl. HV), U (incl. K), JT
  • Minor: N1 (incl. I), X, W, M1
  • Uralic-related: CZ, D (probably later arrivals via Siberia)
  • African flows: L(xM,N) (not going to detail them here)
Y-DNA:
  • Major: R1b, IJ, G and maybe R1a (or a subclade of it)
  • Minor: T, L, Q
  • Uralic-related: N
  • African flows: E1b1b1
Something that is quite apparent is that the number of lineages in each category (specially if we ignore the Uralic and African groups, which clearly belong to different processes) are not too different between Y-DNA and mtDNA, what suggests that each pair represents a founder population, specially among the major haplogroups. Another thing is how to pair them and even if doing that makes any sense at all but it's quite possible that at the beginnings of the colonization of West Eurasia they were coupled to a great extent because of the well known phenomenon of fixation by drift in small populations. 

So I am guessing that there were like up to six or seven founder populations (maybe less as fixation does not need to be absolute, specially not in expansive contexts) some 50 millennia ago in West Asia. If we only consider the major clades, maybe these were just three.

Where did they exist? That I'll leave up to you. The details of the MP-UP transition in West Eurasia are not fully well understood and there seems to be blanks in the research. Some specific regions can be Palestine (clearly so), Central Asia, the Zagros mountains... an oasis area in the Persian Gulf (now submerged) has been proposed recently too... parts of Anatolia surely played a role, though here is where one of the most critical blanks exist...

Anyhow, most if not all these lineages are wildly scattered, so, while their West Asian origin is certain, the exact locations in this area are not at all clear. Maybe I am wrong and they were all part of a single founder population with only some, now blurred, structure. 

As I say, just a working note. Thinking loud so you feel free to provide feedback and opinions. Cheers.

28 comments:

  1. I think you are on track.

    I also think that ancient DNA (particularly North African, Upper Paleolithic European, and early Neolithic European), modern population coincidences, phylogeny, and gene flows on each side that we think we understand, can explain a lot.

    Since we have more ancient DNA on the mtDNA side, that part of population genetic pre-history is better resolved.

    I also think that it can be fruitful to determine what couldn't have happened, even if that means that founder populations are left somewhat vague. For example, ancient DNA can tell us pretty definitively that certain founder lineages were present in LBK farmers, but not in their Upper Paleolthic predcecessors, even if there are some lineages for which that question is unresolved. For example, we can say with some certainty that mtDNA RO was probably not an important component of Upper Paleolithic Central Europe, and that mtDNA N1a and HV were part of early LBK populations.

    North African ancient DNA puts some lineages there much earlier than a lot of otherwise reasonable theories would suggest.

    When you have two different haplogroups that we know were both present at a certain time and place, and they come from very different layers of the same phylogeny, while other lineages at the same layers of the phylogeny are absent, we can infer that the time and place where we find them together was not the founder population for one or both of them.

    Mutation ages, even if not accurate in absolute terms, can provide Baysean input about which mtDNA founder groups to associate with which Y-DNA founder groups.

    The possibility that the minor lineages don't match one to one because some small migration had overwhelmingly men (perhaps invaders) or overwhelmingly women (perhaps with returning warriors as tribute or through long distance marriage arrangements tried to trade routes, from somewhere else) is possible. But, it is hard to see how the major lineages could not have involved members of both sexes in founding populations.

    It may also not make sense to compare lineages at the same levels of notational generality. A full lettered mtDNA lineage may correspond to a mere sublineage on the Y-DNA side.

    ReplyDelete
  2. A problem is that we have now lots of info on the aDNA of North-Central Europe but rather little for South-Central Europe, where LBK originated, specially in regard for the (Epi-)Paleolithic period. Actually we have only some evidence from this (Epi-)Paleolithic mtDNA south of Dresden, which in Paleolithic, and even Neolithic, terms is quite far to the North.

    We know (with whatever certainty aDNA studies have) that mtDNA H has been found in (Epi-)Paleolithic North Italy for instance, as well as in Portugal and Morocco. CRS HVS-1 sequences, which typically are within R0 (incl. HV, H and, most commonly now, H1), have also been found in Gravettian South Italy and Central Russia, as well as Morocco.

