September 9, 2015

Detailed analysis of ancient Atapuerca genomes

Everybody seems to be buzzing about this study on Atapuerca (El Portalón) site's Chalcolithic and Bronze Age genomes and I do think it is indeed worth taking a good look.

Torsten Günther, Cristina Valdiosera et al., Ancient genomes link early farmers from Atapuerca in Spain to modern-day Basques. PNAS 2015. Open accessLINK [doi: 10.1073/pnas.1509851112]


The consequences of the Neolithic transition in Europe—one of the most important cultural changes in human prehistory—is a subject of great interest. However, its effect on prehistoric and modern-day people in Iberia, the westernmost frontier of the European continent, remains unresolved. We present, to our knowledge, the first genome-wide sequence data from eight human remains, dated to between 5,500 and 3,500 years before present, excavated in the El Portalón cave at Sierra de Atapuerca, Spain. We show that these individuals emerged from the same ancestral gene pool as early farmers in other parts of Europe, suggesting that migration was the dominant mode of transferring farming practices throughout western Eurasia. In contrast to central and northern early European farmers, the Chalcolithic El Portalón individuals additionally mixed with local southwestern hunter–gatherers. The proportion of hunter–gatherer-related admixture into early farmers also increased over the course of two millennia. The Chalcolithic El Portalón individuals showed greatest genetic affinity to modern-day Basques, who have long been considered linguistic and genetic isolates linked to the Mesolithic whereas all other European early farmers show greater genetic similarity to modern-day Sardinians. These genetic links suggest that Basques and their language may be linked with the spread of agriculture during the Neolithic. Furthermore, all modern-day Iberian groups except the Basques display distinct admixture with Caucasus/Central Asian and North African groups, possibly related to historical migration events. The El Portalón genomes uncover important pieces of the demographic history of Iberia and Europe and reveal how prehistoric groups relate to modern-day people.

It must be said to begin with that the Atapuerca samples are actually similarly related to Basques as to Sardinians: they have more Paleoeuropean admixture than Sardinians and early European farmers but not quite as much as Basques. The various formal analyses, such as the one displayed to the right (fig. 3-B) confirm this intermediate position between what I'd call First Neolithic and Atlantic Neolithic genetic configurations, which can be directly associated to modern Sardinians and modern Basques respectively. 

Other data such as the mtDNA pool or lack of the lactase persistance allele also places them rather in the First Neolithic group in spite of their greater Paleoeuropean admixture, which is undeniable. The El Portalón samples have much lower frequencies of mtDNA H and U than Neolithic Basques or Burgundians, let alone the "hyper-modern" Neolithic Portuguese with their >80% mtDNA H (Chandler et al. 2005).

Similarly their lack of the T-13910 lactase persistence allele, dominant among modern Basques and many other Western Europeans, and already detected in at least some Chalcolithic Basques from an intermediate area (Upper Ebro banks), suggest that the old archaeological and anthropometric narrative of Mediterranean colonists migrating up the Ebro and establishing intermediate (but still rather Mediterranean) populations in the Upper Ebro banks, somewhat distinct of proto-Basques proper of a distinctive Pyrenean (Keltid?) type, was not completely wrong. It is true that the Pyrenean type cannot be anymore considered a pure derivate from Paleoeuropeans but rather a mixed population with strong Mediterranean Neolithic input but there is still some distinction very apparent in the archaeogenetics we know so far that cannot be totally ignored. 

Not a highlight of the study or the press release but I think it is very worth mentioning that the Bronze Age ATP9 woman also shows strong affinity with Britons, particularly Cornish and Scots. In general there is stronger mainland European affinity for this sample but still Basques and Sardinians, as well as the mentioned Britons, are the closest matches.

The general details of the samples are as follow:

As you can see only four samples had sufficiently good coverage to be considered for most analysis. Mitochondrial DNA, as usual, is the exception (all are good enough) but Y-DNA cannot be reliably assailed from such a poor quality sequence, nor actually much more that is not fuzzy. 

Stopping for a moment on Y-DNA, it must be said that haplogroup H2 was formerly known as F3, being a rare West Eurasian haplogroup (with some presence in the Persian Gulf and Zagros mountains area, as well as a scatter through Europe) which has seen its phylogeny recently refined under H (otherwise a South Asian and Roma lineage). 

I2a2a is not the typical Sardinian and Pyrenean lineage that is generally considered to be part of the Cardium Pottery package, originating probably in the Balcans, even I first thought it would be (thanks to Krefter for the correction). Instead it is some other Paleoeuropean lineage, which is today most concentrated in Northern Europe (→ map).

Genetiker claims that ATP3 should be R1b1a2-M269, while ATP17 would be also I2a2a. However given the very low coverage of these genomes, I would take such claims with great caution. As I've written somewhere else the question is not anyhow if there are some R1b of any sort here or there because the M269 → L11 stage has left only a scattered legacy, except for two large subhaplogroups: S116/P312 and U106. These two subhaplogroups are the big mystery and for them we only have so far late Chalcolithic terminus ante quem dates from Bell Beaker Germany and Corded Ware Sweden respectively. The lack of Atlantic Neolithic samples, be them British, French, Basque or Portuguese surely has a lot to do with this lack of evidence because, you know, follow the trail of modernity traits such as early "modern" mtDNA pools (Neolithic Basques and North French, Portuguese also) or early presence of the lactose tolerance allele (Chalcolithic Basques and Swedes). The answer to this pressing question must be in the Atlantic basin of Europe, just sequence it for Chaos' sake!

As for the mtDNA, the genetic pool is partly typical Neolithic (K, J, X) but with notable highlights: on one side, there are 3/8 U5 lineages, which are surely a legacy of the admixture with Paleoeuropeans, and, on the other side there is this H3 haplogroup that is nowadays important especially in Southwestern Europe. Unlike H1 or H6, which are known to have been carried by Paleoeuropeans, H3 so far has only been found among Neo-Europeans (i.e. from Neolithic onwards) but we do not know so far its ultimate origins for sure, as happens with haplogroup V (similar situation). 

In any case the low frequency of haplogroup H, makes the overall pool not yet "modern", unlike what happens with at least some Neolithic Basques (Paternabidea) and certainly with proto-Basques from Chalcolithic onwards, as well as other mentioned groups like North French Neolithic, Portuguese Neolithic and later also German Bell Beaker peoples. A more complete comparison of various ancient mtDNA pools can be found in fig. S4. 

