March 15, 2011

Provenzal genetic data... and weird speculations

The following paper offers some information on the genetics of Provenzals and some specific populations of Turkey (Foça, Izmir) which is compared with older studies (on Turkey and Greece) to reach quite unfathomable conclusions:


I'm split on this paper: on one side it does provide some interesting data and makes some common sense claims (like Provence having been little affected by Neolithic expansion direct colonization) but then you stumble upon absurd ideas, such as Cardium Pottery stemming somehow from "Anatolia":
Using putative Neolithic Anatolian lineages: J2a-dys445=6, G2a-M406 and J2a1b1-M92 the data predict a 0% Neolithic contribution to Provence from Anatolia.
There is absolutely no reason to be looking at Anatolia: the Neolithic wave that arrived to Provence did not originate in Anatolia but in the Western Balcans. It is very possible that Anatolia was the ultimate origin of Greek Neolithic and this was in turn at the origin of Cardium Pottery Neolithic somehow, but the real origin of the Neolithic wave that arrived to Provence must not be looked for in any case in Western Anatolia - that is a total nonsense.

We know way too little as of yet to explain the exact process of cultural transference from West Asia (Anatolia specially) to the Balcans (Thessaly in particular) and from Thessaly to the Adriatic, where the cultural elements are so distinct anyhow. There is no particular reason to expect any arrival directly from Anatolia into Italy or SW Europe in the Neolithic. Any such migration would have been dampened in two filters: one in Greece and another one in the Adriatic Balcans.

From my ongoing (and slow) work of summarizing  European Neolithic in maps:


Here you can see in brown the first area of Cardium Pottery Neolithic: Dalmatia, Montenegro, Coastal Albania, most of Bosnia, Italy (in a second moment)... It has a precedent in Otzaki (Thessaly) and a derived influence in Biblos (Lebanon) but by no means can it be linked to "West Anatolia" of all possible places.

Universities and grants should require that any geneticist doing historical population genetics hire a prehistorian for assessment, sincerely.

Still there is a very interesting amount of data that is of interest, summarized (as I said before) in figure 2 specially. This is an extensive table that I cannot reproduce here with enough resolution without some previous work. So for reason of its relevance and novelty I'll focus on the Y-DNA data of Provence (n=51, only attested lineages shown):

  • E1b1b1b1a2 (V13): 4%
  • E1b1b1b1c (M123): 2%
  • G(xG2a3a) (M201): 8%
  • I1 (M253): 2%
  • I2(xI2a2,I2b) (M438): 4%
  • J1 (M497): 2%
  • J2a4h1a (DYS445=6): 8%
  • J2a4b(J2a4b1) (M67): 2%
  • R1a1a (M198): 10%
  • R1b1b2 (M269): 59%

Up 26 to 30% (depending on how you evaluate I2*) of the genetic pool is "Eastern Mediterranean" in Provence. E1b1b1b1a2 (V13) is probably from Albania or other Adriatic areas (see Battaglia 2009). That can also be argued to be the case for all the other "transmediterranean" lineages, which agrees well with a Neolithic origin of all them. However it is not impossible that these Neolithic arrived in batches and with intermediate stops in Italy for example or, why not, in Phocaea in some cases. 

But the research falls very short from demonstrating what they claim to demonstrate. If they have demonstrated something at all they have failed to explain it properly. So the only interest of this paper is the raw data, which adds to other such data to be integrated into a careful and comprehensive exploration of all (and not just some) data with proper prehistoric assessment. 

It is in any case important to understand that under the Neolithic colonization hypothesis, E1b1b1a2 should not be expected to originate neither in Anatolia nor in Greece but in Albania, Montenegro and Dalmatia. And, if anything, in Greece rather than Anatolia. Attributing European Neolithic directly to Anatolia or West Asia in general is not an acceptable assumption but a wacko fetish that should be discarded altogether.

62 comments:

  1. I don't understand what are the conclusions of this study. Provençals are 30% neolithic or greek? That's a lot, I wonder if this % is similar in most European countries. And R1b is more paleolithic?

    ReplyDelete
  2. The conclusions of the paper are a mess: they argue for Greek Phocaean origin (in Anatolia) but not for a Neolithic origin (in that same Anatolia). It's confusing and ultimately pointless.

    My first text was very harsh, cruelly sarcastic, but I finally cooled it down and removed that part, not sure if for good however.

    In any case what matters is the raw data. It shows that their small (n=51) Provenzal sample has Y-DNA that appears to be 25-30% from the Eastern half of the Mediterranean. But that's just Y-DNA not overall ancestry which may well be half that or even less.

    The authors argue for all that Y-DNA having arrived with Phocaean colonization but I find it unlikely and looks more like at least two waves: Neolithic and Phocaean (plus some random scrambling and sample uncertainty, I'd say).

    ReplyDelete
  3. "The conclusions of the paper are a mess: they argue for Greek Phocaean origin (in Anatolia) but not for a Neolithic origin (in that same Anatolia). It's confusing and ultimately pointless."

    But 30% seems too high to be that recent! I agree that a Neolithic origin for most of these haplogroups it's more reasonable, unless the Phocaean Greeks killed most Provençal men.

    "My first text was very harsh, cruelly sarcastic, but I finally cooled it down and removed that part, not sure if for good however. "

    No problem with that, only with the aims and conclusions of the study that aren't too clear.

    "In any case what matters is the raw data. It shows that their small (n=51) Provenzal sample has Y-DNA that appears to be 25-30% from the Eastern half of the Mediterranean. But that's just Y-DNA not overall ancestry which may well be half that or even less. "

    Y-DNA tend to vary in % a lot more than mtDNA. For example, the Sardinians have a very high % of the I haplogroup (>30%) , unlike other mediterranean populations where the % is much lower (5%), yet this doesn't mean they're descended from a different population to all their other neighbours.