    LBK people seem to have spread North (always with due caution) N1, H, V, K, T and J. But the data from Italy, Portugal and Morocco make the presence of H (and surely V too) in Paleolithic
    Europe almost certain.

    In fact, I am quite persuaded that the huge star-like structure of H (only comparable to that of M) can only be explained by a massive sudden scatter as the one that probably happened in the colonization of Europe. Diversity and scatter patterns suggest an origin in Central Europe, but this may be an artifact of later recolonization. Other suggested origin for H is SW Europe (but this would hardly explain its expansion to East Europe and Central Asia).

    A possibility is that parts of Central Europe (Panonnian plain?) would have been a reservoir of mtDNA H, even if further North U was (maybe) the dominant clade.

    "North African ancient DNA puts some lineages there much earlier than a lot of otherwise reasonable theories would suggest".

    Molecular clock estimates are NOT reasonable in general. In fact I am proposing that a large part of the error comes from measuring mutations to present, instead of using mutations to ancestral nodes (because larger haplogroups would tend to drift out novel mutated lineages, impeding a regular "molecular clock" in fact).

    "Mutation ages, even if not accurate in absolute terms, can provide Baysean input about which mtDNA founder groups to associate with which Y-DNA founder groups".

    I think that the molecular clock conjectures should be kept out of this because they are totally unreliable. It must be considered on the merits of distribution patterns alone or almost alone.

    "It may also not make sense to compare lineages at the same levels of notational generality. A full lettered mtDNA lineage may correspond to a mere sublineage on the Y-DNA side".

    Not sure what you mean but I am trying to consider the West Asian lineages on their own merits, under the supposition that they all arrived more or less at the same time, i.e. c. 50-40 Ka ago, from a South Asian source (but in some cases with an ultimate SE Asian source in fact, as seems to be the case of mtDNA N and R and Y-DNA MNOPS but not the rest of F, IJK and K). I think they all arrived roughly at the same time (same period) within the same kind of general drive. Then the gene flow largely stopped and in some cases even reversed (U2, etc.)

    ReplyDelete
  3. I think what Andrew was trying to say is that at any particular slice in time, one has to be careful about what levels in haplogroups should be compared - IMO not only between y-DNA and mt-DNA, but also between the same kind of DNA. Firstly, the lettering/level system is arbitrary and organic (growing/changing with better knowledge) from the get-go, and secondly, between two different lineages, there could have been huge differences in past population size, meaning huge differences in the number of accumulated mutations (even if a comparable number of final branches survived). For example, the IJ level may be comparable to some R1b sub-level 50,000 years ago.

    As to origins, I would also look farther north: both in East Asia and in Eastern Europe, some of the earliest sites after expansion ~45K years ago are in the far north. So, one group likely migrated via the steppes (following grazing animal herds) north of the Caspian and Black Seas.

    Another point to remember is the effect of the Younger Dryas. Before that, you have the possibility of huge founder effects but also mixing in the re-population between northern France and ~ the Ukraine. Then the population had to again shrink into much smaller areas (albeit larger than during the LGM).

    From that viewpoint, it seems interesting that R1b and R1a couldn't make it farther east and west, respectively, but I evidently was able to diversify and spread. Perhaps we can try to identify a predominantly I-carrying population that largely prevented the R's to penetrate the central regions. For example, it seems plausible that the Danubian plains east of the Black forest and the southern Czech lands remained hospitable during the Younger Dryas - but I would have to look up direct evidence to back this up. Areas that supported significant populations during the Younger Dryas would then have had a huge advantage after ~11.5K years ago, and would have prevented further expansion from the south (Iberia, Italy, Balkans, Anatolia, Caucasus).

    ReplyDelete
  4. "Maybe I am wrong and they were all part of a single founder population with only some, now blurred, structure".

    I realise you'll be very surprised when I say that I'm fairly sure you are correct.