Going back to the main focus of the study, which is autosomal DNA, I guess that we can continue with the following excerpt from the ADMIXTURE analysis:

Fig. 3(A) - Population structure of ancient and modern-day individuals. (A) Admixture fractions among modern-day individuals from Eurasia and North Africa together with 16 ancient individuals. Only ancient and modern-day individuals from Southwestern Europe are shown (see Dataset S1 for the complete plot with all individuals). Admixture components are labeled based on the populations/geographic regions in which they are modal.

The two main components are (1) the Paleoeuropean of HG modal, whose fractions may well be close to real in this case, at least judging on the strict alignment of Europeans between Paleoeuropeans and West Asians  provided by the PCA with North African samples (see below) and (2) the West Asian or EEF modal. We see that the First Neolithic populations were around 20-25% Paleoeuropean (the orange fraction would be West Asian or mostly so), while the Atapuerca samples show double HG scores, c. 40-50%. Modern Basques are even higher, around 50-60% Paleoeuropean in this analysis (although obviously a small fraction should be attributed to the Early Farmers' inflow).

In any case we see that in general the increase of the Paleoeuropean fraction is very notable in Atapuerca and Gokhem, and was plausibly even greater among other Atlantic early farmers, judging on the modern Basque and Gascon ("French South") data.

The next most relevant component is the Caucasus/Central Asian one, which should be attributed to the Indoeuropean expansion almost certainly. These would also have contributed with a proportional fraction of Paleoeuropean blood (maybe double than the black segment or something in that line). We see that this Indoeuropean influence is most important among the French and much less relevant in Iberia (but still much greater than among Basques or Sardinians). Nothing really new in this after Lazaridis 2014, Haak 2015 and Alentoft 2015, just a complementary perspective on the same issue. 

As for the North African component it seems to follow the same pattern of it being mostly a feature of the Western third of the Iberian Peninsula rather than something attributable to "historical events" (such as the Muslim period or the Phoenician conquest, much more influential in the South and East instead). Again lack of ancient DNA from those regions have hindered the understanding of the origin of this component which can be either (1) a Neolithic founder effect or (2) a Paleolithic founder effect dating from as early as the Solutrean-Oranian interaction around the Last Glacial Maximum. The fact that La Braña in some analysis shows North African or otherwise African affinities suggests that it is a Paleolithic element that has remained basically a Western Iberian thing (excepted the occasional founder effect such as certain district of Northern Wales and some diffuse scatter of related markers like E1b-M81 along the Atlantic coast of Europe).

If this interpretation is correct, then it strongly weights against the hypothesis of widespread (Western) Iberian origin of Chalcolithic (Megalithic, Bell Beaker) founder effects through Western Europe, because we would see it everywhere. Instead I'm much more inclined for a major role of what are now the Western parts of France, which were clearly involved in the (late) Neolithic colonization of Britain. I'd rather advocate for a "French" origin of Y-DNA haplogroup R1b-S116 for example. 

True that studying the Hexagon is sadly hampered by a hostile bureaucracy, legal framework and state ideology but no such obstacle seems to impede the study of ancient Britons, for example. In due time I guess. 

Another highlight is the TreeMix graph. However I must say that I am much more comfortable with the S10 tree rather than the highlighted one in the main body of the study. The main difference is that Ötzi (Iceman) is in this one placed near Gokhem and also with some Pitted Ware type admixture. And this is new and therefore a bit disturbing. Instead the S10 tree is not so surprising:

Compare with the simplified 2-A tree to the right...

You choose, as not enough data is provided in the paper to argument one choice above the other.

It is apparent in any case that Gokhem seems to have strong admixture from a source that is precursor of Pitted Ware Ajv58. This may also be the case of Ötzi (?), although it'd be weaker. 

Also, in spite of claims of La Braña admixture in Atapuerca I do not see that element clear in either graph. In one it is apparent that the Paleoeuropean influence is rather Lochsbour-like, while in the other it seems something rather ancestral to both WHG individuals. 

Confused? Well, what about the pre-Motala 46% admixture in Mal'ta 1? That is indeed new and it is consistent in both graphs. I would think it is suggesting that, rather than Ma1 (ANE) admixture in SHG, what we have is an ancient flow from Europe to LGM Siberia. Which can that be? Gravettian culture only. Oddly enough it makes some sense but, if this is correct, then it should also say that all the ANE buzz was a bit nonsensical after all. Confounding factors at play. 

To finish this review I'll copy here the various Principal Component Analyses, eye-candy, which in this case have diverse sampling strategies. Oddly enough for someone who has asked for Europe-only PCAs, it seems to me that in this case the one including West Asians is a better approximation to the reality of Atapuerca's ancients:

Fig S7

As you can see, the three Chalcolithic samples properly sit in this case right between Basques and Sardinians. The position of the Bronze Age sample should be explained by the British affinity, rather than being truly closer to Spaniards.

All this is less clear in the Europe-only analysis:

Fig 2-C

They do appear correctly between Basques and Sardinians but a shallow look could make them look as almost Spaniards, something that is not quite correct. Again the Bronze Age sample is pulled towards Britain, incidentally (and misleadingly) overlapping Spaniards. 

Someone somewhere asked for a PCA with North African samples. Well, this study also includes one. However no North African influence is apparent in any European, ancient or modern, so the best use we can have for it is using the lineal reorganization of Europeans between the Paleoeuropean and the West Asian polarities as a ruler of sorts to estimate the levels of admixture of the various populations:

Fig S6

And that's all for today. Soon to come (hopefully): new Y-DNA age estimates (I promised more than a month ago, shame on me), genetics of Baltic peoples (new study) and the Vasco-Nubian linguistic connection (something I've been ruminating for more than a year now).

Update (Oct 16): I just stumbled on this review[es] of a Bronze Age weapon hoard in Los Cascajos (Upper Rioja, unrelated to the Neolithic site of the same name in Navarre), not too far from Atapuerca, and the authors mention a practical identity in design between one of the swords and another one from Stratford (England), which is known to be of local manufacture. Hence he ponders that at least some of the weapons could have been imported from Great Britain.

The, very apparently ritual, destruction of the weapons, including the use of fire, before their burial also resembles a lot British rituals of the same period.

All this, along with other well known Iberia-Britain exchanges in the Bronze Age, may help to support the notion of ATP9 being partly of British ancestry.


  1. Personally, I would like to see future studies define North Africa beyond a generic, modern Berber.
    I think pre-Berber will show the similar dynamics as Europe, with communities of hunters, farmers and pastoralists.
    Along with the Atlantic DNA vaccuum, these two areas are the most frustrating.