    ReplyDelete
  4. Maju: here's that study of mediterranean populations the other day I was talking about, but you couldn't access to it because it's restricted:

    http://www.tesisenxarxa.net/TESIS_UB/AVAILABLE/TDX-0128110-091829//03.EGP_RESULTATS.pdf

    The first part is in Catalan, the second one in English. I think it has interesting information, such as an high haplotype diversity in Atlas Berbers, and ancient gene flow between mediterraneans.

    Unfortunately the study is inconclusive in many aspects, because it's very difficult and confusing, every haplotype tells us a different history.

    ReplyDelete
  5. I'd say it's a 25% because I2a is possibly of SW European origin. The apportions are anyhow within what we find in many Iberians (Valencians, Balears, Castilians, Andalusians, Portuguese) or South Italians (the paper mentions estimated 37% in Sicily, though is a questionable figure).

    But this is all Y-DNA, i.e. patrilineages and not overall ancestry. Valencians may be c. 30% Eastern Mediterranean by Y-DNA but they are not more than 20% by autosomal DNA (Bauchet 2007). If we apply this rule to Provence, then we are before not much more than 15% "Oriental" ancestry, which is not that much considering what is sometimes claimed (total population replacement and nonsense of that kind).

    "Y-DNA tend to vary in % a lot more than mtDNA".

    It is more easily swung to highs or lows of whatever kind. Drift is more intense in Y-DNA than in other markers, exaggerating the apparent results (so deal with care).

    The Sardinian case is a very clear case of brutal founder effect: two lineages I2a (found also around the Pyrenees and among some Albanians from Macedonia) and R1b1a (surely from Italy) are wildly overrepresented.

    Thanks a lot for the link to that paper, I'll read it with due attention. :)

    ReplyDelete
  6. I may make a brief comment on that paper, because, even if the markers are kind of odd, the array of populations studied has some interest. Notably I appreciate the inclusion of South French (which cluster systematically between Basques and other Iberians).

    However I'd like to have got those same graphs without the Ivory Coast outgroup. And even without the German and Siwa ones, though these do not probably introduce such a big distortion. I suspect that a W-E European Mediterranean dimension is not showing and that could be revealed by taking out one or more of the extremes (outgroups).

    Do you recall the original link?

    ReplyDelete
  7. My comment has disappeared, and I don't know why.

    The original link is this I think, but the study is not fully available here:

    http://www.ncbi.nlm.nih.gov/pubmed/19918993

    Fully available (pdf):

    http://www.tesisenxarxa.net/TESIS_UB/AVAILABLE/TDX-0128110-091829//05.EGP_BIBLIOGRAFIA.pdf

    ReplyDelete
  8. The stupid Blogger anti-spam filter again. Sorry about that but I cannot deactivate the spam filter. Two other posts were stored there - they are all restored now.

    As I read all comments almost only by email, and "spam-filtered" ones are notified just like the others, I seldom detect such issues unless you tell me.

    "The original link is this I think, but the study is not fully available here"....

    Sure but if I'm going to write an entry I should link to the official page, even if I also hang the paper at ZohoViewer for example and link to that as well.

    Thanks again.

    ReplyDelete
  9. "The stupid Blogger anti-spam filter again. Sorry about that but I cannot deactivate the spam filter. Two other posts were stored there - they are all restored now. "

    Yeah, I could imagine it, I have to deal too often with this so nice anti-spam filter as well :P

    "Sure but if I'm going to write an entry I should link to the official page, even if I also hang the paper at ZohoViewer for example and link to that as well. "

    I think it was published first here:

    http://onlinelibrary.wiley.com/doi/10.1002/ajpa.21161/abstract

    (American Journal of Physical Anthropology)

    ReplyDelete
  10. "But 30% seems too high to be that recent! I agree that a Neolithic origin for most of these haplogroups it's more reasonable, unless the Phocaean Greeks killed most Provençal men".

    It's not necessary that they kill them. The Neolithic settlers may have been concentrated originally and later spread, adopting local girls as their partners.

    "I'd say it's a 25% because I2a is possibly of SW European origin".

    I2a could be Neolithic in southern France though. It may have originated in the Balkans. To me R1b1b2 is the only certainly pre-Neolithic West European Y-hap here. That makes 40% Neolithic Y-haps.

    ReplyDelete
  11. Let it be clear that, according to archaeology, only a couple of colonies are documented in the area of Nice, not really Provence but Côte d'Azur (a territory originally Italian by language). Otherwise there is no known settlement from outsiders in that key early moment when founder effects were most possible. But there is always the possibility that we are misinterpreting the data somehow.

    However Provence may have received more and more fluxes from various times, from around the Mediterranean, not just Phocaea and other parts of Greece but Italy and the Catalan Countries, where Neolithic colonization was surely a bit stronger overall at first.

    "I2a could be Neolithic in southern France though. It may have originated in the Balkans".

    It is possible indeed but, if correct, we are before a very strong founder effect, as in the Balcans it has only been located in an inland sample (Albanians from Macedonia) and at frequencies equal or lower than those in the West.

    "That makes 40% Neolithic Y-haps".

    I certainly do not consider R1a and I1 to be Neolithic. They are Bronze Age (Celtic) surely. However a 12% Celtic-specific lineages may well mean a 20% Celtic paternal input overall (after comparison with the South German genetic pool, where Celts originated).