    ReplyDelete
  5. "I think what Andrew was trying to say is that at any particular slice in time, one has to be careful about what levels in haplogroups should be compared - IMO not only between y-DNA and mt-DNA, but also between the same kind of DNA. Firstly, the lettering/level system is arbitrary and organic (growing/changing with better knowledge) from the get-go, and secondly, between two different lineages, there could have been huge differences in past population size, meaning huge differences in the number of accumulated mutations (even if a comparable number of final branches survived). For example, the IJ level may be comparable to some R1b sub-level 50,000 years ago".

    If that's what he means then I agree.

    It's very difficult for me (and I believe anyone) to measure the temporal structure of Y-DNA but mtDNA, with a much smaller DNA chain, is a lot better known, almost perfectly known by comparison, and hence can be analyzed better that way, even if there are always some limitations. That's why I try to understand Y-DNA from its interaction, "coupling" or "pairing", with mtDNA.

    For example this allows me to understand that the claims about the brutal differences in age between mtDNA Eve and Y-DNA Adam are surely wrong, as Y-DNA A couples very well with mtDNA L0, etc. It's not always that easy but in some cases it's quite straightforward.

    "As to origins, I would also look farther north: both in East Asia and in Eastern Europe, some of the earliest sites after expansion ~45K years ago are in the far north. So, one group likely migrated via the steppes (following grazing animal herds) north of the Caspian and Black Seas".

    The Kostenki group, you mean? I am not really sure it comes from Altai, though it's possible. The "Aurignacoid" continuum anyhow is pretty much universal in West Eurasia in the colonization period and people could also have reached Eastern Europe via the Caucasus (Mezmayskaya shows the same kind of signs of population replacement). Of course the ancient U2 of Kostenki is suggestive of a Central Asian connection but notice that U4 is a close relative of U2 (U2'3'4'7'8'9). Also I tend to associate Altai specially with the Sibero-American clades like Y-DNA Q and mtDNA X2 (secondary home after West Asia for both clades probably).

    ...

    ReplyDelete
  6. "Another point to remember is the effect of the Younger Dryas".

    Sure, though uncertain about which effects it could have right now.

    "Before that, you have the possibility of huge founder effects but also mixing in the re-population between northern France and ~ the Ukraine".

    I'm highly skeptic about West-East flows after Gravettian expansion in fact: Magdalenian never reached to Eastern Europe, nor did Solutrean. The only such flow is the Indoeuropean (Kurgan) expansion, which is just too recent and too thinly apparent.

    So, in general, if I see a haplogroup shared between Eastern and Western Europe, specially at the extremes and specially in mtDNA (less likely to have been carried around by the warrior IEs), I think it's something pre-LGM. A different case might be if the Balkans or West Asia are involved, as these two regions were important in the Neolithic (another good occasion for founder effects of some size at least). This may be the case with Y-DNA I2a, for instance, which may have migrated westwards from a putative origin around Ukraine, via the West Balcans, where Mediterranean Neolithic originated and where this haplogroup is very common.

    "Perhaps we can try to identify a predominantly I-carrying population that largely prevented the R's to penetrate the central regions. For example, it seems plausible that the Danubian plains east of the Black forest and the southern Czech lands remained hospitable during the Younger Dryas - but I would have to look up direct evidence to back this up".

    I suspect I being eastern by origin and I'm still very much unsure about how R1a spread (maybe it just arrived to East Europe in the Neolithic/Epipaleolithic and then actually did spread with Kurgan IEs). A problem here is that we do not know where did the PIE core population of Samara valley had its roots, another is the apparent greater basal diversity of the haplogroup in India.

    But, if I is Central-European by origin, then an explanation may be conjectured re. the differences between Bohunician-Bakhokirian and proto-Aurignacian. Maybe these two West-Asian originated cultural groups somehow reflect the arrival of two different, albeit related, lineages/populations, which then had distinct success in further spread, either within Aurignacian or within Gravettian.

    "Areas that supported significant populations during the Younger Dryas would then have had a huge advantage after ~11.5K years ago, and would have prevented further expansion from the south (Iberia, Italy, Balkans, Anatolia, Caucasus)".

