November 15, 2015

Minor issue with comments

For a week or so the comments feedback, on which it depends that latest comments that appear listed on the right margin of this blog, had been glitched. 

When I checked around yesterday I could not get the gadget to go back to work, however, looking around, I found three comments awaiting approval, something that should not be the case because there is not comment pre-moderation implemented right now. After I approved them all, the gadget seems to be working back again, so I'm guessing that those ill-queued comments were causing a glitch in the feed. 

My apologies for whatever inconvenience.


  1. yeah, I also thought something weird was happening, anyway, will you not comment or post something on the CHG? are you preparing something or not find it interesting ? :).

    1. I will write something "soon [TM]", however I just read the article a few hours ago for the first time and I'm a bit overwhelmed by other issues, like the consequences of the Paris attacks. Sadly enough, I can't dedicate myself to this 40 hours/week, as it'd be the case if it was a paid job.

  2. Can we use this as an open thread? :3

    1. Feel free. I normally don't have issues with off topics unless they go way off hand, so if you think you have something to say that is interesting and don't find the space, put it anywhere.

    2. Thanks. :)

      I don't know if this qualifies as interesting or not, but I and my family recently did one of's autosomal DNA tests. The results were mostly boring - a mix of Germanic and Celtic, but I did get one odd thing - under the Mixed Mode Population Sharing, I consistently come back as ~5-6% of some sort of Caucasian population. My best hits on Davidski's K13 model are 93.8% North_Dutch + 6.2% Tabassaran, 94.1% North_Dutch + 5.9% Lezgin, 94.2% North_Dutch + 5.8% Chechen. It follows that general pattern with either North Dutch or Danish at ~95% and some Caucasian population at ~5% until the ~12th best hit, at which point South Asian (Balochi and Brahui) and Central Asian (Turkmen and Tadjik) start showing up as the ~5% part.

      This seems to be coming from my maternal grandfather's side. He's 3/4 German and 1/4 Czech and born in Bohemia. 88.8% Austrian + 11.2%Tabassaran is his best match, below that it generally follows a pattern of 86-92% Austrian or East German + 8-14% some population from the Caucasus.

      So I was wondering if people here think that signal is real, and if so any ideas on how the heck someone in Bohemia is ~1/8 Chechen. The family tree goes back 4 generations from my grandfather, and all the names seem German or Slavic. Can anyone think of a plausible explanation?

    3. The Huns and the Alans? Oh, well... I really have no idea. People do migrate: they have been doing it all the time, so if your grandpa is 11% Caucasian genetic-wise, then he may have a great-great-grandparent who was 100% Caucasian or, more probably, a great-grandparent that was more mixed... or whatever. That brings us, possibly, to the early 19th century in any case, maybe a Russian soldier keeping the Restoration's "new old order" against the French-led revolutionaries?

      I really don't know enough but I'm assuming that that component does not normally appear in Western and Central Europe, does it? If it does, among which populations?

    4. The Alans (Ossetians) are actually a pretty good fit interestingly enough lol. And yah, mid or early 19th century or so.

      My grandfather shows in Eurogenes K13 as 34.76% North Atlantic, 29.09% Baltic, 11.55% West Mediterranean, 11.50% West Asian, 8.56% East Mediterranean, 1.51% Red Sea, 1.94% Amerindian, 0.73% Oceanian and 0.37% Sub Saharan African.

      Your typical Austrian is 37.21% North Altantic, 29.87% Baltic, 14.58% West Mediterranean, 7.23% West Asian, 8.23% East Mediterranean, 1.56% Red Sea, 0.78% South Asian, 0.07% East Asian, 0.16% Siberian, 0.12% Amerindian, 0.11% Oceanian, 0.07% East African, 0.003% Sub Saharan.

      So the differences for my grandfather seem to be -2.5% North Atlantic, -0.8% Baltic, +5.27% West Asian, +0.43% East Mediterranean, -0.78% South Asian, +0.62% Oceanian, +1.81% Amerindian, and a few other minor ones.

      So I'm looking for a population with a lot of West Asian ancestry, maybe a little bit of Amerindian ancestry, maybe moderate levels of Baltic ancestry, and very little of anything else.

      Here's Davidski's spreadsheet of each population in his K13 calculator:

      At 11.5%, that's more West Asian ancestry than any group in Western or Central Europe. There are some Southern European groups that have decently high West Asian levels - namely some Italians, some Greeks and Bulgarians, but they all have much higher levels of East Mediterranean ancestry, so I don't think they're likely candidates. My grandfather would have to be almost 50% Central Greek to have enough enough West Asian ancestry, and that would his East Med ancestry in the 18-20% range instead of 8.6%.

      So I think this must be something from the East. The time frame does fit with the Russian conquest of the Caucauses, or possibly their conquest of the Crimean Khanate. And a lot of refugees from the Caucauses went in that western direction initially only to be expelled again by the Russians. Apparently some Nogais settled alongside German Mennonites north of the Sea of Azov before getting expelled around the 1860s, and the West coast of the Black Sea was in Ottoman hands until the late 1870s, and had a Nogai presence. So maybe a mixed refugee from one of those areas? A mixed German-Nogai, German-Circassian or German-Tabassaran or some other refugee fleeing the Russians? Still seems a bit strange and convoluted.

      The Amerindian component is small but somewhat strange. The only Eurasian population with higher Amerindian ancestry that doesn't also have higher Siberian ancestry is the Tabassarans, and they're only 0.01% higher Amerindian. Probably just within-group variation?

