Sunday, September 13, 2015

Basque vs Danish


  • Image result for basque flag VS Image result for danish flag


In my last post I discovered large differences between West Asian and European mtDNA with low coverage mtDNA data. Within Europe  and most of West Asia it is impossible to find such large differences between differnt regions/ethnic groups with low coverage mtDNA data.

Now, I have been able to find a great amount of mtDNA differences between two European ethnic groups: Danish and Basque, because I have high coverage mtDNA data from both of them.

Here's my comparison analysis of Danish and Basque mtDNA.
Basque, Danish mtDNA H
Basque, Danish Other mtDNA
Basque vs Danish mtDNA Frequencies

All of the Danish data was fully-sequenced but for Basque only their H was fully-sequenced.  So, I wasn't able to find a lot of differences between most non-H in Basque and Danish.

H takes up nearly half of European mtDNA and 90%+ of H remains undefined with low coverage testing. Fully-sequenced Danish and Basque H  reveals European H is far from uniform. Nearly 50% of the H from each appears to have separated from each other 8,000+ years ago.

Most noteworthy is that about 60% of H1/H3 in Basque and Danish probably separated from each other 8,000+ years ago. This is noteworthy because H1/H3 take up 40-50% of pan-European H and are very rare in West Asia. They're probably the most important part of the history of H in Europe.

Outside of H Danish differentiation with Basque reflects their higher amount of "Steppe" ancestry(U5a, U4, T1a, J1b1a, J2a1a1, and I).  Outside of H Basque differentiation with Danish mostly reflects Basque-specific founder effects(U5b1f1a, J1b1a-16368, J1c5c1) and a higher amount of "EEF/WHG" ancestry(K1a, HV0(xV), T2(xT2b, T2c1, T2a1b, T2f, T2e)).

27 comments:

  1. Lots of data and admittedly hard to process but very much appreciated, because this kind of most interesting info is not readily available, and therefore we tend to use simplified data, what in some cases may not be enough.

    I'm a bit perplex about anyhow about the third sheet, because rather than frequencies it lists a bunch of haplogroups but with no associated figures, except for what seems a synthesis to the right. Is that an error?

    ReplyDelete
  2. Do you still think that mtDNA haplogroups H1 and H3 are likely Anatolian Neolithic farmer mtDNA lineages that spread into Europe with farming?

    ReplyDelete
    Replies
    1. Yes, I tend to think most is.

      Delete
    2. At the very least for H1, that does not fit the data, because there is some obvious H1b in Portuguese Epipaleolithic. Re. H3 I don't have a strong opinion as of yet for lack of data.

      Delete
    3. This is why I think H1 is mostly from Neolithic Anatolia:

      H1 was found at high frequency in two separate sites from Neolithic France. And about 1/3 of fully sequenced Neolithic H from Spain and Germany is H1.

      A recent abstract of an upcoming paper with Neolithic Anatolian genomes says Early Neolithic Europeans had very little hunter gatherer ancestry. So, considering how much U5 Central European hunter gatherers had and how little U5 and high H1(maybe 5-10%) Early Central European farmers had, it doesn't make sense for their H1 to be of hunter gatherer origin.

      All fully sequenced H1 from Germany(including LBK) is H1j, H1e, and H1*. H1j and H1e in the top 3 of H1 clades in Basque. I guess Western hunter gatherers could have also carried H1j/H1e, but it makes more sense Basque got those lineages from Anatolia.

      The same is true for H3. Full-blooded Neolithic Anatolian immigrants in Spain carried H3 in two separate sites. H3 is the next most popular form of H in fully-sequenced Neolithic from Germany, and was found at high frequency in the same two French sites H1 was found at high frequency in. Also, H3 is the second most popular form of H in Sardinians after H1.

      Delete
    4. It is has been proven H1 was carried at decently high frequencies in "EEF". Of course it is possible hunter gatherers in Europe carried H1 to.

      The pre-Neolithic H1b from Portugal is specifically H1b1a-h. I'm skeptical of that paper, because they found several of the same haplotypes(including H1b1a-h which is very rare today) in their Neolithic data. Most of their Neolithic mtDNA was R(xR0) that was labeled as H which is unheard of today. The results just don't make sense.