    I don't know the Younger Dryas particulars but overall in the late UP, it was West Europe, specially SW France, where almost all Europeans lived, it seems. Other important regions were the Low Countries/North France and Moravia. Eastern Europe and the Mediterranean (excepting SE France) were thinly populated in that time.

    However if some of these "thin" populations then expanded dynamically in the Epipaleolithic or Neolithic, they surely could make a large impact. Being many in the past, doesn't mean being many in the present, unless that population effectively expanded further.

    ReplyDelete
  7. "The Kostenki group, you mean? I am not really sure it comes from Altai, though it's possible."

    I just mentioned it as a parallel development (of foraging in the extreme north). Some of the mtDNA connections are still intriguing, to me - but here I was just thinking about the northern steppe exploitation that evidently was one route to Europe.

    "I'm highly skeptic about West-East flows after Gravettian expansion in fact: Magdalenian never reached to Eastern Europe"

    Apparently, it reached Poland, and the Hamburgian and Ahrensburg complexes (reindeer exploiters at the edge of the ice sheets all the way to the Baltic from just before and through the Younger Dryas) appear to be derivatives - but with some important differences that may indicate mixture with different peoples, perhaps from the east.

    Relatively early Magdalanian is well established at the northern edge of the German "Mittelgebirge" - the lowish mountains just south of the northern plains. This recent review provides great background information about the era in the North.

    Given the Magdalanian in much of northern Germany, I would associate that with either R1b or perhaps R1b with some I admixture. I may be more associated with the northern Balkans and Bohemia/Moravia before expansion - but I need to read up on Younger Dryas finds in that region and in Southeast Germany.

    I agree that R1a seems to enter later - but not necessarily much later; it also has had much diversity and large geographical extent for millennia. I seems to be found all the way around the edge of R1a - perhaps indicating that it once was wider spread and since then replaced by R1a.

    ReplyDelete
  8. When I said Eastern Europe I meant East of Poland and Hungary, specifically the Dniepr-Don area, which is of Gravettian tradition and seemingly has a Neolithic based on those local traditions. These must be one of the founder populations of Europe and a quite clearly distinct one.

    These Eastern Europeans spread in the late moments of the UP into Romania and then, in the Epipaleolithic to the Caucasus and, for what I have read, beyond: the Zagros area (Zarzian culture), which was apparently deserted since the LGM. This culture is directly behind Zagros Neolithic. So they may well also be one of the founder populations of West Asian Neolithic.

    In addition, they influenced Northern Europe in the Chalcolithic (or Neolithic, as there's no meaningful copper in that area), a few centuries before Indoeuropean migrations. While Denmark was more densely populated and more solidly agricultural, Sweden was not and was almost without doubt highly impacted by this flow from the Baltics and ultimately Ukraine-Don. These are the "famous" Pitted Ware peoples and the related peoples (also Pitted Ware in the books I read a decade ago but termed often Eastern Funnel Beaker in the online materials I read now), which sometimes are used as examples of European hunter-gatherers. But they were regressive hunter-gatherers with pottery and pigs and are better termed as peripheral Neolithic. It is important to understand that they do have a clear connection with East European Neolithic (pottery, burial customs) and you can more or less track them as a bunch of "poachers" moving NW from Neolithic Ukraine, even if they may have also incorporated Baltic forager remnants of Magdalenian tradition as they reached the Baltic, where they probably participated in the genesis of Funnel Beaker as a whole (along with Megalithic farmers, centered in Denmark and Southernmost Sweden).

    I think this is important to understand in the context of how Northern Europe was populated (then add an Indoeuropean layer and the Finno-Ugric layer even further north and that's it, I understand). Eastern flows seem to have an specially strong impact in Sweden and Poland but the Polish distinctiveness may be more of Kurgan origin, while the Swedish one seems to retain a larger pre-IE Eastern European layer, which I identify typically with Y-DNA I.

    ReplyDelete
  9. I will read now your linked paper, which looks very interesting. But I am aware of epi-Magdalenian presence in all that area and even further north and NE (Denmark-Scania, Poland, the Baltic countries) but the population densities achieved by hunter-gatherers were no doubt limited and there were further layers thrown upon them by prehistory, mostly from Eastern Europe (but also add Megalithic Neolithic, which, whatever the particulars, is no doubt a Western influence). On the other side, the LBK (Danubian) influence shows up nowhere as far as I can tell, though it may have co-influenced Danish Neolithic (but to a limited extent in any case).