    ReplyDelete
  12. I
    I haven’t yet had the time to read the paper, but ,watching the diagrams, I, that have always refused the Aegean origin of Etruscans, had to say that it is more likely that the inhabitants of Alalia (where Etruscans defeated Greeks) and of South France, which has had a massive Italian (above all Tuscan) immigration, are of Etruscan/Asian descent more than Greek.
    II
    Probably the paper hasn’t demonstrated its assumptions, but has put at our disposal many interesting haplotypes. Amongst the R-M269 from Smyrna and Phocaea (23) 6 have DYS385=11-11. Argiedude was the first to speak of a Balkan haplotype of R1b1b2a (now R1b1a2a), found also in Ciulla, a Sicilian of probably Arberesh origin. I think having demonstrated that the Balkan haplotype is very uniform and found in Italy an haplotype with more variance (Risso: ySearch EK3WY), then more ancient, and spoke of an Italian origin also for this haplotype. See now this haploype found in England (Beacham: ySearch TN2ZH), which demonstrates the origin of this cluster in Western Europe (I of course think in Italy) rather than in the Balkans.

    ReplyDelete
  13. These samples are weird. If I had analyzed them without knowing their origin I would've sworn they're from Greece, Albania, or Macedonia.

    ReplyDelete
  14. Oh, God, the stupid blogger filter again. Ok, to summarize, and I hope this gets through, I posted a 5 paragraph explanation at Dienekes blog, but I had to post it paragraph by paragraph, 5 different posts, because of blogger. Jesus. I'm not going to do that again. Go to Dienekes blog and read the detailed analysis I made, concluding that the samples are either mistaken, meaning they erroneously analyzed some Balkan samples instead of the intended Smyrna samples, or we have something very special here, which definitely requires further investigation.

    ReplyDelete
  15. test

    paragraph break, will it make it through?

    ReplyDelete
  16. test

    paragraph break, copy/pasted from text editor

    ReplyDelete
  17. Jesus Christ, these assholes at blogger know how to infuriate people. I've written a detailed explanation at Dienekes blog about something very important regarding these samples, I had to post it paragraph by paragraph. This blogger bug is really bugging me. I hope this post makes it through. Second attempt.

    ReplyDelete
  18. @Gioello: the maps, as much of the paper is totally confusing. There is only one sample of 51 Provenzals and no Corsican, as far as I can tell.

    The red dots only mention Greek (Phocaean) colonies but we all know that Alalia was Greek only for a brief time as an Etruscan-Phoenician coalition expelled them a few decades later for good.

    There is, I insist, no Corsican sample. Only a Provenzal sample and I have not been able to discern from exactly where.

    "Probably the paper hasn’t demonstrated its assumptions, but has put at our disposal many interesting haplotypes".

    That's true and that's why I mentioned the paper and contained myself from going on verbal rampage against it. I believe I managed to remove totally the word "junk" which featured prominently in my first three drafts.

    "a Balkan haplotype of R1b1b2a (now R1b1a2a)"...

    Since when? Per ISOGG M269 describes R1b1b2. I do not care what internal babbling they may go on about the nomenclature at FTDNA but it's not official unless ISOGG says so or unless ISOGG is deposed as the reference on the matter. None of which has happened as far as I can tell.

    ReplyDelete
  19. @Argiedude:

    Greece or Albania would typically have 20-35% E1b1-V13. This has just 8%, which is pretty much standard for many Western Mediterranean regions.

    I hope you manage to dribble Blogger. Otherwise, email me (email address at my profile, remove antispam protection DELETETHIS) and I'll post in your name.

    ReplyDelete
  20. Ok, Argiedude, on your request (above) I'm replicating here what you wrote at Dienekes:

    These samples are weird. If I had analyzed them without knowing their origin I would've sworn they're from Greece, Albania, or Macedonia.

    R1b1b2 - Albanian cluster. This cluster isn't formally recognized yet, it doesn't have its own SNP, but there's no doubt about its validity, it has one of the most easily recognizable and divergent modal haplotypes of any R1b1b2 cluster. 6 of the 89 Turkish samples in this study belonged to this cluster, or 7%. The cluster makes up 15% of Albania's y-dna, 4% of north Greece, 4% of south Greece, 0% in northeast Greece (the panhandle), 2% in Macedonia, 16% in south Serbia (Kosovo), 2% in central Serbia, 1% in Bulgaria, and 0% in Romania. In Italy there are 0 out of 1000 samples, in Croatia there a few out of 1250, and in Turkey there are 0 out of 1500 samples, underscoring the extremely limited range of the cluster. Yet this study found 6 out of 89 in northwest Turkey. Also, the 6 samples of this study have a low diversity, which is normal for this cluster, but not an ultralow diversity indicating a very recent "Ashkenazi-style" bottleneck.

    The study found M78+ V13+, aka E1b1b1a2b, at a whopping 13 out of 89, or 15%. Incredible. V13+ is also centered in Albania. It's found in Turkey at barely 1%, at 7% in Greece, 6% to 8% in south to north Italy, and 25% in Albania (this last figure calculated with just 50 samples).

    This is so strange, I think we either have a huge sampling mistake, for example, they mistakenly analyzed a set of Balkan samples instead of the intended northwest Turkey samples, and if not, then this is a very interesting discovery that requires further investigation.

    ReplyDelete
  21. @Argiedude:

    So you are saying that, within the R1b1b2 of Provenzals there is a haplotype that, as far as you can tell is Albanian (or alternatively Greek from peninsular Greece) and hence very closely associated to the core of E1b1b1a1b (V13) which is also found among Albanians and Greeks.

    If so this would indicate (in a positive, rather than a negative way) that effectively the E1b1b1ba1b of Provence has its origins in the SW Balcans and is therefore probably Neolithic by origin (not Phocaean).

    "aka E1b1b1a2b"

    Not too important but right now ISOGG calls it E1b1b1a1b.