    The linguistic connection you mention is very interesting. Never heard of that.

    1. Well, sequencing Taforalt would be quite interesting indeed. Not sure if possible though.

      The Vasco-Nubian thing is something that I just happened to say: well, what's up with the "Basal Eurasian", sometimes "Dinka" thing? Probably nothing but let's have a look with Swadesh mass comparison tools anyhow. I found a Nubian list online (incomplete) and did the job out of mere curiosity and THIS is the result. IMO the percentage of apparent cognates is too high to be a mere coincidence, and also there are a handful of very interesting "oblique" cognates (things like the Nubian word for "hot" does not mean "hot" in Basque but instead "fire", not sure if it's exactly that way: from memory right now). So AFAIK Basque has two vocabulary relatives: one is proto-Indoeuroeopean (quite notable also and partly researched separately by linguists like Roz Frank, who calls the alleged common root "Paleoeuropean") and another is, seemingly, Nubian languages (part of Nilo-Saharan).

      If this is correct it would mean that Nilo-Saharan (or at least its Nubian branch) was once more important towards the North, with presence maybe in Mesolithic and Neolithic Levant, now and historically replaced by Afroasiatic Semitic. All this, be Afroasiatic or Nubian, is clearly related in my viewpoint to the presence of "Basal Eurasian" (African-like) component, Y-DNA E1b-M78, etc. not just in West Asia but also in the mainline European Neolithic. Obviously Basques as such do not show any obvious genetic connection to Nubians (even E1b is most rare around here) but Basques are only the last link in a long chain, yet the only one that remains unbroken. In cultural aspects there might be some surviving elements however: stuff like bells in ankles for dancing or, more obscurely, maybe carnival elements... but researching that would need of a quite seriously dedicated cultural anthropologist, falling a bit beyond my scope.

    2. This is really, really fascinating. Obviously modern Basques and Sudanese are very different but going back to the Aterian and beyond, the regions at times were part of the same industry. I also came across some papers on evidence of impresso farmers in Northern Sudan. And the obvious connection with T13910/P297/M269/V88

      Given the two never succumbed to IE or AF, it seems possible

    3. It has nothing to do with Aterian (which is a NW African culture mostly and way too old, from the OoA epoch) nor with the more real Oranian-Solutrean interaction between South Iberia and NW Africa (that's almost certainly a confounding factor here but not the real connection). IMO this has to do with mainline European Neolithic, E1b-V13, the "Basal Eurasian" thing and the Palestinian (Naqab Bedouin) connection of the putative source population.

      What happened in Palestine before and in the Early Neolithic? That it did have some influences from the Nile area, probably traceable to Sudan/Nubia. Until this analysis I though we were talking of the Northern branch of Afroasiatic only, which does seem related to Capsian culture in NW Africa (with possible source in Upper Egypt or Lower Nubia) and these influences in Mesolithic Palestine. But now I suspect that almost necessarily Nilo-Saharan (specifically the NE branch, which is Nubian) was also involved. As you probably know there is a complicated overlap between Afroasiatic and Nilo-Saharan in the area of the Middle and Upper Nile, where both families surely originated. Can't say much more, really but in any case I don't think there's any NW African connection but rather one via the Eastern Mediterranean and the mainline European Neolithic.

    4. Maju the 'nilotic' genetic component is really saami-like and probably comes from a migration to north africa from iberia 20k years ago.

      The riffians were very tall people who used northern european fishing implements.

    5. Nilotic = Sámi? LOL.

      Keep up with your buffoon role (but find a comedy club, this is a serious blog). I say that because I can't think of anything else that sounds positive. My first impulse was rather: get lost!

    6. Pretty ironic for you to call someone a buffoon considering the respect you have among others.

      You are an ignorant fool in anything that goes outside your manic ethnic self-obsession, and this is yet another area you open your stupid mouth to insert a foot without bothering to check the most basic facts.

      Ancient north african mtdna has been sequenced and found to have many of the same markers, plus riffian artifacts show the same type of fishing implement and sami show 'niltotic' component in their genetics similar in nature to the 'american indian' in europeans. How you can't know this is completely beyond me, but like I said tunnel vision on your self obsession can be a terrible thing.

      As for france and briton megalithic, well of course. The areas where r1b spike being where they formed? Which is also where YOU SHOULD KNOW all the archaeology says the megalithic as we know it formed?! What an amazing, shocking event!

      This obvious fact is never 'discovered' because those countries don't want it discovered, they have a much different narrative to tell, where europeans have no place in the past or the future.

      Sorry to see you can't keep it civil, but you are the one with the head full of rocks and the inability to act like a normal person, not me. A guy with a banner begging for donations calling me a buffoon and putting on airs, truly a wacky world we live in.

    7. Now that you have vented off, I hope you do get lost.

      As for Taforalt's mtDNA, the study is not available (never was, just a power point presentation) and I've been one of the most interested in it in the past. Assuming the data is correct, all it shows is a massive impact of ancient Iberian genetics at the Oranian (alias Iberomaurusian) genesis, including lots of H, something that modern mtDNA data seems to confirm. Old news.

      As for Megalithism, it is a thousand years older in Portugal than in Brittany. You should know that.

  2. The calls Geneticker got for ATP3 making him R1b1a2 has been confirmed by several reliable people.

  3. @Maju,
    "I2a2a is the typical Sardinian and Pyrenean lineage that is generally considered to be part of the Cardium Pottery package, originating probably in the Balcans. "

    You're confusing I2a2a-M223 with I2a1a1-M26. I2a2a has also been found in Megalithic Spain, Yamnaya, and Bronze age Hungary. Today it is most popular in Central Europe, but is scattered basically everywhere.

    Also, H2 has been found in Neolithic Germany and Hungary, and there's a debatable one from Megalithic Spain.

    So, yeah you're that the Y DNA gene pool was still "First Neolithic". Looks like the ancestor of R1b-L151 may have been present, but I guess massive founder effects after 3500 BC lead to its dominance.

    "The answer to this pressing question must be in the Atlantic basin of Europe, just sequence it for Chaos' sake!"

    IMO, Lactose tolerance alleles in a few sites and a high amount of H isn't good evidence of gene flow from the Atlantic into other parts of Europe. A finding of R1b1a2 from 3500 BC is good evidence though/

    IMO, frequency of H isn't a good way to test modernity. mtDNA in Europe probably isn't very uniform, so there probably isn't a such thing as modernity.

    The possibility of an Ancient people with insane frequencies of mtDNA H who migrated across Europe raising frequencies of H isn't realistic IMO. All the H clades split 8,000+ years ago.