    "I agree that R1a seems to enter later - but not necessarily much later"...

    Honestly I am puzzled by R1a, which I believe deserves more in-depth research in order to clarify its structure. With the current data there are two possibilities, IMO:

    1. That it arrived somehow (Central Asian Neolithic?) to Easternmost Europe (East of the Volga) and spread only with IE (Kurgan) migrations (in Europe). R1a1a7 seems to fit well a particular sub-group of this expansion: the Polish Kurgan sub-center of Eastern Baalberge-Luboń-Globular Amphorae-Corded Ware but I am unable to say much with certainty for the rest of the haplogroup.

    2. That it was part of the Epigravettian Eastern European ("Ukrainian") populations (along with I?) and spread with this haplogroup.

    As I and R1a scatters only partly overlap (specially anomalous is I2b), I tend to think that option 1 is the most likely but I may be wrong.

    At the current understanding of R1 phylogeny and diversity patterns, I am rather for discarding that R1 as a whole spread from a European or West Asian origin, as the R1b/R1a divide might suggest without further research. Rather R1 and R1a seem of South Asian origin, while R1b looks of West Eurasian origin instead (it may even be from SE Europe but West Asia still stands as most likely candidate).

    This makes me think that the R1->R1b transition is very old, as we know of no flows from South Asia after the colonization process some 50-40 Ka ago.

    R1a would be, if so, the main (or most successful) South Asian remnant of R1, which then spread somehow into Central Asia and East Europe, then again spreading as part of the IE (Kurgan) expansions. A big problem is to discern what is Kurgan and what is pre-Kurgan in R1a and what role played Central Asia in its pre-IE spread.

    ReplyDelete
  10. After reading the paper, an issue that seems to be raised and a key one is the understanding of Hamburgian (and derived cultural groups of Low Germany and Denmark) as distinct from Magdalenian (albeit contemporary in the Central European region). However no alternative source is suggested.

    In any case, this Hamburgian cultural area might be behind the distinct and Nordic-specific R1b1b2a1a1 (U106) sub-haplogroup, which does not look Magdalenian-related. The expansion of this haplogroup southwards may be directly-related with Germanic expansion. However the highest frequencies (>30%) are reached in the Netherlands rather than North Germany or Scandinavia. Sweden, as expected, is low for R1b in general and also for this haplogroup.

    Another haplogroups that might have been related to Hamburgian would be I (I1 and I2b specifically) and mtDNA U5 and maybe U4 as well. But earliest researched Danish/South Swedish sites (all Neolithic or later) show already other lineages as dominant in the mtDNA aspect.

    ReplyDelete
  11. To me it looks like both R1a1a and R1b are of South Asian origin. The main difference is that apparently R1a1a took a northern route north of the two big lakes - although this could have been anytime before LGM, not necessarily 45,000 years ago.

    There seem to be old pockets of different strains or R1a1a in the Balkans and in the non-Slavic speaking countries (Romania and Hungary), and it is also attested in Germany ~4,600 years ago. That to me means that it likely was the main haplogroup of the Ukraine/Crimean LGM refuge, and may also have been partially present in the northern Balkans, and populated eastern Europe long before the advent of agriculture or IE spread.

    I think the Caucasus and Iran have some R1*, which could indicate a number of migrations to the west, with R1a1a being perhaps the latest.

    R1b seems to be South Asian, again.
    R1b1 may have derived from that in Turkey and entered Europe from there (there still seems to be some R1b* left in the Middle East). Again, it is not clear if the first AMHs had yet other haplogroups, but the simplest explanation to me is R1b1 was the main haplogroup entering Europe ~45,000 years ago. Perhaps the migrating groups picked up some IJ in Turkey along the way, but the fraction diminished on their way west. This would have left relatively little IJ/I west of Italy, and would have created some "hot spots" in the Balkans and surrounding areas - which seems to agree with its distribution.