      Delete
    5. Considering H finds hunter gatherers of Karelia and Spain, I think it's possible some H1 is from before the Neolithic. The fact H1 is so diverse leaves room for that possibility. But as I said I think most is Neolithic, also I suspect we'll see H1 in Neolithic Anatolia in upcoming papers.

      Delete
    6. My working hypothesis is that EEFs picked most of their H on the march, because on this issue of mtDNA there is data from West Asia and presumptive H was not that clear either (R*, just as we have from Europe, but no R*-CRS, which is most likely to be H, unless it is that hyper-rare U* with CRS reverse mutations that has appeared unlikely often in Paleolithic Germany).

      The Chandler 2005 paper definitely did not sequence any clear-cut R(xR0), they only tested for HVS-I and with that data on hand we just cannot conclude if R is or not R0, unless there are clarifying downstream markers, as happens with the various H1b sequences. Per the latest PhyloTree version, R0 is defined by G73A (HVS-II) and A11719G (coding region). You can discard that it is R0a'b but you can't discard it is HV and, within it, H (or some rarer HV subclades).

      As for what you say of several individuals sharing the same lineage, this should not surprise us if they came from the same site, because they could well be relatives.

      IMO Central-North Europe was exceptional in their Late Paleolithic genetic pool, as once and again data from other regions seem to demonstrate, and H was therefore common in all the rest of Europe, be it Karelia or Iberia (both demonstrated beyond any doubt) or other regions like Italy or Britain (where it is suspected) or the Balcans, where data is absent altogether, except for Franchthi cave, where it is unclear if it is Neo- or Mesolithic (but has the longest Meso- persistence in the area and where even farmers were probably of Paleoeuropean stock, as the habitation is continuous).

      IMO everything points to at least part of the farmer H being part of the Paleoeuropean signature in EEFs.

      BTW the paper in the making you mention only seems to suggests that some Western Anatolian farmers were similar to EEFs but that is of little relevance because they did not originate there (no known Mesolithic in Western Turkey) and instead were most likely part of the same overall "Aegean" genesis of European first farmers, probably borrowing their Paleoeuropean blood from the European coast. There must be a prior West Asian Neolithic stage without any Paleoeuropean signature, similar to modern Naqab Bedouins or peninsular Arabs and not too distant from modern Palestinians. In that case, the Paleoeuropean signature in EEFs is more notorious.

      Delete
    7. Also what we see in some data from Central Europe is no or low H1 and instead sizable fractions of H5. This makes even harder to equate H1 with Neolithic.

      Delete
  3. "The Chandler 2005 paper definitely did not sequence any clear-cut R(xR0), they only tested for HVS-I "

    They tested 73. There are a lot of R-CRS with G73G not R0 version G73A.

    "IMO Central-North Europe was exceptional in their Late Paleolithic genetic pool, as once and again data from other regions seem to demonstrate, and H was therefore common in all the rest of Europe, be it Karelia or Iberia (both demonstrated beyond any doubt)"

    It looks like Iberia had a lot, but not Karelia. There's one example of H(H2a2b?) from Karelia. The rest is U5a, U4, U2e, C, and Z1a(spanning 3500 to 7500 YBP). Also, in Siberian HGs all their mtDNA was East Asian and U5a/U4/U2e. So, it looks like H was a small minority lineage in most Eastern hunter gatherers.

    "As for what you say of several individuals sharing the same lineage, this should not surprise us if they came from the same site, because they could well be relatives. "

    There are three examples of H1b1a-h in Portugal from the same study. One is from ~8000 BC, one from ~5000 BC, and one from ~3200 BC. That spans 5,000 years. Family relations isn't a likely reason. Coincidence or contamination are most likely.

    "Also what we see in some data from Central Europe is no or low H1 and instead sizable fractions of H5. This makes even harder to equate H1 with Neolithic."

    Of Neolithic Central European Hs tested for H1 6/24 have it, which would equal a total frequency of about 6%. That's below modern frequencies. Although samples from Neolithic France show modern-like frequencies of H1. Considering farmers from France in 4900-4500 BC were probably 80-100% Anatolian/East Mediterranean(based on recent teaser), it makes much more sense for their H1 to not be from Europe.