    5. "Amerindian" should mean extra ANE, and hence Siberian of the kind of Kets... or maybe their Ugric neighbors. Could it be something Hungarian that has been randomly fixated in your family? The Magyar invasion seems way too old to leave such a big chunk but maybe there's some not studied population that retained some of that. You know that isolate populations in mountain areas often keep strange ancestry fractions. The population may be even extinct by now, especially considering the many demographic changes that Central Europe suffered after WWII, but maybe is still alive, hiding somewhere beyond the interest of population geneticists.

    6. Re: Amerindian, there's a separate Siberian component too, and while the two components do go together, in virtually every Eurasian population the Siberian component is larger than the Amerindian component by a lot. My grandfather comes back as 0% Siberian and 1.94% Amerindian. Kets come back as 57.56% Siberian and 7.7% Amerindian for example. Even West Greenlanders come back as 38.7% Siberian vs 37.1% Amerindian. The only samples where Amerindian was over 1% and higher than Siberian were Karitiana, Pima, Maya, North Amerindian, MA-1, Motala, Tabassaran, Brahmin_UP, Kalash, Punjabi_Jat, Icelandic and Kol. Only Tabassran and above were higher than 1.94%.

      So yah, I agree that it's ANE, but the question is which population is ANE but Amerind-ANE and not Siberian-ANE. Barring a time machine a MA-1 stepping in to a time machine... yah. Those new CHG samples may be a better proxy.

      ANE isn't that high for him though. 17% - pretty normal for that part of Europe. The bigger excess is with WHG-UGH - 56.7%, which is higher than any of David's European samples for K7. I'm actually higher than that - 61%. Not sure what's going on there, but he warned that that K7 was pretty noisy.

      I think the Hungarian suggestion may be a good one. The Jasz people in Hungary are Alans that migrated there from Ossetia in the 13th century along with many Cumans.

      They're in the middle of the Hungarian plains, but who knows. My mother's been working on our family tree, and while nothing jumps out in the first 3 generations, maybe something will show up as she goes deeper into their records.

    7. It could also be just noise: ADMIXTURE (or actually any other autosomal analysis) is not exact science and a 2% can always be just noise. What you are asking to the always limited inference power of the algorithm is: how can you classify this genome according to pre-determined "zombie" populations (which even in the best case are not perfect references for ancestral populations). Try running it freely instead of using supervised mode (that is what a "calculator" does) and see what it produces.

      You can do that at home for free, although I think you need a Linux OS (Ubuntu or Mint are good Windows-like distributions "for dummies", that you can always install as secondary OS easily). Razib had a very good entry years ago explaining how use ADMIXTURE:

    8. I think you're right about it being noise, and that the Amerind zombie population is just splitting off a piece Caucasus heritage and labeling it "Amerind." Though it's interesting that it does that.

      Yah, been meaning to do some like that re: ADMIXTURE. David has a post on how to model 4 way admixture that I was going to give a shot too. Probably will have to wait until after this project from hell is complete.

    9. So, my mother went back through her family tree work and found an error. My grandfather's paternal grandfather was illegitimate. Born 1866 in southeast of Děčín in what's now the Czech Republic. We were planning to try 23 and Me at Christmas, and since this is my grandfather's direct paternal ancestor, maybe the Y-DNA results will shed some light.

  3. This comment has been removed by the author.

  4. Maju,
    Thx, so since you have brought up the issue which is making people think worldwide, my questions for you are, how we can stop this advent of extremism in Europe and how this incident will affect the innocent Muslims there?.
    I am eager to know your views :) .

    1. I think not marginalizing those groups to begin with would be a good start. Daesh is counting on Europe's xenophobia to help them recruit in Europe.

    2. I have another political blog: by allowing "off-topic" stuff, I meant within the scope of this blog. You can comment there but expect pre-pub moderation because, for some reason, I almost only get right wing extremists to comment, and lots of spam too. Something that does not happen here.

      I took the decision in 2010 to split my old generic blog in two and, in retrospective, I think it was an excellent idea. So please go an comment there. I recently made a handful of entries on the matter, that are the ideal place for this discussion.

      Thank you in advance.

  5. A discovery of Mesolithic art in Spain depicting a hunter-gatherer campsite might interest you.

    1. Well, curious, thank you.

      The paper is this one: (haven't read it yet).

  6. This comment has been removed by the author.

  7. Maju,
    I'm really interested to know, about your opinion on the latest Irish aDNA, which also show steppe signals around ~2000 BC, which Rob linked.
    do you think this goes along with this?.

    Metallurgy arrived in Ireland with new people, generally known as the Bell Beaker People, from their characteristic pottery, in the shape of an inverted bell.[9] This was quite different from the finely made, round-bottomed pottery of the Neolithic. It is found, for example, at Ross Island, and associated with copper-mining there. It is thought by some scholars to be associated with the first appearance of Indo-Europeans in Europe (possibly Proto-Celtic),[10] though this theory is not universally accepted.

    The Bronze Age began once copper was alloyed with tin to produce true Bronze artefacts, and this took place around 2000 BC, when some Ballybeg flat axes and associated metalwork were produced. The period preceding this, in which Lough Ravel and most Ballybeg axes were produced, and which is known as the Copper Age or Chalcolithic, commenced about 2500 BC.
    Greetings for upcoming 2016!.

    1. Reading the paper right now and trying to figure out what part of the R1b tree does DF-21 hang from, from what I see in ISOGG, it may not be part of "Southern" S116 but from "Northern" U106 (right?) and therefore the Rathlin island individuals are not ancestral to most modern Irish (who are 90% S116) via Y-DNA but would rather appear as "intruders" from some other area.