    "The study found M78+ V13+, aka E1b1b1a2b, at a whopping 13 out of 89, or 15%. Incredible".

    Where? There is no n=89 sample anywhere that I can see nor 15% apportions of that lineage either. In Provence E-V13 is only 4% (2/51).

    Either you or me are looking at the wrong place. I am looking at fig. 2, are you exploring some obscure supplemental material or something?

    ReplyDelete
  22. [I use livejournal and they've sent out messages over the last week informing users that they're having some recent trouble with their site being placed on a spam-block list and they're working to remove this. Maybe that's my posting problem. Anyone else use livejournal to post here?]

    Maju, the % of V13 in northwest Turkey is 13/89 = 15%. Look at the SNP table in the pdf, it shows 6 V13 from Phokaia and 7 V13 from Smyrna.

    Meanwhile, I've noticed that the 16 yhrd samples from the Greek island of Chios, very close to Smyrna, have 5 M78 samples, indicating a very high rate of V13, close to 30%. Keep in mind that M78(xV13) is very low in both Greece and Turkey, so at best, maybe just 1 of them is M78(xV13), but that still doesn't change the overall picture of a very high V13 percentage. This is very interesting.

    ReplyDelete
  23. Just incredible, there's just no reason behind why some posts are blocked, it's just random. My 2nd post, which I lost, wasn't allowed, that's why I posted the 2 tests afterwards. Let's see if this goes through.

    ReplyDelete
  24. Second attempt to follow up my original post.

    Yesterday Dienekes told me my percentages for V13 for Greece, 7%, were way off, and he was right, it's more like 25%. I was using ysearch samples from Greece, but now I'm realizing as much as half of these samples could be Jewish Greeks, and from what I've seen of the dozen samples that explicitly indicated they were Jewish Greek, they're haplogroups are indeed notably different from Greeks. Together with the high rate of V13 in the island of Chios, these new samples from Smyrna and Phokaia are starting to look like they're really ethnic Greeks in coastal Turkey.

    [Oh, God, here we go, don't f*** me blogger crapola...]

    ReplyDelete
  25. Look at WOT's livejournal card. There are 23 comments of spam, 34 of bad customer experience, 27 of adult content, 7 spyware or adware, 4 of annoying ads or popups... It has some positive comments as well but the negative ones clearly dominate.

    It does have a good overall rating now but it's obvious that there's some people pissed off with them. For example PeterGut says that many subdomains are used by spammers... I do not see any mention of blacklisting but it's clear that it has poor reputation in some corners.

    What message do you get anyhow? Because random errors make sense but "blocking" implies getting a message that says your post is too long or something.

    :(

    ReplyDelete
  26. "Maju, the % of V13 in northwest Turkey is 13/89 = 15%. Look at the SNP table in the pdf, it shows 6 V13 from Phokaia and 7 V13 from Smyrna".

    How confuse!

    There is a region called "Northwest Anatolia" (T1), which surely corresponds with Marmara Sea area. Smyrna and Phocaea actually correspond to Western Anatolia, which also has a sample (T8, these are from an older paper, Cinioglü?) Smyrna and Foça are Smyrna and Foça and they do of course look Greek in the "Chios" style of Greekness (not all regions are equal).

    I thought you'd be talking of Provence.

    ReplyDelete
  27. Phokaia and Smyrna are the new samples from this study, T1, T7, T6, T8 are rehashed from Cinnioglu's 2003 study of Turkey, retested for some new SNPs, such as V13. The table also includes 3 populations from Greece, each has exactly 57 samples, and they're also rehashed, from King (2008). Finally, at the bottom of the table we see Provence, which the study says are brand new sparkling samples.

    I'm ok with blogger's anti-spam attempts, despite my cussing at it. I've read up on the problems of spamming on blogs, and wow, it's for real. But anyhow, I think I just figured out the mechanics of why I was being blocked in an apparently random manner, I've already managed to post 3 comments in a row here and in Dienekes blog without a problem, let's see if this will be the 4th...

    ReplyDelete
  28. Woo-hoo! Yes. It's very simple. Anything that you send once is cached by blogger. The 2nd time you attempt to resend even a part of that original text... Block!

    Gioiello, the Italian sample with surname Ciulla has stated he is indeed Arberesh, aka Albanian Italian, and the other Italian sample, ySearch EK3WY, doesn't belong to the cluster, because the most important feature of the Albanian cluster is 459a/b = 8/10, and he has 9/9.

    ReplyDelete
  29. Waht I say is that Foças and Izmir are distinct samples from NW Anatolia (T1) are not even in NW Turkey (Bursa, Istanbul) but rather by the Mid-West part of it.

    "Woo-hoo! Yes. It's very simple. Anything that you send once is cached by blogger. The 2nd time you attempt to resend even a part of that original text... Block!"

    Maybe. So when you just copy-paste your comments at Dienekes here, it blocks you, but if I do (a different agent) it does not (it's a quote).

    Still, it should send to the spam folder not just "block" it with a random error, right?

    However it may be the reason why I do not perceive many issues with spammers here. :/

    ReplyDelete
  30. @ Argiedude
    “Gioiello, the Italian sample with surname Ciulla has stated he is indeed Arberesh, aka Albanian Italian, and the other Italian sample, ySearch EK3WY, doesn't belong to the cluster, because the most important feature of the Albanian cluster is 459a/b = 8/10, and he has 9/9”.