    A people with a high frequency of closely-related H clades(H1 clades???) that they spread across Europe is realistic IMO. So, we need high resolution ancient samples to get evidence of such an event. A high frequency of generic H doesn't in an Ancient people isn't good enough.

    When looking at deep subclades there's a lot of sharing between Neolithic and modern Europeans. J1c, T2b, H1, H3, HV0 are all lineages that take up almost 50% of European mtDNA, are rare everywhere else in the World, and popular in Neolithic farmers.

    I collected almost 1,000 Basque mtDNA samples(all H was fully-sequenced) from a recent study, and H3c is the most popular form of H3 in modern Basque and was found in El Portalon. K1a2 is considered a European branch of K1a and two were found in El Portalon. U5b3 is most popular in Iberia and Italy and was found in El Portalon. The other mtDNAs U5b3, K1a2, and H3, are rare but exist today. There's nothing that suggests El Portalon mtDNA is non-modern.

    1. I'll correct the I2a2a thing. So do you think it's a locally adopted HG lineage?

    2. "IMO, Lactose tolerance alleles in a few sites and a high amount of H isn't good evidence of gene flow from the Atlantic into other parts of Europe. A finding of R1b1a2 from 3500 BC is good evidence though"...

      1. In order to find it, you will first have to search for it. Hence my desperation for lack of Atlantic genomes... yet.

      2. R1b1a2-M269 without further definition is meaningless. It is too far upstream of S116 and what expanded was S116 (and U106 in the North). Others talk about L51 (M412), same thing although at least it'd be already within the European specific branch. What I want to find is the oldest S116, and so far it has not been found before German BB. But it should be there in the West, where nobody is looking.

      "Modernity" traits are not direct evidence of not sequenced Y-DNA, of course, but they are good indirect evidence that only seems to require that someone goes and take a look where they do exist. In Germany's BB it worked totally according to expectations, so I don't see why it won't work in all or most of the other regions I mentioned.

      It's a theory that makes a prediction, just go and test it...

      "mtDNA in Europe probably isn't very uniform"...

      In Western Europe it is. There is some cline towards the East (less H, more U) but I'm not too interested in the East for this issue. "Modernity" in mtDNA means ~50% H, maybe 20% U (mostly U5) and the rest "Neolithic" stuff (K, X, W, JT, N1). It's very approximate but when you have 70% "Neolithic" and 25% H, it's clear that it's a pre-modern population as much as one that is 100% U.

      The T-13910 allele is also very important. We know it was there but not among First Neolithic peoples nor among the later Indoeuropean intruders, nope. We know it was only in the Atlantic fringe, with a rare exception in one of Gotland's Neo-HG individuals. This is like flashing lights about modernity and it is also flashing lights for R1b-S116 or U106 surely.

      "All the H clades split 8,000+ years ago".

      Are you even serious? We know for a fact about subclades of H, H6 specifically, before the end of the Ice Age! Your prejudices impede your vision.

      Anyhow I have investigated H a lot and it must be extremely old, roughly first Upper Paleolithic. Just that HVS-I methodologies don't allow us to discern it for sure in most cases and, when accidentally does, as with the H17'27 of Sunghir, the recentist chorus (Jean Manco directing) cry so hard that hide the evidence under their ugly noise. But for me it's clear that UP Europe is full of H, HV and stuff like that. And in some instances it has been demonstrated beyond any doubt. It's not just all U, that would absurd.

      Also H is one of two massive star-like nodes in the human genome, the other one being M. Exactly the same that M corresponds to the very first expansion of H. sapiens in Asia, with a dozens of niches ready to be occupied without any resistence, H must correspond with the early UP colonization of West Eurasia, when nearly the same happened. U probably expanded later, with Gravettian.

      But you have to see outside the box, as is my rule, to see this. You have to count mutations from the root, not present day, because a established haplogroup with minimal effective population (around Ne>10) will evolve very very slowly, with novel groups being reabsorbed by mere drift in most occasions unless the population is continuously expanding or a major fluke happens.

      While you keep thinking inside the box all you'll see will be those four walls. They are not there: there is no box but it's just an imaginary barrier that blocks your thought. It's like Plato's cave but nobody prevents you from walking outside.


    3. ...

      "A people with a high frequency of closely-related H clades(H1 clades???) that they spread across Europe is realistic IMO".

      Yes of course. But H1 is also found in places where R1b almost is not present like the Baltic. That's probably because it's the oldest established H subclade and probably has two centers a Western one associated with R1b and a Baltic one not associated with it.

      I most cases we don't have enough data but of course I do look for H1 or H3 presence when available, H1 particularly is a good signal. But lots of H is just H*, today and in the past. So it's more a matter of apportions.

      H is not "Neolithic" in Europe anyhow: it was clearly present in Paleolithic Europe, just that not in Germany apparently (but Germany is not the measure of all things). While H did probably re-expand sometimes with Neolithic populations, most of its expansion is from later times Chalcolithic notably, coincident with the increase of HG component in Western and Central Europe, coincident with the Megalithic expansion.

      "There's nothing that suggests El Portalon mtDNA is non-modern."

      Too low H, too much U. It may be "modern" for Russia but not for Western Europe. I don't want to make too much of it but it's clearly very low in H. It may a fluke anyhow.

      But they are not modern in lactose tolerance either, nor in the discernable Y-DNA... For me they are transitional of sorts but not modern in a typical Western European way at all. Definitely not modern in Basque way either, not at all.

      It can be argued however that they are not First Neolithic either (too much U5 particularly, too much Y-DNA I2) and all that is almost for sure due to the extra WHG admixture implicated.

      We see also some of this stuff in the Basque Neolithic of the Ebro: lots of U*, but very little U5, which is almost the only kind of U sampled in HGs in Iberia. It may still be HG but from non-local populations, maybe carried from some mixture in the Mediterranean or whatever. Instead piedmont Paternabidea was already very modern with almost exactly the right amount of each haplogroup. And yet a bit further North, by the coast, the mtDNA pool was 100% HG in this first Neolithic (U5 plus some H). So in the very narrow space of the Basque Country and in the very original time of the first Neolithic we see three different pools: one in the more Mediterranean area of the Ebro, another (the modern one) in the intermediate area and yet another one in the Atlantic area that was still purely HG in appearance.

      Similarly we also see later a lactose tolerant population and a lactose intolerant side by side. We cannot pinpoint their origins because they are all mixed in a single "military" cemetery but IMO the lactose tolerant ones were northerner "proto-Basques", while the intolerant were more in the line of the First Neolithic.