    So, in my mind, R1a1a, R1b1, and IJ were the only significant haplogroups that made in into Europe before ~25,000 years ago. The recent ideas of making these neolithic completely disagree with all known migration events and paths, IMO, and also disagree with the western star-like cluster of R1b1.

    ReplyDelete
  12. "To me it looks like both R1a1a and R1b are of South Asian origin".

    R1b does not even exist in South Asia, except for a nomadic ethnicity which look totally immigrants. So R1b is simply not South Asian, even if R1, R and P are. R1b represents an R1 flow westwards, probably the first one of its kind and maybe related to that of Q, which also looks West Asian by origin (though more easterly).

    If R1a and R1b would be strictly parallel, then there shoul be no R1a in South Asia and certainly not the older branches (some may have poured back in with the Indo-Aryan and other steppary flows but should be quite derived by then.

    In SA there'd be "R1c" or whatever else but not the same clade as in East Europe. This I say because the downstream structure of R1a has been highly reluctant to show up. Even in the paper that discovered the South Asian origin of the clade, most R1a still was clumped into a single, derived, node and some lesser branches, strongly suggestive of a single rather fast expansion at that level.

    I suspect that R1a moved into Central Asia and East Europe at a later moment but not sure exactly when for lack of clear archaeological evidence of such migration across Central Asia.

    "There seem to be old pockets of different strains or R1a1a in the Balkans"...

    Can you point me to the relevant paper or related source?

    "... and it is also attested in Germany ~4,600 years ago".

    That's trivial because it is in an Indoeuropean context: it only confirms that R1a seems to have expanded in Europe with the Kurgan migrations, whose origin is at the Samara valley.

    But what before? We just do not know.

    "That to me means that it likely was the main haplogroup of the Ukraine/Crimean LGM refuge".

    Unsure but I have "always" tended to favor I in that role. I have no good references but someone told me that the highest diversity of I (or at least I2) seems to effectively be in Ukraine, what supports this hypothesis.

    Btw, are there any archaeological sites in Crimea? I am aware of the Dniepr-Don (Ukraine-South Russia) area with expansion to the Upper Volga in favorable climates but no idea about Crimea.

    ...

    ReplyDelete
  13. ...

    "I think the Caucasus and Iran have some R1*, which could indicate a number of migrations to the west".

    Well, R1* is found also in Europe. As such does not say much (unless haplotype structure clarifies what does it mean) because they may be remnants of the pre-R1b arrival.

    "R1b seems to be South Asian, again".

    No. Why do you say that? All R1b is non-SA and seems indeed to have a West Asian, Balcanic or even maybe Italian origin. There's essentially no R1b in South Asia and when you look at the whole picture the highest diversity is in that NE Mediterranean arch.

    "... the simplest explanation to me is R1b1 was the main haplogroup entering Europe ~45,000 years ago. Perhaps the migrating groups picked up some IJ in Turkey along the way, but the fraction diminished on their way west".

    I can agree with that. But that precisely sends I (and not R1a) to the eastward expansion.

    However the haplogroup which entered Europe, at least north/west of the Balkans and Italy, may not have been R1b1 but a much more derived clade: R1b1b2a1. This seems pretty straightforward from the haplogroup's structure. R1b1b2 may have coalesced in the Balkans rather than Anatolia, as it seems to be found mainly there in its root form. The Anatolian clade would hence be a monophyletic clade, derived from the same Balcanic root as West European R1b1b2a1.

    In parallel maybe some R1b1a also arrived to as far as Italy, while Central Asian R1b1b1 instead might have have never gone through Europe (though maybe it did), because there's no trace of that branch over here and, considering West Asian diversity, this is not necessary at all.

    "This would have left relatively little IJ/I west of Italy"...

    I suspect all Western I are recent arrivals (Neolithic and such). But not totally sure because there's a good deal of unresearched I*. The largest portion however looks derived from Western Balcans' I2a, which strongly suggests for a Neolithic flow within Cardium Pottery Neolithic.

    However I may be ignoring something because I still lack a mechanism for the flow from Ukraine/Romania westwards into former Yugoslavia.