    Of course I noted it too, but…
    You presuppose that DYS459=8-10 is the fundamental datum of this cluster, but the most important is above all DYS385=11-11, that you can say it is diffused in other haplogroups.
    But we are speaking of R1b1b2a (if you don’t like R1b1a2a like FTDNA). How many samples have you of R1b1b2a with DYS385=11-11 in other clusters? Probably none, and my Italian sample and the British one are. Risso is clearly R1b1b2a, having DYS461=10, which presupposes a previous 11, and it has then a more variation of the “Albanian” one. Probably Dienekes is right in saying that this haplogroup in the Balkan isn’t “Albanian”: We have the Perovic (J6BPG) from Croatia. on the ht35 project of Vizachero, and are R1b1b2a with DYS385=11-11 also Seguer from Spain and Zold from Hungary, which maintain DYS393=12, whereas the “Albanian cluster “ has 13. I suppose that Risso belongs to this haplotype, amongst other reasons, for having DYS393=14, which presupposes a previous 13. At this point DYS459=8-10 couldn’t be the discriminant point. We can presuppose that also Beacham is of the cluster of Risso for having DYS393=14 and DYS388=14 (Risso has 13). Beacham has also DYSD461=12 I presuppose for 11, like Risso who has a mutation from 11 to 10. Then these haplotypes, if are of the same cluster of the “Albanian” ones, have more variation.
    Of course we should test them for these fundamental SNPs.

    ReplyDelete
  31. Maju, sorry about that, I see the confusion now. When I think of northwest Turkey, I'm thinking of the northwest of the peninsula. My bad.

    Gioiello, there's only 1 legitimate Albanian-cluster sample amongst all those you mentioned, and that's the Croatian sample. Having 385a/b = 11/11 is a rare but not alien feature of R1b1b2 in general, I myself have 385a/b = 11/11 and I'm L21. This cluster is very tight, like M222 of Ireland, so it's not going to drift too far from the modal. The modal is 393 = 13, 385a/b = 11/11, 458 = 16, 459a/b = 8/10, 461 = 11.

    ReplyDelete
  32. NW Anatolia Peninsula: it's the same. Specially, since the Cinnioglu nomenclature is there calling NW to the NW and not to the mid-West (Izmir, etc.), which it calls simply West.

    But never mind. :)

    ReplyDelete
  33. Let's see if this message will pass the spam blocker, because I posted 5 minutes ago in another blog.

    Check this out:

    V13 and R1b-Albanian cluster map

    Note that the Albanian cluster has its highest frequency in the Penopolesse peninsula, at 35% (n=93).

    ReplyDelete
  34. The map is interesting, Argie.

    "Note that the Albanian cluster has its highest frequency in the Penopolesse peninsula, at 35% (n=93)".

    Yes. It must be part of a founder effects' pattern anyhow, because it's also high in Thessaly, specially in the key coastal area of Sesklo/Dimini, which is so far the main archaeological reference of European Neolithic origins. There it reaches 29% and may have been even higher before Indoeuropean and other older (Vinca-Dimini) invasions maybe.

    More arduous is to explain how that lineage arrived to Greece and Albania but once there its expansion within the Neolithic flows (primarily at least) looks fairly obvious, specially as it spread by the Mediterranean coasts (Otzaki-related Cardial Neolithic) and the Morava-Tisza axis (Sesklo-derived Balcanic Neolithic).

    Now, maybe to Izmir and Foça it spread with Greek colonization, that may well be the case because the nearby region of West Anatolia has, it seems, 0% E-V13.

    ReplyDelete
  35. @ Argiedude

    If a SNP for your “Albanian cluster” was found, I think you should take in consideration not only

    Da Silveira (Brazil): 58HRZ: 13, 11-11, 16, 8-10,11

    but also
    Anonymous from Slovenia (Slovenia): B2A4W: 13, 11-11, 16, 8-11, 11
    Anghel (Romania): SXRNE: 12, 11-11, 15, 8-9, 11
    Yuran (Croatia): YARC5: 12, 11-11, 15, 9-10, 11
    etc. etc.

    ReplyDelete
  36. Gioiello, you've inadvertently made an interesing discovery. The Romanian sample doesn't belong to the Albanian cluster. Notice that it has the very weird value 392=11. There are several other samples just like this one, all with 393=12, 385=11/11, and 392=11. They're all from southeast Europe. I only knew them from yhrd, so I didn't know their values for 459 and 461. Now thanks to SXRNE we do, and the results link this cluster with the Albanian cluster. It has 461=11, which together with 392=12 means it's almost certainly L11-, like the Albanian cluster. And it has the unusual value 459=8, found in only 3% of R1b1b2 samples, but which is modal for the Albanian cluster. Both clusters are L11-, have 385=11/11, and 459a=8. They're sister clades. The rarer one is distinguished from the Albanian cluster by having 393=12 and 392=11.

    ReplyDelete
  37. And judging from the yhrd results, this rarer L11- clade (392=11) is found at around 0,5% in Greece, Albania, and Romania. Perhaps also in Bulgaria, but there aren't enough samples from that country, just 100.

    On the other hand, it's completely absent from Turkey, despite having 1000 total samples in yhrd. And from Italy, despite having 3000 total samples. This pattern, of pan-southeast-European distribution together with total or almost total lack of presence in both Italy and Turkey, is also observed for the Albanian cluster.

    ReplyDelete
  38. Could that be the possible Balcanic root of R1b1b2 apparent in Morelli's paper?

    ReplyDelete
  39. @Argiedude

    I have found many haplotypes with DYS392=11 all over the world but I didn’t put on ySearch because I thought they were R1a1…, but they have DYS438=10. There is a convergence of this haplotype with some R1a1…, but this Romenian seems just an R1b1a2a. Anyway probably this haplotype, you call “Albanian”, seems like the R-L277+, i.e. diffused from ancient time all over the Indo-European world and only a SNP could say us which variations can have.