      En fin...

    4. "The possibility of an Ancient people with insane frequencies of mtDNA H who migrated across Europe raising frequencies of H isn't realistic IMO. All the H clades split 8,000+ years ago."

      Maybe there was selection in place as well i.e. what if mtdna U had an HG advantage and H had a farmer advantage?

    5. mtDNA H was in Europe well before the Neolithic. Some basal H clades were already formed back the Epipaleolithic.

  4. H1bz mtdna was noted by brotherton in 2013 in neolithic central Germany, it was Genbanked by ian logan.
    In 2015 Haak found that the ydna for this H1bz sample is T1a at 5500BC to 5500BC in can you make a claim that this is not central european but a baltic marker?

    1. I didn't make any such claim. Who did? How?

      What I say is that, for what we know, mtDNA H was absent in Central-North Europe in the late Paleolithic and Epipaleolithic, while it was present in Iberia and Eastern Europe (and probably also the Balcans, maybe Italy too). So it's resurgence in the Neolithic and particularly Chalcolithic probably owes to the various Paleolithic "refugia" and not just one. And that's probably also the case of H1.

  5. Hi there! First time commenter, occasional reader...

    As someone who has only cursory knowledge of this field, I would be interested in your opinion: Do the recent findings of Haak, Laziridis, and this paper have any affect on the migrationist-vs-diffusionism debate regarding origins of agriculture in Europe? Looking at the ADMIXTURE analysis, it seems clear that the French and Iberian farmers mingled with hunter-gatherers quite a bit, supporting the diffusionist hypothesis. Is that interpretation completely off-base?

    1. Obviously the recent studies, but also previous ones on ancient mtDNA, have changed the Neolithic paradigm quite radically strongly redefining it in favor of migrationism. However the issue is not without subtleties, gradations and problems of its own, such as the back-flow of mixed populations from the West, possible early admixture in the South too, etc.

      But in general terms the Paleolithic continuity models have just been smashed to pieces. Even the most Paleoeuropeans of all clearly have major West Asian components and the very Paleoeuropean component is surely from various sources (some picked in the Aegean by the Asian migrants, some absorbed upon arrival to certain regions, some backflow from the West, some later arrival from Eastern Europe...)

      I don't think we can talk of diffusion anymore as a general pattern for Neolithic in Europe. Of course locally there may have been some instances of it, but more commonly admixture rather than mere contact, but the general pattern is clearly primarily migrationist.

  6. Hi there! First time commenter, occasional reader...

    As someone who has only cursory knowledge of this field, I would be interested in your opinion: Do the recent findings of Haak, Laziridis, and this paper have any affect on the migrationist-vs-diffusionism debate regarding origins of agriculture in Europe? Looking at the ADMIXTURE analysis, it seems clear that the French and Iberian farmers mingled with hunter-gatherers quite a bit, supporting the diffusionist hypothesis. Is that interpretation completely off-base?

  7. "If this is correct it would mean that Nilo-Saharan (or at least its Nubian branch) was once more important towards the North..."

    I have also read claims by some scholars of possible similarities between Basque and Mande languages; Fula; the Chadic and Cushitic branches of Afro-Asiatic [closer than Berber and Semitic]; and possibly also the Tuareg [rather than non-Taureg] branches of Berber. Now you have mentioned Nubian, a Nilo-Saharan language. Which is interesting.

    "IMO this has to do with mainline European Neolithic, E1b-V13, the "Basal Eurasian" thing and the Palestinian (Naqab Bedouin) connection of the putative source population..."

    I think that the Basque/Nubian linguistic connection and the 'basal Eurasian'/EEF/Dinka genetic connection may be linked in more than one way. Your E-V13 hypothesis is undoubtedly a good fit.
    However I cannot ignore the additional likelihood of a link to a Holocene 'green Sahara' source. As I said, other scholars identified linguistic connections between Basque, and Mande, Fula, Chadic, and the Tuareg branches of Berber.
    I believe that there are genetic parallels with this also. E-V13 accounts for only about half of all 'E' in Sardinians, and about one third in Sicilians [two European populations with the highest levels of 'EEF'/'basal Eurasian' component]. In addition to the diverse African clades of 'E' in these populations, there are of course different clades of haplogroup 'A'; different clades of R1b R-V88; and also haplogroup 'T' ['T' being found at notable frequencies in several African populations - Berbers, Tuaregs, Chadians, Egyptians, Sudanese, Eritreans, Ethiopians, Somalians, Kenyans, Ugandans, Tanzanians, Rwandans, Zambians, South Africans, Angolans, Cameroon Fulani, Cameroon Kanuri].

    HLA haplotype A*30:02--B*18:01--C*05:01 is found in Europe at a very high frequency in Sardinians [with lower frequencies in Basques, Balearic Islanders, Catalonians, S.French, Swiss, and Italians]. However the haplotype itself and all of its components are more diverse in Africa. There the haplotype is found in Berbers; Senegal (Mandenka); Gambians; Ghanaians; Cameroon (Uldeme - Chadic speakers); Sudanese; Kenya (Luo); Mozambique [probably many more - compounded by insufficient sampling].

    In addition to your E-V13/Palestine hypothesis I would also add the following:-

    Green Sahara [Niger/Chad/S.Algeria/S.Libya] --> Tunisia/N.Libya --> Sicily/Sardinia/Corsica/Balearics --> Catalonians/Basques/S.French/Swiss/Italians

    I believe that in addition to Basque, there have also been claims of an Afro-Asiatic-like substrate in Welsh [something to do with v-s-o word order]. I don't know whether 'Morris dancing' or carnivals represent cultural remnants of north African migrations into southern Europe but this could be worth investigating.

    1. I remember old claims about relations with Mande, which was one of the reasons I began this personal quest to understand Basque (and by extension Europoean) origins: nobody seemed to have a clue, so I needed to make up my mind on improved knowledge. That was back in the 20th century and I have never again hear of it.

      More recently I have also read vague claims about Dogon affiliation specifically but all under paylock and I'm not going to pay for something that looks like an erudite sort of scam. To begin with Dogon is at least two different language families and to continue mainline Dogon is part of Niger-Congo, so there should be a general relation with all the family and not just a branch. Otherwise it's like saying "Quechua has a strong relationship with Albanian", the obvious question arises: what about the rest of Indoeuropean.