      More interested on the Ballynahatty sample therefore personally, although so far I haven't found which Y-DNA lineage they belonged to.

      There's a lot of noisy data about hair color and such which I believe quite irrelevant and I'm trying to understand the substance of the matter. Neither Dienekes nor Davidski dwell on the details, so I have to read the paper in full first.

      Thank you for the heads up. I reckon I'm not paying attention these last weeks, rather months. I haven't even checked my RSS feed in a long time, so I'm getting stuck.

      Another interesting study that I got a heads up for is: (about Basque R1b-DF27, which is shared with Irish and others - but very especially with the Irish).

    2. I correct: only one of the Rathlin samples is claimed (but very unclear) to be DF-21, in general the only thing clear is that they were all R1b-L21 (M529), what is fully in the Southern clade of Western European R1b and hence within modern Irish expectations. I still can't find anything re. the Y-DNA of the Neolithic Irish individual, which would be a tie-breaker.

    3. Forget the previous (was thinking loud mostly, also hangover): the Neolithic sample is a woman with mtDNA HVO, while the BA samples are almost certainly relatives, with mtDNA U5 and J1 and Y-DNA R1b-M529. This lineage is the "Irish" subclade of S116 (SW European R1b), which is not just Irish but also found at high frequencies in Great Britain and West France (also less frequently in other parts of France and some Iberian areas).

      So re. Y-DNA Raithlin demonstrates that "modern" and "Irish" clade R1b-529 existed (in at least some insular parts of Ireland) already in the late Chalcolithic (c. 2000 BCE, Bell Beaker era). Hence it totally dismantles they hypothesis of R1b expanding in that period: it must be significantly older, and, judging on the available data, NOT steppe-derived but rather Western.

      Another issue is autosomal DNA, which does suggest quite strongly a shift towards Corded Ware type configuration in this early date. We can't confirm if this affected all Ireland (and Britain?) or just this particular island or even a particular family of this particular island, but it's also apparent that the change they embody all modern Ireland and the Western and Northern parts of Britain, so they are not just a fluke.

      A possible explanation would be that the Western Bell Beaker province suckled a lot of ancestry from Indoeuropeanized parts of Europe (Germany and surroundings). However the Y-DNA points to the pre-IE south instead.

      Intriguing. We still need more data.

    4. The three individuals from the Bronze Age
      Irish site are R1b1a2a1a2c-L21, which is typical of British isles and absent in Eastern Europe, not to say of the steppes, but quite strong in Basque country:

      I think this is related with Bell Beakers, they have some green/teal component as other Bell Beakers but they do not come from the steppe, rather from Iberia...
      Maybe it is not a case that Ireland was called Hibernia. I also doubt that these people
      were IE speakers, rather Vasconic. See here some connections of Basque and Irish:

    5. L21/M529 is not "strong" in the Basque Country (5-13% in one study, just 2-3% in another), although it does exist. Most common are "brother" clades (DF27 particularly, which also exists in Ireland). In general all them are offshots of R1b-S116, which seems to stem from Southern France.

      "I think this is related with Bell Beakers"...

      May be. Notice that the teal component is to this day absent (or tiny) among Basques anyhow. Until now we were only sure that it was present in Central European post-CW peoples (BB or not), and the big novelty is that now it also appears in Ireland.

      I can't say anything about language at this point. I know that Germanic, Celtic and Italic have Vasconic substrate elements but all these families seem to stem from Central Europe at a latter date.

    6. David's K10 run has Basques around 15-20% CHG, which would be on par with these Bronze Age Irish samples.

    7. Hi Maju,
      About L21, I have found no frequencies in the links you have given. I noticed in the Eupedia map a concentration in the Basque country, at least 15-20%. I think that the intensification in that region is significant. Here there is a slightly different map, with some presence also in NW Italy, a region of Celtic presence but also Bell Beaker!:

      I have seen in fig.3 in the paper a low affinity with Basques, but apparently this involves also mtDNA, isn't it? Also the affinities with the steppe are very low, by the way. Chalcolithic in Ireland arrived with Bell Beakers, originally from the continent, so it is not excluded that they brought some steppe DNA arrived in the third mill. BC, but I don't understand from the paper where is the proof of EHG and Yamnaya, because the rose component is also WHG and the green component is originally Caucasian or near, as admitted by the paper itself??.
      Also to see that the closest triangles in the PC map B are of Bell Beaker and Unetice, a culture heir of Bell Beaker according to some scholars. It is curious that Sintashta/Andronovo are closer than CW and

      About Celts in Ireland, the main theory is that they arrived in the Iron age in small numbers:
      It seems there was also a recent memory of Gaelic invasion, and curiously they have a Scythian origin:
      Genetically, it is rather R1b-U152 which is connected with Celts (and apparently also Italics),
      with 5-10% in certain parts of Ireland. The fact that it is not found in the Bronze Age site can be interesting.

    8. "I have found no frequencies in the links you have given".

      Try again, seriously. Notice that L21=M529, but that only applies to the second link, in the first one there is a Dienekes' spreadsheet with frequencies, right near the top, which clearly displays the frequencies in each Basque & Gascon region.

      I think the Eupedia map is based on the first dataset, it would not work with the latter. The two dots with 15-20% should be very specific towns whose source I don't know about.

      "I have seen in fig.3 in the paper a low affinity with Basques, but apparently this involves also mtDNA"...