    Other candidates could be
    Alves Dos Santos (Brazil): ySearch M57PY: 13, 11-11, 16, 9-10, 11
    Lawson (United Kingdom): ySearch YJ8U7: 13, 10-11, 16, 9-10, 12
    Dore (England), ySearch RC95S: 13, 11-11, 16, 9-10, 11
    Thornton I haven’t yet put on ySearch.

    ReplyDelete
  40. @Argiedude

    By watching all these clusters of R1b1a2a I think it is clear that the core is my haplotype (KV7Y2) if you correct my DYS392=12 to 13 and DYS447=24 to 25 (see Del Badia: HFXBJ), demonstrated by my closest relatives (Giancarlo Tognoni (HG6HX), separated m me during the 15th century and an Anonymous Brazilian (N92C5) and now by the Fluckigers /Flickingers (GJKPY and S6EVJ), separated from me during the first centuries after Christ. I am testing a Fluckiger/Flickinger for S136 c/o EthnoAncestry to understand when this SNP was born).
    From my haplotype derived both the Balkan cluster (you call “Albanian” but now we can call “Balkan”) and the West European ones (Portugal/Brazil and British Isles). It is also clear for me that the subclades of R1b1a2a were born from my haplotype and not from others. Now the L150- of Romitti is involved again.
    Italian R1b1a2a is in the core of all subclades, beginning with the Italian ones: see Belgieri for R-U152 etc. The Risso cluster (EK3WY), I have linked with the Albanian/Balkan one, I think it is actually a remnant of those ancient cluster developed in Italy and eventually migrated to the British Isles (see Beacham: TN2ZH).
    At the same time I think that your R-L21 could be the remnant in Italy of those subclades which migrated to North-West, where developed intensively.
    As you can see my idea of the Italian refugium is stronger than ever.

    ReplyDelete
  41. This comment has been removed by the author.

    ReplyDelete
  42. This comment has been removed by the author.

    ReplyDelete
  43. Is that the same root(?) haplotype found by Morelli in the Balcans, Gioello?

    She did not test peninsular Italy but she clearly oversampled Sardinia and did not find it there.

    If so, it may be shared by parts of the Balcans and parts of Italy but it'd be interesting to know which ones.

    Whatever the case, the lineage does not imply, IMO, refugium, but original colonization or even ultimate origin, together with R1b1a maybe. LGM refugium is best described by R1b1b2a1a2 (S116), a much more derived lineage. Hence the "root" haplotype discussed here, which probably is R1b1b2* (very apparent in this map) would be the shared root of all the lineage - which, as in the case of R1b1a, could be Balcanic, Italian or West Asian.

    ReplyDelete
  44. This comment has been removed by the author.

    ReplyDelete
  45. You did post, Gioello. It just went to the spam folder (but those subscribed read it in our mailboxes).

    It is corrected now.

    ReplyDelete
  46. I apologize for my fourfold posting, but it didn’t depend from me. Probably by my posting I have already answered your question. This at least is my thinking so far.

    You say: “LGM refugium is best described by R1b1b2a1a2 (S116), a much more derived lineage”.

    But thinking that R-S116 is so ancient, with Vincent Vizachero and others who give it a few thousands of years, perhaps is hazardous. The true question is the time. If you are thinking to an R1b Paleolithic, I could agree with you about its origin from Asia and a development of its in some European refugium, but others pretend that it is arrived to Europe only a few thousands of years ago and that Europeans are only Middle Easterners and have nothing of our Paleolithic hunter-gatherers. This would agree with the presence of R-S116 above all in Hiberia, but if the times are shorter (R1b1b2 at the Younger Dryas) your hypothesis becomes hardly sustainable, seen that R-L51 seems having being irradiated from Italy (this was the map of Argiedude and I don’t know if he changed his mind).
    About the Albanian/Balkan cluster I think that Argiedude is correct: its presence is where he said. But the fact that it lacks in Asia Minor (and not only in Italy) makes my hypothesis likable: an irradiation from the Italian Refugium and above all the origin of the R1b1b2a subclades from an haplotype like mine. Many questions are yet open: where was born the R1b1b2/YCAII=17-23, present above all in Italy? Was really the modal of R1b1b2/YCAII=18-23 and not 19-23? Has the SNP L150 all the importance I gave it, now that FTDNA has recognized it? But I should cite what I have written in these last years. Anyway when I write something I remember very well everything.

    ReplyDelete
  47. "But thinking that R-S116 is so ancient, with Vincent Vizachero and others who give it a few thousands of years, perhaps is hazardous".

    I'd say it's "hazardous", rather ridiculous, to give it just a few thousand years, sincerely, contradicting most serious researchers on the matter, btw (neither Vizachero nor Dienekes are professional geneticists, mind you: their opinion is as good, or as bad, as mine or yours).

    Here we must apply common sense: if R1b1b2a1a2 is Neolithic, then:

    (1) It must have originated in Yugoslavia/Albania (Cardium Pottery culture) - reality check: nope, the highest diversity by far is in SW Europe, probably in the Franco-Cantabrian region (we lack clear data about Iberia proper ATM).

    (2) As Cardium Pottery does not imply a total population replacement, as both archaeology and archeogenetics show, but rather a quite mild localized colonization at most, there must be other "older" lineages which represent the surviving and more or less assimilated Paleolithic peoples. There is no such thing, much less at the levels we would expect (i.e. dominant or at least a large minority). Only R1b1b2a1a2 can represent that survival.

    "About the Albanian/Balkan cluster I think that Argiedude is correct: its presence is where he said. But the fact that it lacks in Asia Minor (and not only in Italy) makes my hypothesis likable: an irradiation from the Italian Refugium and above all the origin of the R1b1b2a subclades from an haplotype like mine".