      Talking about Quechua that is another of the claims I've read about. But something is common about all these fringe claims: they rely on particular interpretations of the "comparative method" (which in principle should be more solid but then depends heavily on individual linguistic dedication and self-honesty, because it's not an easy task). I don't go on such complicated endeavors: in theory better, in practice highly subjective, instead I use the mass-lexical comparison which at the very least is easily reproducible by anyone and gives you approximative figures of affinity. And then again I don't put my findings behind a paylock but freely share - "evivva il communismo é la libertá!" You judge.

      Finally I want to say that Africa is big and the same that I don't see any particular similitude with Afroasiatic (at least not Semitic), I doubt that a Nubian connection automatically leads to any other one in the continent. It'd be like claiming that a relation with Japanese leads to a relation with Dravidian or Chinese or Cambodian... Africa is that kind of big and diverse, like Asia or more (because it's much older in what respects to Human paleohistory). And the kind of genetic and archaeological links we can see in North Africa or West Asia are all with the Sudan-Ethiopia area and not or very limited with West Africa. Now, prove me wrong... do a mass lexical comparison and go clearly above the ~7% likely noise (coincidences) threshold. :)

    2. As for the NW African connection, when you go beyond immune markers (which are strongly subject to selective pressures!!!) there is only a very weak link Sardinia-North Africa, almost invariably mediated by Iberia (where it is also quite weak, although probably a bit stronger). In a few cases it can be, I guess, a more direct relic of the Carthaginian colony period too, but in any case the link with NW Africa is very weak in the island, although not zero. For most purposes, after the LGM (it seems to me: there was a period of intense interaction but it seems most ancient) the Western Mediterranean has behaved as a barrier rather than a corridor, even in the areas where the distance is very short like Gibraltar Strait. This is something that has been discussed time and again and the hard facts are there: J1 is almost not found in Europe, J2 or other European Y-DNA markers are almost not found in NW Africa. This Y-DNA J duality, which has a Neolithic or Mesolithic timeline probably in both shores, defines the temporal limits of the trans-Mediterranean interactions in the West: what there is, appears pre-Neolithic, surely even pre-Mesolithic. In fact the best candidate for such interactions, which left most of its legacy in form of SW European mtDNA in NW Africa (~30% is H1, H3, H4, H7 and V), is the Oranian genesis of quite apparent Iberian Gravetto-Solutrean roots (but with minor backflow, which is probably the cause of the lesser NW African signature in West Iberia).

    3. BTW, notice that in the ADMIXTURE analysis above, La Braña is again showing a small but apparent North African component, absent in the other populations (except modern Spanish, among whom it is even larger). Only NE1 (Hungary's Neolithic) shows something similar, which may well correspond to the thin NE African ("Basal Eurasian") legacy of EEFs (but that explanation does not work for La Braña). So for me La Braña is documentary support for my LGM NW African interaction in Iberia. However it's quite possible that this NW African component grew later, either because of a second NW African flow in the Neolithic or because of internal rearrangements in Iberia favoring southern and particularly SE sources of new population (I'm thinking in Portugal particularly, of course).


    First google result talks about 'nilotic' component in sweden. holy shit, what an intellectually lazy fucking buffoon you are.

  9. «In fact the best candidate for such interactions, which left most of its legacy in form of SW European mtDNA in NW Africa (~30% is H1, H3, H4, H7 and V), is the Oranian genesis of quite apparent Iberian Gravetto-Solutrean roots (but with minor backflow, which is probably the cause of the lesser NW African signature in West Iberia).»

    If we isolate Berber groups, do you think that the SW European mtDNA percentage would be bigger in Northwest Africa?

    1. There's no fundamental difference between Berbers and "Arabs" in NW Africa, the latter being in essence nothing else than Berbers who have lost their language through the last centuries. It's a process still happening today.

      Said that, I vaguely recall someone telling me that it was indeed the case but only (?) with specific Berber groups such as those from Northern Morocco (Tuaregs of course too but that's a peculiar founder effect).

      In any case Tamazight being an Afroasiatic language, is almost certainly not the language of Oranian but that or Capsian culture (in fact Berber is too homogeneous, suggesting that an specific dialect expanded in historical times at the expense of older proto-Berber languages, probably within the Mauretanian and Numidian kingdoms). So Berbers, as relatively isolated populations, are equally likely to preserve Oranian or Capsian traits, or even older ones, they will just normally represent some more local genetics relative to the (normally) more cosmopolitan "Arabs", which will have less localized recent roots in many cases.

      Anyhow I remember back in the day it was claimed that Taforalt ancient mtDNA was almost an exact match to the modern local genetic pool. However the claim probably needs revision.

  10. "hyper-modern" Neolithic Portuguese with their >80% mtDNA H (Chandler et al. 2005).

    What do you mean with «hyper-modern» here?

  11. HI Maju
    I wonder if you still remember this from Athanasiadis 2010 paper :
    "Regarding population structure, when all of the populations were analysed, PCA identified three clusters, corresponding to: the Mediterranean groups; the two Bolivian groups; and the Ivory Coast (Figure 3a). In the Mediterranean cluster, all South European groups appeared separated from the North African groups along the first PC, except for Tunisia, which was closer to the South European groups. A similar result for the same Tunisian sample was reported in a previous study of Alu insertion polymorphisms on the X chromosome"

    the Tunisian sample was from the coastal city of Monastir,and since I am familiar with the looks of my fellow North Africans, I find it interesting that no one cared about the importance of the gene flow in the center of the Mediterranean,mostly in a north-to-south direction via the Italian corridor..
    in my view, the coastal Tunisians are the most Europid looking of all North Africans (the coastal Algerians are a close second )...therefore, my instinct tells me that, there has been more episodes of Euro-migrations to North Africa via Italy than via is true the strait of Gibraltar is narrow, but,it is also known for its dangerous currents which makes the crossing risky..
    I think the reason why the experts in the field are not looking into the Italian-Tunisian connection ,is probably because of the "Taforalt narrative",, after all, there is no equivalent to Taforalt specimen in Tunisia , therefore,everybody "reasonably" thinks that the "Sardinian component" in North Africans must have came via Iberia , it is true in the case of Morocco and most likely Western Algeria, but, not in Tunisia ,Eastern Algeria and Tripolitania..where I do speculate that the "Sardinian component" is more Italic...

    1. Alu insertions were once very talked about but I remember they were dropped at some point (around 2007?) because it was becoming apparent that they were not reliable (don't ask me why). I'm a bit surprised that a 2010 study still used them therefore.