      I think it's autosomal only, right? I will tell for sure when I have time to deal with the piled-up work but that was my impression when I read the paper first yesterday.

      R1b-U152 is most frequent in Switzerland and NW Italy (Liguria, Piamonte and Tuscany particularly, so not looking Celtic at all). It may have been scattered to some remote areas of the Balcans with Celts but otherwise it seems that like its "brothers" R1b-M529 (West France, Britain and Ireland) and R1b-DF27 (largely associated with Iberia but also notable in Ireland) it should originate from France, rather towards the South.

      A comment in one of Davidski's endless debates, provided a map showing that the "ancestral" paragroup R1b-S116* was most common in and around the Massif Central, what, if confirmed, would be a most important evidence for the origin of this R1b sub-haplogroup, the largest one.

      Basically it seems that R1b-S116 originated in Southern France and that it had three sub-expansions:

      · R1b-M529 (L21 for FTDNA fans): expanding from Brittany to Ireland and large swathes of Britain (looking Neolithic but maybe needs reconsideration after these findings).

      · R1b-U152 (S28 in Eupedia's map series): expanding from SE France probably, and looking coincident with Chassey-La Lagozza late Neolithic or early Chalcolithic cultural group, also with later Ligurians. Notice how important it is in Corsica also, which cannot be "Celtic" either.

      · R1b-DF27, expanding from somewhere near the Pyrenees and much more common in all Iberia than Eupedia admits to (see second link above). The process of expansion of this clade is as of yet unexplained in terms archaeological (could be First Neolithic but early Neolithic samples lack it so far, so maybe it has to do with BB and/or the Megalithism-related mixed-HG Atlantic farmer reflux).

      Finally these three major sub-haplogroups have a Northern "uncle" R1b-U106, which should have expanded from somewhere near the Netherlands, although more research is needed.

    9. Thank you for the detailed information, Luis. I have a remark about U152 and NW Italy and Corsica.
      Ligures/Ligurians of Liguria and Piemonte and Corsica are considered Celtic or Italo-Celtic speakers, and Corsica was inhabited by Ligures and
      later Romans:
      The present dialects of Liguria and Piemonte are Gallo-Italic, and Corsican is a sort of Tuscan.
      U152 is quite strong also in Central Italy and Sicily, so it seems both Celtic and Italic.

      DF27 is clearly Iberian and Basque, but I suspect it is more recent than the others since it did not spread much?.
      U106 is mainly Germanic, also in Italy it is associated with Langobards/Lombards.

    10. I know the issue of Ligurians and there's absolutely no reason to imagine that they were Celts (although they were Celt-influenced after they were cut off from their Vasconic relatives in the 1300-550 BCE period, i.e. between Urnfield culture expansion and Marseilles-influenced Iberian reconquest).

      Corsica has no apparent relation with Ligures other that it might be tracked to the (late) Megalithic period (i.e. totally pre-Celtic). Most Romans obviously never stepped in Corsica at all (why would they unless they were bureaucrats, traders or soldiers?)

      Imagining everything as "Celtic" is just an Indoeuropeanist fashion that has no basis but is, or rather used to be, very popular among historians. Pre-IEs (excepted maybe some more civilized ones like Iberians) had no "prestige", so they were dissed out without any objetive reason and everything was forced into "Celticness", almost always without any single piece of evidence whatsoever. Today we know better... or we should.

      DF27 did not spread much? Wow! All Iberia is "not much" to you? It's the second largest peninsula of Europe after the Balcans, with 5% of the area of the subcontinent (of which 40% belongs to Russia) and by population (some 55 million: 10% of European Union's) is as important as the Balcans, France, Italy or Britain. I know that geopolitically doesn't weight so much but it still represents a very sizable fraction of Europe and approximately 1/3 of Western Europe (which can be divided in Iberia, Greater Gaul between the Rhine and Pyrenees, and Britain-Ireland, roughly of similar relevance each).

      So to my eyes the three sublineages have roughly similar patterns of distribution, neither is more significantly more important in terms of numbers or area.

      U106 has indeed a Germanic-like pattern but this is difficult to extrapolate that simplicity to Britain, where it's just too dense to be merely Germanic. Compare with the distribution of R1a for example. So IMO it's a North Sea clade that may have indeed re-expanded with Germanics in southward direction.

    11. "Corsican is a sort of Tuscan".

      You know that Tuscan = modern standard Italian, right? IF Corsican derives from Tuscan it must have been a very primitive form of Tuscan, with conservative -u endings, now (and in the Renaissance already) totally lost. I'm a bit skeptic, although Tuscan late Medieval influence seems undeniable.

      Anyhow, the concept of Gallo-Romance has seemingly fallen out of fashion. It was used to group Italian and French groups but these two branches are very different from each other. Nowadays Italian dialects, also Corsican, are grouped in an Italo-Dalmatian category, which makes better sense (any Spanish or Italian would agree that their respective languages are more similar to each other than French is). IMO, to be overly simplistic, -u ending Romances (Romanian, Corsican) probably derive from an early phase of Vulgar Latin, -o ending ones (Italian, Spanish) from a later phase and -e ending ones (French notably) from the Dark Ages (very marked Germanic influence). Call me simple-minded but there's probably some truth to this, there's also some evidence for an -o ending Vulgar Latin in the 3rd century or so.

      Interestingly, reading on Corsican language:

      "Seneca the younger, reports that both coast and interior were occupied by natives whose language he did not understand"...

      He would have identified Celtic, no doubt.