    "BTW the paper in the making you mention only seems to suggests that some Western Anatolian farmers were similar to EEFs but that is of little relevance because they did not originate there (no known Mesolithic in Western Turkey) and instead were most likely part of the same overall "Aegean" genesis of European first farmers, probably borrowing their Paleoeuropean blood from the European coast. "

    Paleoeuropean could have also been native to West Asia. Bedouin do have such ancestry. H1 may have existed in Greek Mesolithic, but I think most H1 in most of Europe arrived with "EEF". It's purely based on Ancient mtDNA evidence. There's no way to explain H1 frequencies we see in Early Neolithic remains and Sardinians, considering how little WHG they have.

    ReplyDelete
    Replies
    1. Apparently they did test locus 73, although all the data is only described in fig. 6 of the conference paper. But in that figure locus 73 is clearly placed to separate N(xR0) from R0, so everything to the right side should be what they say it is: H and V.

      Have you even read the paper? Because what you say does not correspond with what I'm re-reading right now but rather sounds like something Jean Manco could say.

      Delete
    2. "Considering farmers from France in 4900-4500 BC were probably 80-100% Anatolian/East Mediterranean(based on recent teaser)"...

      If you mean the ones from Burgundy, surely not: their mtDNA is too modern, only comparable to what we see among (some?) Neolithic Basques and maybe later East Germany's Bell Beaker samples. They should be something Basque-like in their overall genetic pool, i.e. more HG than MN samples available so far, or at least Gokhem-like.

      In these cases we're probably talking of 50% or more HG admixture, of which more than half is novel, i.e. adopted in the Atlantic regions. For instance if you compare EEFs with Basques (using Günther & Valdiosera's ADMIXTURE data), the former have a 3:1 Neo:Paleo apportion but the latter have a 2:3 (with ATP or Gokhem having 1:1), so we are talking of major increases of Paleoeuropean genetics: 50% in the case of modern Basques and around 38% among ATP or Gokhem. The Burgundy and other Atlantic Neolithic French should be in those ranges IMO. And some haploid genetics should be related to it, for example U5b but also H1 or even maybe H3, why not?

      "Paleoeuropean could have also been native to West Asia. Bedouin do have such ancestry".

      How can you claim that? Naqab's Bedouin B are the best approximation to EEF-all WHG. As happens with ANE, there are no absolute component ancestries, just relative to something else. What is that comparison? Because of course West Asians of any time will always cluster with Europeans of any time versus third parties in further Asia or ultra-Saharan Africa. That's because we share a Paleolithic origin and not just a Neolithic one.

      Delete
  4. "If you mean the ones from Burgundy, surely not: their mtDNA is too modern"

    Gok2 for example was at most 30% WHG(I've seen estimated at Haak 2015 and Eurogenes). So, even Middle Neolithic folk were 70%+ Anatolian. The WHG already existing in their Anatolian ancestors isn't from "Europe"(everything outside of Greece).

    Also, I think lots of modern and Neolithic WHG is expressed via the paternal line more than maternal line. Founder effects R1b-L151(WHG?, EHG?), R1a-Z283(EHG), and I1-M253(WHG) have erased the former high frequency of I2a. The maternal lineages and autosomal signature of I2a-rich people is still strong today though.

    We have prove H1 existed in Early Neolithic farmers who were about 100% Anatolian. It could have existed in hunter gatherers in Europe but there's no prove. I believe there's very likely unsampled hunter gatherers with H1. Although I still think the defining-popularity and expansion of H1 is a Neolithic thing.

    ReplyDelete
    Replies
    1. "Gok2 for example was at most 30% WHG"

      Compared with what? EEF quite apparently. EEF already had some WHG/UHG (25% maybe per the latest study).

      "So, even Middle Neolithic folk were 70%+ Anatolian".

      EEF, not "Anatolian". I won't accept calling that baseline "Anatolian" because just taking an still unpublished sample from near Izmir, which probably had back-influences from Greece and whose exact similitude with EEFs is not yet properly quantified in public, as reference is just not acceptable. More so when all that region of West Anatolia has not produced any Epipaleolithic or Mesolithic of any kind. Call it Aegean and we'll be in agreement. But anyhow that's the same as saying EEF, i.e. Stuttgart, etc.