    "So, in my mind, R1a1a, R1b1, and IJ were the only significant haplogroups that made in into Europe before ~25,000 years ago".

    Makes sense but I'd exclude J (half of IJ) and most likely R1a1a too. And I'd specify R1b1b2 and R1b1a (or rather two of the four known subclades of this haplogroup, which look Italian-specific). So I'd exclude R1b1b1 and the African-specific clades of R1b1a.

    ReplyDelete
  14. PS- Also central Ukraine is strangely low in R1a. What doesn't fit with your hypothesis for R1a because this area should be the most conservative district of the East European province.

    ReplyDelete
  15. Maju, according to the latest table published in Wikipedia, S Asia has ~5% R1b1 sub-clades - 3% (most)of that unique! Pakistan has 3% R1b1b2 - none of it in the European, derived U106. R1b* may have been found in the middle East.

    It is hard to make a case for R1b* because it is so elusive. However, S Asia has great R1 diversity and unique or un-derived R1b subclades - which makes it suggestive to seek the R1b origin at least very close to it.

    ReplyDelete
  16. Let's see: I am 99% sure that the R1b1b2 is all concentrated in a nomadic group original from West Asia, the Lambadi. Another thing may be R1b1b1 (the Central Asian clade), which I overlooked (my bad).

    In any case they do not have R1b1a nor R1b* (only documented in West Asia), what effectively excludes them from overall R1b origin.

    "S Asia has great R1 diversity and unique or un-derived R1b subclades - which makes it suggestive to seek the R1b origin at least very close to it".

    That is not what you can see in those tables at Wikipedia. You must have mistaken something:

    Southern Asians (n=288)- total R1b: 4.8% - R1b*: 0.0% - R1b1a: 0.0% - R1b1b2: 1.7% - R1b1b1: 3.1%

    All you see there is some derived R1b, specifically R1b1b and mostly R1b1b1.

    ReplyDelete
  17. @ Maju :

    "Other suggested origin for H is SW Europe (but this would hardly explain its expansion to East Europe and Central Asia)."

    I think you're right, but Some specific mtDNA H subclades could really have come from the western part of Europe as strange as it may sound.

    a/ I can't help noticing that The recently discovered megalithic site near Chelyabinsk * - east of the Urals correspond the highest frequency of R1b1b2 east of the urals, concentrated near Chelyabinsk (1st map : http://1.bp.blogspot.com/_x6Y4ZgFsZdY/THayWhv-w7I/AAAAAAAAAYw/ew7YwZrSlJI/s1600/Myres+maps+R1b1b2a1+frequencies.jpeg).

    * (http://www.russia-ic.com/news/show/10889/ - Chelyabinsk Archeologists Discover Ancient Ritual Complex of Menhirs)

    It also makes me think some of the mtDNA H among Mansi look kind of "western" IIRC for instance there were some H3 which is rather western IIRC.

    b/ The oldest Xinjiang mummies (all R1a1a) had some mtDNA hg K and H and this H had modern matches in modern European individuals. Most of these mmatches were actually more frequent in the west of Europe (4 english, 1 Icelandic (I assume it originally came from England or Ireland), 1 Portugal, 1 Italian, 1 german and 1 hungarian).

    ReplyDelete
  18. "Some specific mtDNA H subclades could really have come from the western part of Europe as strange as it may sound".

    I'm almost positive it is the case for at least H1, H3, H4 and H7. But H2 for instance is split and scattered and H8 or H11 are specific of Central Asia, so we must be careful with not taking the part for the whole.

    "I can't help noticing that The recently discovered megalithic site near Chelyabinsk * - east of the Urals correspond the highest frequency of R1b1b2 east of the urals, concentrated near Chelyabinsk"

    Curious indeed. Anyhow careful because the Bashkirs are known to be the R1b "island" in East Europe and they don't seem to have been included in those maps.

    "It also makes me think some of the mtDNA H among Mansi look kind of "western" IIRC for instance there were some H3 which is rather western IIRC".

    Totally.