      "There was probably a sub language that is still visible in the toponymy or in some words, for instance Gallurese zerru 'pig'".

      Ha! Compare Basque 'txerri' (← *zerri) and Spanish 'cerdo', meaning the same and obviously from the same Vasconic root, totally unrelated to Lat. 'porcus' and very possibly *zerri (the ending in -u, -o is clear Romance influence). So, Ligures or (most likely) not, Corsicans were also Vasconic speakers almost certainly. No surprises here.

    12. Also interesting in the Italo-Dalmatian entry: "Based on mutual intelligibility, Dalby lists four languages: Corsican, Italian (Tuscan–Central), Neapolitan–Sicilian, and Dalmatian".

      This, if correct, would make Corsican a distinct branch of the Italo-Dalmatian Romance sub-family, regardless of Tuscan influence. Basically it means that it's difficult for Tuscans (or Italians) and Corsicans to understand each other, much as I (Spanish speaker with only some very shallow knowledge of Italian) can understand an Italian (Tuscan) speaker but definitely not a Sicilian speaker (I can tell you from experience, man, that language is terribly obscure, harder than French!)

    13. Today Gallo-Romance seems to include only French and Franco-Provençal (spoken around Savoy but moribund in most areas). Dalby again proposes the following Romance sub-classification:

      · Ibero-Romance: Portuguese and Galician, Mirandese and Asturian-Leonese, Spanish, Aragonese;
      · Occitano-Romance: Catalan, Occitan;
      · Gallo-Romance: Langues d'oïl (including French), Franco-Provençal;
      · Rhaeto-Romance: Romansh, Ladin, Friulian;
      · Gallo-Italic languages;
      · Venetian;
      · Italo-Romance: Corsican, Italian, Neapolitan-Sicilian;
      · Sardinian;
      · Dalmatian (extinct);
      · Romanian: Daco-Romanian, Istro-Romanian, Aromanian, and Megleno-Romanian.

      Makes some sense but I'd seriously consider the -u, -o and -e/-ø endings as a key layering issue. Languages with the -u ending (more similar to Lat. -us) are only found in remote isolated areas: Asturias, Corsica, Sardinia and the Balcans, they probably reflect one way or another a primitive Vulgar Latin root, still attached to Latin proper. Those with the -o ending are the most common ones: most Iberian and Italian lects, as well as Venetian and Dalmatian, the -o ending is attested in c. 3rd century scattered Vulgar Latin inscriptions, so it probably represents a second stage of this proto-Romance evolution. The loss of the ending sound (-ø, as in Occitan-Catalan, or unspoken -e, as in French) is probably Germanic influence. French however is clearly much more influenced by Germanic (Middle German especially) with sounds like /ü/ or gutturalization of "r", sounding almost as /g/, which make it probably the most distinctive of all Romances. I don't think that Celts played much of a role in the evolution of Romances, on the other hand Vasconics did have an influence in vocabulary and in some cases pronunciation (/v/ → /b/, loss of initial "f", persistence of the 5-vowel system... all these evident in Castilian/Spanish).

    14. Re: the Ligurians - obviously we don't have a lot to go on linguistically, but simple geography dictates that they would have at least been under very heavy Celtic influence. If the defining feature of Celtic is Indoeuropean with a strong Vasconic substrate though, it would be tough to separate the two.

    15. "Basically it seems that R1b-S116 originated in Southern France and that it had three sub-expansions"

      There's nothing to support an origin in West Europe. You need to let go of ethnocentric theories, and I mean no offense. Essentially everyone who cares about this, from academics to amateurs, thinks it came from the East.

      I guess you're probably afraid this means Basque have Steppe ancestry and R1b, which has been seen as the original marker of West Europe, isn't from West Europe. Maybe you're also afraid Basque language was brought from the East.

      The whole idea of Basque being a isolate happened only after linguistics discovered they spoke a language unlike any in the world. Then early genetic work, before ancient DNA, made theories they're Paleolithic isolates. The concept is purely from modern academics.

      Before that Basque were just one of many ethnic groups and no one knew anything about ancient history. There's no force stopping them from being mixed. I know Basque are isolates in historical times, but still no one thought "They speak a non-Indo European language" in AD 1600.

      No one in 2300 BC knew anything about the past, if they wanted to have sex with someone from another tribe nothing would stop them. Latin America is an extreme example, and most did know about the past(That Spanish were invaders, etc.). New people came into Western Europe from the East, and they mixed with the natives. After several hundred years no one knew who was native or who wasn't.

      Ancient DNA has essentially proven R1b-P312 came with the new people from the East. IMO, it expanded mostly before 2000 BC. We already have U152 near the Alps and L21 in Ireland dating to circa 2000 BC.

    16. "There's nothing to support an origin in West Europe. You need to let go of ethnocentric theories, and I mean no offense."

      S116* peaks in frequency in Iberia for one. And sorry, when I said P312 before, I mean L-11.

    17. "S116* peaks in frequency in Iberia"...

      In the Basque Country and Ireland as far as we know to this day, Ryan. It might be just one shared Basque-Irish haplogroup or two (one Basque, another Irish) or more but, in any case, it reinforces the general pattern described by the three major subhaplogroups of S116/P312, which points to Southern France as shared origin, or rather it pulls it even further to the West.

    18. In this comment, Krefter, you don't seem to understand my thought: Y-DNA genetics must be considered on its own merits. Autosomal DNA, mtDNA, LCT and Rh are other markers: complementary viewpoints for the overall population history but not the same thing. You guys just jump from one to the other at caprice and, well, let's be serious.