      "Also, I think lots of modern and Neolithic WHG is expressed via the paternal line more than maternal line".

      Expressed? We are talking raw genetics here not epigenetics or phenotype, what is what "gene expression" is about. You may mean "carried by". That remains to be demonstrated and anyhow most unlikely. That is in any case what unifies the "kurgan gang": fringe claims like: (1) all R/R1 comes from the steppe, (2) Basque 90% R1b is in fact IE, (3) mtDNA doesn't matter. All that is nonsense, not just to me but to anyone who considers the matter without a dogmatic bias.

      "The maternal lineages and autosomal signature of I2a-rich people is still strong today though".

      Really? U5/U4 have collapsed from 100% in the regions where I2(a) was found to be dominant in the Epipaleolithic to some 10-20%. What are you talking about? Most of that collapse however was surely caused by EEFs, whose lineages are not the ones you are discussing.

      "We have prove H1 existed in Early Neolithic farmers"...

      And also U5. What means what? Borrowing from the Paleoeuropean stock, nothing else.

      Delete
  5. "How can you claim that? Naqab's Bedouin B are the best approximation to EEF-all WHG."

    Bedouin have something closely related to WHG. Maybe it isn't exactly the same, maybe it split from Loschbour/La Brana 20,000+ years ago, but it's still closely related.

    ReplyDelete
    Replies
    1. Closely related relative to what? Mbuti? Onge? Sure, all West Eurasians are "closely related" since 50 Ka ago.

      Delete
  6. "Apparently they did test locus 73, although all the data is only described in fig. 6 of the conference paper. But in that figure locus 73 is clearly placed to separate N(xR0) from R0, so everything to the right side should be what they say it is: H and V. "

    All my data is coming from here. Look at Table S4.

    https://docs.google.com/spreadsheets/d/1jF2dE4AhdZLIROXgbak2L_csN9vUYmsK63_bYa324PQ/edit#gid=1403851040

    It's from the Supp. info of Brandt. 2013.

    ReplyDelete
    Replies
    1. I don't see the problem, Brandt says it's G37A, what clearly indicates it is within R0.

      Delete
    2. My bad. It's actually Hervella et al. 2012. There are many R-G73Gs. I'm going to update my mtDNA data. I'm especially going to look at pre Neolithic HV/Hs. There are two convincing Upper Palaeolithic and one Mesolithic HV(not tested for H) from Spain in recent study. There are a lot of H1-subclades that vary by region and any of them could be pre-Neolithic, considering how old H1 probably is.

      Delete
    3. Hervella did not test for locus 73, instead she used the old school but reliable method of testing RFLPs (Alu and such). Again denying the validity of those results is only caused by Jean Manco's personal hostility to the very idea of presence of H in Paleolithic Europe altogether. She had just published a book with such claims when... reality check knocked even more strongly at her door. Instead of correcting, she chose the path of dishonesty. Sad but true.

      Delete
    4. In the link I gave you it gives Hervella 2010, Hervella et al. 2012 as their source and calls for 73. I'll check their original sources.

      Delete
    5. The findings of Hervella 2012 were when I fell off with Jean Manco (or rather she rejected to keep discussing with me, when we had previously kept a productive correspondence). I may still keep some of the emails. On her insistence that allegedly Hervella et al. had only considered HVS-I, I initially was led to think the same, so I wrote to the authors to clarify the doubts and got clear answers, which were posted HERE as "edited section". I believe it is of your interest as self-proclaimed mtDNA librarian to know all the details.

      Delete
  7. Here's the post-Mesolithic data from Hervella 2012. I wish they mentioned in the paper they tested RFLPs. The results are interesting. I'm going to look back at other papers I've ignored for a while.

    https://docs.google.com/spreadsheets/d/1rkbllOZG8Eif9hGmG97H_DCOihY00RL9kWNT78J0Ru4/edit#gid=2042621177

    Just as the authors didn't mention they tested RFLP markers in the paper I think they forgot to mention other key information. Because....
    >Only a handful of U HV1 haplotypes fit in U-clades 90%+ of West Eurasian U fits in. Maybe the authors didn't mention most Us got little coverage in HVR1.