    I can't say. Judging from Morelli's haplotype graph, the "root" haplotype is shared by Balcans and (very limitedly but also) Turkey.

    Overall the R1b1b2* "root" clade does extend at levels above 5% between the Balcans and Iran, with smaller penetrations into Italy and parts of Central Europe, which may be old or recent (Neolithic).

    It does not look any signature of any Italian refugium.

    Also there is an archaeological problem with your hypothesis of an Italian refugium: Italy remained Epigravettian (with the exception of some parts of Liguria maybe) until the very arrival of Neolithic, yet there is no Epigravettian technoculture in Central Europe after the LGM as far as I can tell. Of course it is not impossible that some people could have migrated northwards and adopted Magdalenian tech instead but with the still frozen Alps in between, the migration looks pretty much unnatural, to be honest.

    ReplyDelete
  48. Maju, what are you saying about the Morelli study? I'm looking at it, and I think there are just 2 L11- samples, with 393=12 and 461=10. Sardinia doesn't look like a hotspot for L11-. On the other hand, it has a very interesting array of R1b1* (whose technical name is no longer R1b1*, I have to get up to date).

    ReplyDelete
  49. You have archaelogical knowledge and you think by an archaelogical point of view and it is possible you are right, but your map isn’t in contrast with what I said. It seems that actually Italy has the highest level of R-L51 and that all the European subclades of R-L51 presuppose my haplotype of R-L23 and not others isn’t disproved. You are thinking by an archaeological point of view, I by a genetic one.
    My theory was (and is) that the massive presence of the previous haplogroups downstream R-L51 in Middle East was due to the first expansion from the Italian refugium. I think having demonstrated that some mitochondrial haplogroups was born in Italy and diffused to the regions where now we find the massive presence of R-L23, which is above all linked there with the Armenians, and they were Indo-Europeans come certainly from the Balkans. My theory was that the first expansion was to the Balkans till the Black Sea and we were speaking above about this Albanian/Balkan cluster which lacks in Anatolia, but my theory is that it is present with its more ancient haplotypes in Italy, in Hiberia and in the British Isles. Of course only a SNP could link all these haplotypes, but we already have L277+, M405, my S136 (so far private, but I have said above that I am doing (and paid) a test c/o EthnoAncestry). Of course all this is linked also with the origin of the Indo-European. My theory is that there was a link between the most ancient phase of IE and Etruscan-Rhaetian-Camun, and that IE was born in Central Europe and that the Centum languages were spoken by people above all R1b and that there was in Central Europe a link of Germans, Celts an Italic peoples.

    Actually I am studying the R1b1b2* haplotypes: as I have said above the most ancient for me is that with CDYII=17-23, above all Italian, with a presence on the Rhaetian region and with an only Lebanese I put on ySearch: Jlelaty. By chance, a relative of mine, born in my conservative region (the Monti Pisani from where I have had my K1a1b1 (99932A) ancestor of the Ashkenazi K1a1b1a (is the Jew Ted Kandell to say this)) is R1b1b2* ancestor of the Jewish one. It is also possible that we all descend from the Jews who migrated from Italy to the Rhine Valley. It is this I am trying to understand.

    ReplyDelete
  50. This comment has been removed by the author.

    ReplyDelete
  51. This comment has been removed by the author.

    ReplyDelete
  52. @Argiedude:

    Sorry, we are confusing each other: Morelli did not sample nor report any STR haplotypes other than the Sardinian ones (which do not include any haplotype like the one you say), the rest are all from Cinioglu 2005 (now pay per view).

    But, as you (or is that Gioello) seem to claim that the rare "Albanian" haplotype is somehow at the root of R1b1b2, I was wondering if it could be the same displayed at the center of the haplotype structure at fig. 2A, which I argued in the relevant post at Leherensuge could be the root of R1b1b2, with a Western branch R1b1b1b2a1 and an Eastern (Anatolian mostly) branch unnamed but that I suggested that could be a distinct subhaplogroup "R1b1b1b2a2" (??).

    If we exclude the Neolithic origin, particularly, there's absolutely no archaeological reason to exclude a "back-migration" from SE Europe into Anatolia. It remains in the real of hypothetical ideas at the moment anyhow.

    "Sardinia doesn't look like a hotspot for L11-".

    I was all the time talking of the Balcans, not Sardinia. Look at fig. 2A or my edited version of it to see how the Balcans can become the center of R1b1b2 as a whole, which I understand is the same idea suggested by you (and/or Gioello) above.

    ReplyDelete
  53. @Gioello:

    Sorry about the Blogger Spam Filter charging against you again. I'd suggest you write a comment of "protest" to me instead of repeating the same post once and again next time it happens. I'll fix the matter as soon as I read it.

    Anyhow, can you go over duplicated posts and delete them. Because I'm not sure how exactly duplicated they are and anyhow it looks better if deleted by author than if deleted by admin, which may look like censorship. If you don't I will have to but I prefer you do it yourself if you can. Thanks.

    ReplyDelete
  54. @Gioello:

    "It seems that actually Italy has the highest level of R-L51"...

    You must mean R-L51/M412(xR-L11), of course. I guess that may be correct (by frequency) though have you contrasted with Slovakia? Though not so high also parts of Spain (and SE France) have noticeable frequencies of this transitional paragroup.

    Thinking genetic primarily: sure: R1b1b2a1* (L51/M412) indicates a remnant of when R1b1b2a1 coalesced, no doubt. So it's reasonable to think that R1b1b2a1 coalesced in the Spain-Italy-Hungary arch and I'd agree that we can probably discard Spain here. And an Italian (or Slovak?) origin is indeed possible.

    But what does that mean? The Aurignacian (or maybe Gravettian) colonization of West and Central Europe, including Italy.