    Is it possible that some western lineages spread eastwards with the Megaliths? It's a quite mysterious flow but it's also clearly there: they left a lot of monuments, so they are anything but invisible.

    ReplyDelete
  19. @ Maju : "H8 or H11 are specific of Central Asia"

    That doesn't seem to be the case for H11 (even though it is indeed found in Central Asia - up to the Udege of south-east Siberia actually).

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2660437/

    "while others are virtually absent in Iberia but are significantly more prevalent in Central Europe (e.g. H11)."

    Look at the frequency map here (bottom right) :

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2660437/figure/pone-0005112-g003/

    It's more true of H8 but it's not central Asia specific either.

    BTW, Any thought on the strange fact that the mtDNA H from from 2,000 BC Xiaohe R1a1a population of the Tarim Basin apparently matches better the wester European population rather than eastern ones?

    ReplyDelete
  20. God catch, Wagg. H7 is also not specific of the West even if it is found there and from there probably arrived to North Africa. I was saying from memory and got some wrong.

    ReplyDelete
  21. Maju : "God catch, Wagg. I was saying from memory and got some wrong."

    No problem, you're still an encyclopedia in the matter :)

    Yet... I think I spotted another little mistake :

    The Uralic-related mtDNA hg in west Eurasia is Z not CZ which is found mostly in eastern Siberia IIRC.

    http://en.wikipedia.org/wiki/Haplogroup_CZ_%28mtDNA%29

    Today CZ is found in eastern Eurasian populations.[2] It is most common in Siberian populations.


    http://en.wikipedia.org/wiki/Haplogroup_Z_%28mtDNA%29

    "Its greatest clade diversity is found in Korea, northern China, and Central Asia. However, its greatest frequency appears in Russia and among the Saami people of northern Scandinavia."

    ReplyDelete
  22. I meant C and Z (both 'sisters' in CZ, and 'nieces' of M8a in M8) and you seem to mean CZ*. AFAIK both C and Z, as well as D are found in that area.

    ReplyDelete
  23. Hey Maju, long time no see. Don't know if you noticed there are some changes in the new Y-DNA tree provided by FTDNA.

    They found that K1, K2, K3, K4 all share the M526 with MNOPS, so that makes it KMNOPS.

    And also, they found L and T share common Y-SNP, P326 and L298, so there's haplogroup LT under K now.

    http://ytree.ftdna.com/index.php?name=Draft&parent=root

    ReplyDelete
  24. The first part I could understand based on Karafet's paper on ISEA Y-DNA. I understand however that K1 has a South Asian distribution, what makes it a second MNOPS west of the Ganges Delta (along with P).

    I do not like the name KMNOPS because it seems to imply that all K is part of it and that is not correct: L and T are not and probably some K* is not either. I prefer to use MNOPS and, if need be, change K1-4 to MNOPS1-4. Makes sense, right?

    What I did not know (but is not totally unexpected) is that L and T are brother clades. Is there a paper on this matter?

    ReplyDelete
  25. "They found that K1, K2, K3, K4 all share the M526 with MNOPS, so that makes it KMNOPS".

    That's very much what I suspected. In fact the original paper showed a KMNOPS haplogroup. So the Ks too originated in SE Asia.

    "And also, they found L and T share common Y-SNP, P326 and L298, so there's haplogroup LT under K now".

    I didn't expect that, however it's not too surprising.

    ReplyDelete
  26. As far as I know, no paper, but here's the FTDNA draft tree:

    http://ytree.ftdna.com/index.php?name=Draft&parent=root

    (Also notice that R1b1b2 is becoming R1b1a2 and so on.)

    ReplyDelete
  27. M73 and M269 now become R1b1a1 and R1b1a2, and African V88 now becomes R1b1c. It makes M73 and M269 closer to each other.

    ReplyDelete
  28. "Also notice that R1b1b2 is becoming R1b1a2 and so on".

    I do not find that serious: R1b1a is another haplogroup. Someone is playing with the nomenclature and that is wrong unless there is consensus. That's why ISOGG exists, where scientists try to blend the research into a consistent phylogeny and nomenclature after due preliminary discussions. This seems amateurish and confusing to me.

    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).