    >Three Hs are an HVR1+2 match with a U. They have rCRS in HVR1 and extra mutations 146C! 285T C312T C313T in HVR2. Maybe the U didn't get a call in the H SNP they tested, and the authors wrongly assumed the sample was U.

    Something that defends the authenticity of the results is.
    An independent study(Lacan 2011) found a Neolithic U5b which shared several rare HVR1 mutations with a Mesolithic U5b from Hervella 2012. Contamination can't explain that.

    ReplyDelete
    Replies
    1. They do mention it in the paper but the language they used is sorta technocratic, mentioning them almost by passing, using a different term and directing the reader to the reference footnotes like: "do you want to know what are we talking about?, read the bibliography". I agree that is not the best way to report science but that's what they did and that's how Latin university works way too often (snooty, distant, hierarchical and bureaucratic) - as counterpoint, Anglo university tends often to fall into sensationalist journo style, what is not ideal either.

      Delete
    2. Fu 2013 has a H1 from the Upper Palaeolithic in the Southern tip of Italy!! No date for the sample was given.They say contamination is more likely for this sample than others. If the result is legit I still think H1 is mostly expanded out of the Neolithic East Mediterranean. Wasn't Italy and the Balkans connected by land during the Ice age?

      http://www.sciencedirect.com/science/article/pii/S0960982213002157

      The two other Palaeo Italians had U(U5b2b1, U2'3'4'7'8'9*). Another paper found haplogroup HV1 in Mesolithic Sicily, based on HVR1 coverage. HV1 is mostly in SouthWest Asia today.

      Delete
    3. "Fu 2013 has a H1 from the Upper Palaeolithic in the Southern tip of Italy!!"

      Thanks for reminding me. It supports my idea of (early?) UP expansion of H and H1 in Europe.

      "Wasn't Italy and the Balkans connected by land during the Ice age?"

      Only in the Northern Adriatic (which was then probably a marshy plain), the Southern Adriatic and the Strait of Otranto were still sea, even if narrower than today.

      While Uluzzian (now believed to be made by H. sapiens) was wiped by the Campanian Ignimbrite mega-eruption, all Europe experienced the sudden expansion of "true" Aurignacian, which also affected Italy and even probably Sicily (needed to cross a strait as well) but probably expanded from Central Europe (oldest Aurignacian I is Istallosko in Hungary).

      Later Gravettian (quite apparently intrusive, maybe from West Asian precursors) expanded on top of Aurignacian, again beginning by Central Europe (Moravia notably). Then came the LGM which shifts the centrality to SW Europe, particularly Dordogne (= Perigord), which shows habitation densities 4x other densely populated districts. From this "paleo-metropolis" of SW France expanded first the Solutrean culture and later the Magdalenian one. Neither of these shows a penetration in Italy or Eastern Europe, which remain epi-Gravettian but rather are primarily limited to the Franco-Cantabrian and Iberian provinces, although Magdalenian also expanded since c. 15 Ka BP into Central Europe.

      An Epipaleolithic culture of Poland has been claimed to be Solutrean-related however and some late "Solutrean" presence has also been claimed for Hungary. The Eastern Adriatic shows rock art comparable to that of Western Europe but remains AFAIK poorly defined in the techno-cultural aspect, some of this rock-art influences could have reached as far as Southern Anatolia (Beldibian) but there is another parallel rock-art expansion wave via North Africa, reaching Egypt/Nubia and very similar to Beldibian typology, so unclear.

      Also there was a late (Epipaleolithic) epi-Gravettian expansion from Eastern Europe into at least Romania (and maybe further into the Balcans, in Greece some sites have been labeled "epi-Gravettian"). Additionally a distinctive culture arose in Nordic Europe, Hamburgian-Ahrensburgian, crossing the North Sea to Norway, Eastern England and Scotland. These people initially also lived on Doggerland for sure. The Magdalenian affinity of this group is controversial but can't be discarded. They persisted in Denmark/Scania until the Neolithic (Maglemöse culture).

      Hope this helps but I reckon that it's all a bit blurry, notably re. the late groups, on which only shallow research is available way too often.

      Delete