    Notice also how the R1b1b2a1* (L51) is in the North, not the South, automatically demolishing the Neolithic hypothesis (because the North did not play any relevant role in the general spread of CP Neolithic, the South and Center did instead).

    It is very clear case of signature of the time of the colonization of Europe by H. sapiens, which went through North Italy most likely into the Rhône valley and beyond.

    I can't decide if the origin is in North Italy or Slovakia on genetic reasoning nor on archaeological ones.

    It's anyhow too fine threading, I fear, considering the limited knowledge overall that we have.

    "My theory was (and is) that the massive presence of the previous haplogroups downstream R-L51 in Middle East was due to the first expansion from the Italian refugium".

    There's no "massive presence" of almost anything downstream of L51. All West Asian R1b is R1b1b2a-L23(xR1b1b2a1-L51) or upstream R1b1b2*. Only West Turkey has some of European-derived R1b (in any case not "massive").

    I do not understand where you get that idea of "massive presence" of Euro-derived lineages in West Asia. There are some but limited and mostly in the Westernmost range.

    "we already have L277+"...

    What do we know about L277? I missed this new discovery and looks intriguing.

    "Actually I am studying the R1b1b2* haplotypes"....

    Honestly I get lost. Now it's R1b1b2* and earlier it was R1b1b2a1* (L51). Which is what you have in mind and where you belong by ancestry? They are two different categories that I understand very differently: different stages in the long process of coalescence of the much larger Western haplogroups and of what may be an Eastern parallel haplogroup as well (following Morelli's fig.2A haplotype structure).

    ReplyDelete
  55. @Maju

    Of course I did mean "upstream" and not "downstream". It was a mistake. I'll write to you this afternoon, now I must go to work. Anyway many thanks for your replies (you and Argiedude), I think very interesting.

    ReplyDelete
  56. "Of course I did mean "upstream" and not "downstream". It was a mistake".

    Ah. That changes everything, of course. :)

    ReplyDelete
  57. @Maju

    I ask you to delete my double postings as I cannot do it.

    Of course I am R1b1b2a (FTDNA R1b1a2a): L23+,L49+,L150+. But my interest on R1b1b2 (R1b1a2) is due simply by the fact that it is the direct ancestor of my haplogroup and Italy has interesting haplotypes like the CDYII=17-23. My relative Federighi belongs to the cluster of R1b1a2 with DYS462=12, a slow mutating markers (one mutations every 10000 years), found among Jews, but as usual with a little variance, but also in Iraq (practically one haplotype with no variance (one or two mutations) and with the weird DYS452=28 and not the modal 30, this is a slow mutating marker too) and also one guy in Benin, we don’t understand where come from. There are also some guys from Hiberia (Spain. Portugal an Latin America) I have put on ySearch. This haplotype is interesting, I think, because probably is the closest to my R1b1a2a and to all its subclades. For this I think that where it is born, probably there was the refugium. Of course I am not sure that it was Italy, but I found this by chance in a relative of mine, one of the two persons I was able to test. For this I think that my region, above all in the mountain and isolate lands, probably has many interesting haplotypes. My relative is also mtDNA R0a, found in some papers on the Etruscans and interpreted by Brisighelli et alii like a sign of a Middle Eastern origin, but this haplotype is very different from the Middle Eastern ones and probably it is in Tuscany at least from 10000 years and perhaps from more. It would be interesting to do an FGS.

    About this new SNPs you could read some threads on Worldfamilies: my “R-L227+ and the Jewish clades” and now “R1b1b2a2: L23+ L277+” by Wayne Kauffman, who announces the new SNPs around R-L23: L479, L482, L483, L500, L502, and L506.

    ReplyDelete
  58. "I ask you to delete my double postings as I cannot do it".

    You did it already. Thanks. :)

    I'll delete the remnant message if you wish, though now is already a lot clearer.

    ......

    If the haplotypes are shared by Italy and Iberia, what we see is not a LGM refuge, which should leave a south-to-north flow pattern (post-LGM recolonization), but something else: probably an East to West flow.

    This may have happened in two processes:

    1) The early colonization of Europe (proto-Aurignacian, Aurignacian and/or Gravettian), in all of which North Italy is important.

    2) The Neolithic flows (Cardium Pottery particularly), where the main role belongs to South and West-Central Italy.

    As the key transitional stage R1b1b2a1* (L51/M412) is found in North but not South Italy, I think this is evidence supportive of the pre-LGM colonization process being the one.

    ...

    As for mtDNA R0a is of course a West Asian clade, almost unmistakably so. But there is nothing impeding it to have arrived in the Neolithic, specially if, as you suggest, it is a quite distinctive clade.

    But rare lineages are difficult to track. There are probably thousands of people out there carrying them but almost no one is tested.

    ReplyDelete
  59. @ Argiedude

    If DYS385=11-11 and DYS459=8-10 define this Albanian/Balkan cluster, then we have probably a North European version with DYS393=13, DYS461=12 etc.
    See Kneebone (Usa, but probably of Northern European descent, living in Norway, Michigan) ysearch : 7PM8V and Tienken (Germany) ySearch: NWX7A.

    They have also some weird results, that probably demonstrate their ancientness:
    Tienken DYS441=8
    Kneebone DYS389=13-26.
    If they belong to this cluster they should be R1b1b2a (R1b1a2a). Of course only a SNP test could make us sure.

    ReplyDelete
  60. @Argiedude

    I think you should test yourself for this new SNP L459. They are
    searching for the presence of this SNP upstream R-L21, but perhaps
    more interesting could be to search downstream R-L21, i.e., if you are
    amongst the most ancient R-L21, you could be L459-.

    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).