      "Basque have Steppe ancestry"...

      Well, actually, Basques lack the tell-tale IE "teal" or Caucasus component (or, when it shows up in some individuals, it's tiny), so I'd say that Basques lack the most clear signature of steppe ancestry in fact. Also this signature is probably clinal in Europe, with negative peak among Basques particularly (Sardinians even do have some, although in this case it should probably be attributed to separate Mediterranean flows or affinities, most notable in Sicily).

      So, the real situation in the autosomal genetic aspect is not at all as you imagine it. Basques may have some EHG (or "ANE") element but without the IE-specific Caucasus element, it must be telling us some other story, not the IE one via Central Europe: you can't mix one and the other happily.

      Finally, I'd like to underline the "pro-Western" weight of the mtDNA pool (first "modern" ones in Paternabidea and Gurgy) and the LCT allele pool (oldest TT one among Basques again). So all markers point West, except that EHG autosomal thingy, that may include some Caucasus component also in the case of Irish and such. It's that aspect the one that falls off the overall Westernizing tendency and may just imply, for example, absorption of IE or Uralic type blood in the context of Bell Beaker expansion (from the Southwest again), but in any case without any apparent major impact in the Y-DNA, mtDNA or LCT marker histories. Some extra mtDNA J and HV0 (some loss of U5 also) is the only thing I can detect in the Basque mtDNA history, so maybe that stuff should be attributed to matrilineal admixture? Can't say until we have more data.

    19. Hi again,
      Keeping the discussion alive, some words appear to be Celtic, like the name of the capital Genua. On the other hand, there are the -asco toponyms (and names) which recall Basque words. An Italian wiki says that the pre-IE thesis is still the most accepted, and that also
      genetics has confirmed the peculiarity of people of Liguria and South Piemonte and their kinship with Basque and Welsh... However, they were in the middle of Celtic people and so had strong influences from them,
      according to Villar there are 5 linguistic strata in Liguria: Latin, Gallic, Lepontic, ancient European (IE) and pre-IE.
      If we see the Eupedia map of U152, it shows actually that it is stronger in Tuscany and Lombardy than Liguria,
      so it is not especially associated with Ligures. About Corsica, I start with this quotation from Seneca, the great
      Spanish-Roman philosopher of the 1st century CE:

      "this very island [Corsica] has often changed its inhabitants. Not to mention more ancient events, which have become obscure from their antiquity, the Greeks who inhabit Marseilles at the present day, when they left Phocaea, first settled here, and it is doubtful what drove them hence, whether it was the rigour of the climate, the sight of the more powerful land of Italy, or the want of harbours on the coast: for the fact of their having placed themselves in the midst of what were then the most savage and uncouth tribes of Gaul proves that they were not driven hence by the ferocity of the natives. Subsequently the Ligurians came over into this same island, and also the Spaniards, which is proved by the resemblance of their customs: for they wear the same head-coverings and the same sort of shoes as the Cantabrians, and some of their words are the same: for by association with Greeks and Ligurians they have entirely lost their native speech. Hither since then have been brought two Roman colonies, one by Marius, the other by Sulla: so often has the population of this barren and thorny rock been changed. In fine, you will scarcely find any land which is still in the hands of its original inhabitants: all peoples have become confused and intermingled: one has come after another: one has wished for what another scorned: some have been driven out of the land which they took from another."

      So, Ligures apparently lived there, and there are also toponyms with -asco. Actually, Corsica was also under Genua in recent times, but earlier under Pisa, that colonized it strongly, so that the Corsican language is a form of Tuscan with
      some phonetical peculiarities like retroflex d, according to someone because of the Southern Italian origin of Roman colonists, a scholar suspects a Neolithic substrate.
      Anyway, it seems that U152 is connected there with the Roman/Tuscan people, according to a theorist of Italian genetics U152 is the Roman SNP. I think it is also Celtic, because of the concentration in NW Italy, France and Switzerland, in the core Hallstatt territory (
      Maybe also its presence in Turkey is due to Celtic Galatians...

      On DF27 and Iberia, I don't want to say that Iberia is insignificant or small, but the others similar SNP are not so confined, and if it was there in the Bell Beaker period, it is strange that it is absent in Italy where Bell Beakers arrived.
      So, I think it should be post-Bell-Beaker!.

    20. @Maju,
      "Well, actually, Basques lack the tell-tale IE "teal" or Caucasus component (or, when it shows up in some individuals, it's tiny)"

      Now that we have DNA from the actual Teal people(CHG) Basque do score in Teal/CHG.

      "So, the real situation in the autosomal genetic aspect is not at all as you imagine it. Basques may have some EHG (or "ANE") element but without the IE-specific Caucasus element, it must be telling us some other story, not the IE one via Central Europe: you can't mix one and the other happily. "

      Basque have Eastern ancestry that didn't exist in their region in 2800 BC, this much we know. This is why Sardinians usually come as closer to Neolithic and Copper age Basque than modern Basque do. It is hard to believe Basque's Eastern ancestry can be unrelated to the people from the East who were moving into West Europe after 2800 BC.

      "Finally, I'd like to underline the "pro-Western" weight of the mtDNA pool (first "modern" ones in Paternabidea and Gurgy) and the LCT allele pool (oldest TT one among Basques again)."

      LCT and mtDNA are shaky evidence of West>East gene flow. I say shaky because LCT can rise in frequency without a lot of gene flow and unless we have high coverage mtDNA we can't be very confident of anything.