      In any case it is quite apparent that Tunisians concentrate in most studies a more extreme Europoid or West Eurasian tendency within North Africa. However there is also the fact that the usual Tunisian samples (Berbers from the SW) have very low in diversity (see Henn's papers for example) and often produce artifacts in analyses. So I just tend to take any Tunisian anomaly as a systematic problem that may well be solved if new Tunisian samples would be produced from Tunis or any other less endogamous population. They can still be more "Europid" for whatever reason anyhow.

      In any case, much of the "Taforalt narrative" as you call it, derives precisely from Cherni 2008, who claimed that Tunisian mtDNA H was derived from Iberian one. This is complemented by other studies like Enafaa 2010, which show that North African H is made up of basically H1, H3, H4 and H7 (all them being likely original from SW Europe, although H4 seems to demand an "French" rather than strictly "Iberian" origin).

    2. Concerning you proposition of sampling Tunis, I'd say it'd be beter to sample some other less arabian influenced rural regions, because cosmopolitan Tunisians contain a significant deal of such recent influence... Also keep in mind that Cosmopolitans are also high in other kinds of mixture, with european of course being one of them.

  12. H is really complicating the paleolithic mtDNA composition of Europe I guess. I still can't see how it managed to reach similar frequencies all around the ''continent'' while being represented by a lot of different subclades each with their own geographic pattern. The fact that it's high in North Africa doesn't help us either. I really wonder what's the source of and what triggered this huge western eurasian star-like explosion...

    1. One of the main problems of understanding mtDNA H is that the "molecular clock" school generally claims it is much younger than it really is. If you count mutations from the root H should not be significantly younger than U but if you count from present day forms, then H appears very young, because it has accumulated very few mutations since its coalescence. This last is what most "molecular clock" models use but not what I believe is correct.

      My hypothesis on the matter is based on the fact that mtDNA novel mutations (at least those that manage to consolidate) only appear every many many generations (maybe several thousand years each), what really allows for a lot of issues to happen. In the case of "paleohistorically large" lineages (i.e. those that show star-like shapes, H being the largest one of those after M), I understand that drift will tend to erase the novel mutations or keep them quite small in numbers, barring new founder effects such as that of H1, that in order for novel mutations to consolidate as novel lineages very small population sizes (Ne<10 maybe) are needed, else statistics oblige and small novel lineages are "reabsorbed" into the ancestral norm almost systematically. In other words: a sufficiently large effective population size (Ne>10 maybe) blocks novel lineages from consolidating, unless they are lucky with founder effects: the "molecular clock" gets frozen.

      So it's best to ignore "molecular clock" guesstimates altogether or use a "clock" that ticks from the shared root and not backwards from present time.

      North Africa does help anyhow: the fact that all those SW European lineages being so important in the region (~30%) but not anymore associated to European Y-DNA (although these were surely more common in the past, as attested by Guanche mummies, which carried I and R1b in significant amounts, as well as other K-derived but ill-defined lineages), strongly suggests that they arrived to North Africa in the Iberomaurusian genesis and later suffered a major backslash with the Capsian culture, which brought NE African Y-DNA lineages such as E1b-M78 or J1, wiping out almost completely European Y-DNA but not quite the mtDNA.

      "I really wonder what's the source of and what triggered this huge western eurasian star-like explosion..."

      Early UP colonization of "the Neanderlands", either first UP ("Aurignacoid"), Aurignacian proper or Gravettian at the latest. The H sublineages should have been consolidated when they reached North Africa c. 22 Ka BP. I tend to associate Gravettian rather with U5, so I'd think that early UP, including Aurignacian proper is the best candidate.

    2. Of course H is very old, regardless of the the currently erroneous molecular clock has to say about it. I really don't give much importance to the estimates we often see in published papers. It's really the number of mutations from the basal node that matters and this, together with diversity and geographical distribution, is what help infere likely models of diffusion that a given clade went through. And as far as H is concerned, it is really deeply rooted in western Eurasia. But we would probably agree that it's still younger than U, which appeared to have spread first to the west shortly after having mutated from R somewhere in or around South Asia. The problem is that for none of these two big clades I'm able to adequatly identify an original male counterpart. I mean, I was always under the impression that U5 was strongly associated with Y-DNA I but the former looks a lot older. Same thing for H, being even even weirder this time because not only I do not think Y-DNA R could have come with it due to being too young, but I also can't figure out how it seems to have exploded on its own, changing the mtDNA pie charts of every european régions. My initial comment about North Africa was in fact in reference to this exact same phenomenon. One theory I tend to support is that some established north african E-M81 bearing population could have back-migrated, introtucing these southwest european H lineages to North Africa as R-M269 certainly couldn't have been responsible for such an introduction. But that doesnt't slove U6 at all though and this one requires even more thorough analysis because it's presence in Africa is extremely old, as attested by its diversity on the continent. And then we have M1, part of the far eastern M1'20'51 which is definitely the most difficult to explain.

      And what we said about the molecular clock would also be said about haplogroup X. I dont't think it's a neolithic lineage because its two subclades clearly are not distributed in a same geographical pattern. I would argue that it could have come with C-V20 in the first human colonization of Europe, because, knowing that X is not derived from R but instead ''directly'' from N, what other lineage would pretend to better candidate if not maybe N2/W (wasn't N2a found in australia?).

      If you say Capsians wiped out previous inhabitants, than why E-M81, which, through L19, clearly has a long history in North Africa, wasn't knocked off as well? And J1 most likely wasn't part of this expansion given the predominance of the recent arab J-L222.2, altough I did Identify some potentially older J1 in north Africa such as the ''basal'' J1-DYS388=13 (that's J1a in the new nomenclature) and some other P58s. But these seem to rare to have been associated with the spread of E-V65, by far the best candidate for the arrival of ''Hamites'' from East Africa.

      As for the Guanches, I'd suggest you to take these results with a grain of salt, because if you're not aware, there's always a lot of misclassified haplogroups in modern, let alone ancient, Y-DNA studies and unless a good set of Y-STR is provided, I won't not draw any conclusions on the guanches. If you want exemples of these misclassifications, I have tons of them, with one that comes to my mind right now is one presumed M-P256 in Afghanistan that really is nothing more than Q1b as clearly evidenced by its Y-STR haplotype... In any case, a study of canary islanders would certainly help us identify the original guanche lines.