      Those two lines of evidence aren't as strong as autosomal evidence for East>West gene flow. We aren't yet able to distinguish Middle Neolithic Hungarians from Spanish. We'd need 1,000s of genomes to do that. So, with autosomal DNA it is impossible to confirm West>East gene flow in the Neolithic. That's not evidence against West>East but it's not evidence for it either.

    21. @Krefter

      "Now that we have DNA from the actual Teal people(CHG) Basque do score in Teal/CHG."

      This kind of argumentation is pathetic and facile.

    22. This kind of argumentation is pathetic and facile.
      I think the real teal is yet to be found, Teal is perhaps a part of the CHG but CHG is not equal to Teal. So, i quite agree with you Marnie.
      Happy and Prosperous 2016.

    23. @Nirhar: I'll pass on Seneca, Herodote, etc. What they say surely refers to lesser coastal colonies, whatever the believed. Archaeology comes first, ancient posh opinions later.

    24. I'm interested in the teal or anything else that may identify Corded Ware people and such, Caucasus people and components are accessory here: tools for pattern identification.

      "Now that we have DNA from the actual Teal people(CHG) Basque do score in Teal/CHG".

      CHG are also high on other "ingredients", notably the so-called Basal Eurasia, and are therefore somewhat related to EEF, so it seems to me that you're confusing the marker of Indoeuropeans (the "teal" in Corded Ware) and the marker of Vasconics (the EEF affinity in CHG). There are no absolutes in autosomal genetics: it all depends on what you compare with. I'm all the time keeping CW and Yamna as IE reference, not some Caucasus hunter-gatherers whose relation with them is a best very oblique.

      This is obvious, K. You knew that before you wrote your misleading claim. Then why did you write that? Did you deceive yourself? You are not going to deceive me, you should know that by now.

      "This is why Sardinians usually come as closer to Neolithic and Copper age Basque than modern Basque do".

      I'm unaware of any ancient autosomal "Basque" sample of any kind. The Atapuerca and La Mina genomes (both from Castile) are clearly not good enough: they are LCT--, when some Basques of the period were LCT++, they are low mtDNA H when Basques were already at higher frequencies. Anyhow, K., we are not discussing Basques, are we?

      A sentence I hear a lot these days is "shut up and calculate", what was what apparently Bohr said to startup physicists. It applies here: less talking, more sampling. It's pointless to try to push theories when we are actually awaiting for the data. We should all shut up a bit, we can wait until hard data comes, as it will no doubt.

      "LCT and mtDNA are shaky evidence"...

      Less shaky than all you just said. They are important clues to consider and you do wrong disdaining them. "Shut up and calculate"...

      "Those two lines of evidence aren't as strong as autosomal evidence"...

      Autosomal data is the most slippery one to analyze in fact: each time a new paper appears the previous results change (a bit or a lot) and not because of new data but because different strategies to statistical analysis of massive datasets. So there's no reason to be too arrogant based only on autosomal DNA: it's the less clear of all references in fact.

    25. Nirjhar said: "I think the real teal is yet to be found"

      There's no such thing: those autosomal "components" are just statistical aggregates. Even for whatever they may have of "real" they are too slippery: they aggregate and disaggregate as time passes because people do fuck and bear children, at least some do, and that breaches the useless (actually harmful) "racial purity" of clonation. The cummulative effect of sexual reproduction through centuries and millennia is of course brutal, so no wonder that the farther back in time we go, the blurrier it all gets in terms autosomal.

      This does not happen with haploid genetics. You can dive back to even beyond the edges of the genus Homo with them. So I request a bit more respect for haploid genetics and a bit less of blind faith on the capacity of statistically inferred autosomal "components" to answer all questions. Thank you.

    26. @Maju

      "There are no absolutes in autosomal genetics: it all depends on what you compare with. I'm all the time keeping CW and Yamna as IE reference, not some Caucasus hunter-gatherers whose relation with them is a best very oblique."


      "There's no such thing: those autosomal "components" are just statistical aggregates. Even for whatever they may have of "real" they are too slippery: they aggregate and disaggregate as time passes because people do fuck and bear children"


      Luis, somehow these obvious realities have escaped many in the population genetics community, the reviewers at Nature and PNAS, and many supposedly highly experienced scientific journalists.

    27. Maybe but we are all humans and therefore susceptible to error. There are two things I believe I know about academic people (doctors and such):

      1. They have access to resources I don't, so I'm borrowing from their work (which we usually pay one way or another but anyhow, it's them who do the actual field work).

      2. To have a degree often does not mean much, just (in most cases) that one is mildly intelligent, disciplined and does not cause much trouble (respects the hierarchy, etc.) The last two aspects are usually more important than being bright in order to get your degree and further advance in the academic bureaucratic hierarchy. So in the end, much as in other facets of socio-economic life, the system tends to favor obedient mediocrity rather than rebellious brightness. Of course you can still be obedient and bright but that's statistically less common and anyhow, for as long as obedience, discipline and being part of a hierarchy is important, challenging the tenets of your field faces an extra layer of obstacles. That's why academy tends to scholasticism: because flattery will get you where confrontation usually won't.

      You may have a successful career in the security and monetary aspects but it's much less likely that you will make a breakthrough that way. And that's why often science advances against scholasticism, be it religious as in the times of Copernicus, or secular, as happens now.

      3. The degree of freedom allowed by not being part of the academic establishment is the right to bear the flaming sword of criticism and the telescope of Galileo to look beyond what is "allowed". Luckily they usually don't burn us at the stake these days, we have progressed so much.

  8. Hi I hope that you are well. I posted the link to my mass lexical comparison on your Basque/Nubian thread. It's on a Google Doc - if you can send me a request to view it, and let me know your thoughts when you've had a chance to skim through it. Thanks in advance.

    1. Hi, Chris. Sorry, I have it in my to-read or to-do list of things, which is getting endless. I swear I'll read it soon, today maybe?

  9. @Maju,

    R1b-P312 came from the East. There's no point in arguing anymore.

    The Late Neolithic Irish samples were autosomally identical to Eastern Bell Beaker. 100% of Eastern Bell beaker is R1b so far, all tested are P312 positive. No surprise these Irish share R1b, and specifically have the British/Irish version of P312.

    R1b-L21 could not have arrived in the British Isles before Steppe ancestry did. This is because the Late Neolithic Irish genomes can fit as 100% Eastern Bell beaker. They didn't have any Neolithic Irish ancestry. R1b-L21 came to the British Isles with people heavy in Steppe ancestry and they came from Central/East Europe(they weren't native to the Atlantic region).

    0% of Copper age Spanish are R1b-P312. One is possibly R1b1a2, but it is of low coverage and some of its positive calls leading to R1b are unreliable SNPs. Yet from around the same era, we see a population in Central Europe very high in R1b-P312 and high in Steppe ancestry(higher than almost all modern Europeans). Anyone could connect the dots and realize P312 is from the Steps.

    1. Steppe-like ancestry doesn't necessarily mean Yamnaya itself though, or IE.

      "R1b-L21 came to the British Isles with people heavy in Steppe ancestry and they came from Central/East Europe(they weren't native to the Atlantic region)."

      I haven't seen Maju deny this. His view IIRC has always been that P312 came from central Europe (please correct me if I'm wrong here lol). I don't see why any of us should expect R1b to have been bottled up in a single village in the Steppe. V88 must have already been in the Sahara by this point, so that suggests rather that R1b was widespread and diverse long before Indoeuropeans ever showed up.

      Bell Beakers could have just as easily gotten R1b from a group adjacent to the Steppe. Like the Danube.

      More ancient DNA should tell us.

  10. @Ryan,
    "I haven't seen Maju deny this. His view IIRC has always been that P312 came from central Europe "

    No he thinks P312 is from SW Europe.

    1. Not last I saw, but I'll let him speak for himself. I certainly think it spread from Central Europe, but that doesn't make it an IE marker.

      I have to say though - Dene Caucasian is looking better these days.

    2. Krefter understood me better: S116 = P312, so all what I just said or in synthesis using the synonym term for the marker and haplogroup: this lineage seems to have originated in France, rather towards the south: one branch went South (DF27), another to the Northwest (M529 = L21) and another to the East (U152). That's what you get if you "join the dots" on the map and seems confirmed by all that S116* that might exist in the Massif Central (unconfirmed but heard of it) but also in the Basque Country and Pyrenean region (confirmed) by Valverde et al.

      That's the apparent internal structure of R1b-S116/P312 and I've seen nothing yet that counters this model. The appearance of some M529 in Ireland in BB chronology is just consisten with it, as is the appearance of some S116* in Germany in the same period.

      What may not be so consistent (??) is the "teal" element but that's not Y chromosome genetics.

    3. Notice that S116* (i.e. S116(xDF27,M529,U152) is also very important in Ireland. So far no other substantial pool of S116* is known, certainly not in Central Europe. It may be just one or two subhaplogroups but, even in that case, they add up to the Western or Atlantic basal diversity of R1b-S116 and not to the Central or Eastern European one, which seems quite limited (basically some U152, again check Valverde's study).

    4. Yah, as I mentioned above, I meant S116`s parent, not S116. Oops.

      I`m going by Myers 2010 when I say Iberia, but I guess he didn't go into as much detail as later studies. I think there's a pretty clear path of R1b moving from the Black Sea, up the Danube, with a final expansion of S116 from the Atlantic coast.

    5. Ah, OK. The problem with Myres 2010, which is still a key referential study on this matter, is that she did not study DF27, therefore inflating the S116* in Iberia a lot, now known to be in essence just DF27 (Valverde is very clear here). But there is still a remnant S116*, which seems to be most common among Basques and Irish but could also be quite important in France, maybe Britain as well (not yet properly studied AFAIK).

      The parent of both S116 and "Nordic" U106 are M412 and L11 (or its equivalent, not sure if L11 is still considered a major marker) and they seem to point (Myres) to either the southern reaches of Central Europe, Italy or Iberia, judging on basal paragroup presence. In any case, these are just "thin trails", the really expanding lineages were S116 ad U106, all the rest is just minor and rather noisy stuff.

      For a most simple reconstruction (just doing it right now), I'd place the S116 "centroid" in Aquitaine, the U106 one in the Netherlands and then it depends what you do with the upstream M412* remnants: I'd place a dot in SE Iberia and another in Slovakia (my best hunch). The resulting overall centroid goes to the Upper Rhone (not too far from Gurgy incidentally) but it's anyhow very approximative. If we only consider U106 and S116, then the centroid goes to Normandy instead. So somewhere in France again, although seems to tend towards the North rather than the South when we include all M412.

      But in any case, this is a most thin trail. In fact the precursors of U106 and S116 may well have arrived to their respective expansion centers by two different routes in the context of, for example, a much more Y-DNA-diverse Neolithic expansion.

    6. I'd suggest to you that this structure is the mark of several bottlenecks related to crossing from the Danube basin to the Rhine basin and the Rhone basin.


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).