    3. You're probably right that U is older than H, while probably younger than R0 (which is I guess what I had in mind rather than H as such, because R0 looks like the oldest Western subclade of R, if any one is). In turn H might be a bit older than U5 (which is the main European subclade of U, along with U4). But all very tentative because we only have so much data. My big issue with H is its massive star-like structure, which seems to indicate a sudden expansion on "virgin land", much as what should have been the case of the only other haplogroup comparable in this aspect: M. That's why I tend to associate it with Aurignacian but always very tentatively. This association would also explain its apparent disappearance in Central Europe afterwards, which may be attributed to Gravettian expansion and maybe associated with some U subclades (U5, U4 but also U*-CRS), but its persistence in the periphery (Iberia certainly but also quite possibly Eastern Europe, while HV* can partly be associated to Italy).

      My issue with (Y-DNA) E-M81 migrating from Europe is that there is no trace whatsoever. I have considered such scenario for a plausible mtDNA "couple", U6, but E1b only seems to have relatives in Africa, notably to the Northeast (Horn, Nile), so I'm rather tempted to infer an early UP East → West migration via North Africa. However such flow is at the moment unattested archaeologically ("Aurignacoid" remains exist in Cyrenaica but not further West), so either there are still discoveries to be made in this aspect or I'm very much wrong.

      "If you say Capsians wiped out previous inhabitants"...

      Obviously they did not in a direct way but the migrations and replacements associated with Capsian may have significantly altered the genetic landscape of NW Africa, partly by introducing new lineages (Y-DNA ones particularly) but maybe also by reinforcing in some areas some older lineages like E1b-M81, at the expense of European-type lineages in any case. How exactly? Why? I don't know that much nor I think anyone does but it's perfectly possible that within Iberomaurusian there was ethno-cultural substructure (some more African, some more European) and that parts of it (African ones, remotely related to the Capsian newcomers) got less damaged or even locally favored than others. Speculation on the puzzle pieces we have anyhow...

      As for J1, all I know is that NE African J1 is very diverse, almost as much as Asian one, and that it looks therefore: (1) very old in the Nile region and (2) having experienced some forth and back secondary migrations between the Nile basin and West Asia. Instead NW African one seems derived from the NE African one and having very low diversity, what demands a rather dramatic founder effect. It is also geographically associated to E1b-M78, of the same origin. Hence both for me look Capsian, not "Arab" (Arab genetic legacy seems almost invisible, unlike language and religion). I wonder why do you think that that J-L222.2 is "Arab", as I'm not familiar with the particulars.

      About the Guanches I did find very interesting that the haplogroup I disappeared in the colonial remains, what is very indicative of replacement (as I is rather low frequency in Iberia) and instead R1b increased. This seems to give some credibility to the study. I have no reason to think that the findings are spurious although I agree that caution is always required.

  13. So would you agree on a close association between Y-DNA I and mtDNA U? Because I looks younger, with a TMRCA of the surviving branches, that is I1 and I2, that could be as recent as 30kya as far as we know, although its separation from J dates from about 45kya. But if U isn't associated with I or IJ, then I can't think of any Y-DNA line that would be since we know C-V20 is brother of japanese C-M8 which has absolutely no connection to U or to any R-based clade for the matter. It is clear that C clades expanded before the colonization of eurasia by mtDNA R and, most likely, Y-DNA HIJK. On the other hand, some N(xR) clades like the mysterious X and N9/Y (very common in japan, especially among Ainus, and even more common among amuric people of Siberia) could potentially be associated with C-V20 and C-M8 respectively. Anyway, what are your thoughts on that?

    As for E-M81, what I said was that a portion of it could have left Morocco for Iberia sometime during the mesolithic or the early neolithic. Then, after centuries of mixing with Iberian females, it could have returned to the Maghreb, thus introducing the new gene pool there. Plus, for all we know, it doesn't necessarily require E-M81 or E-M78 to have wiped out any previous european population. Remember the polynesian scenario in which C-M38 melanesian males quickly absorbed incoming autronesian female lines before expanding into remote Oceania. It is thus not impossible for North Africa to have experienced a similar scenario, with E-M81 absorbing incoming female lines without much I-M170 or even R-M269 (eventhough this one looks far too Young) only to make a local maghrebi expansion. An expansion of this sort would also easily explain why E-M81 looks to have a reduced diversity. Everything is possible.

    '' I have no reason to think that the findings are spurious although I agree that caution is always required.''

    Seriously, there are many people out there who have been fooled by haplogroup misclassification, leading some to even continue to claim an asian origin for the clearly african E-M96. I think you know who I'm specifically thinking of right now. I myself got fooled multiple times before I realised the importance of self-checking Y-STR haplotypes. Some papers out there have a tremendous amount of missclassifications. It's really a shame the such errors exist because many studies don't include STR haplotypes so that we could verify the results by ourselves. Unfortunately, I would probably never be able to know whether the two reported DE* from Tibet are really so or not...guess we can still have hopes though..

    And J-L222.2 is arab, trust me there's no doubt about that. Back in the days I also believed J1 was neolithic in the Maghreb and was dissapointed when it turned out as a recent arabian line. I still don't know how it could have imposed itself to extent of representing a whopping 25% of tunisian Y-DNA but my impression so far is that it's largely confined to large urban poles, which is in line with what we know of the hilalian invasion touching mostly big coastal towns. Rural Tunisians, be them berber or arab, tend to be as high in E-M81 as even Moroccans. That is not to say some J1 isn't more ancient than that in the Magreb as (like I told you in my previous post) I've successfully identified basal J1a (mostly in Morocco) on a few occasions and also some other non arab lines(In Algeria and Tunisia). So there's still some room left for a neolithic J1 migration but the bulk of it is certainly arab.

  14. "So would you agree on a close association between Y-DNA I and mtDNA U?"

    I don't dare to have a strong opinion but it seems quite possible, at least based on the Epipaleolithic data we do have. It would be some sublineages of mtDNA U (U5, U4...), and not all U, of course.

    "Because I looks younger, with a TMRCA of the surviving branches, that is I1 and I2, that could be as recent as 30kya"

    Well, 30 Ka is old enough to be there in the Gravettian. But anyhow I take most TRMCA estimates as worthless junk. If you think you are using a particularly good one, please reference and explain why this one is supposedly any better.

    "we know C-V20 is brother of japanese C-M8 which has absolutely no connection to U"

    Lineages get mixed: people make friends (non-relative allies). This European rare C lineage obviously (to my eyes) arrived to the West in the early Upper Paleolithic and has no more relation with East Asia than Y-DNA D may have with Africa: extremely remote. If you try to read lineages as peoples, especially in a too strict way and without further support for the notion, you will end up in conceptual cul-de-sacs all the time.

    Same for mtDNA N/R. True that R seems more strictly related to Y-DNA K2 (MNOPS) but nothing impedes some N sublineages to have migrated West with a majority of R. Similarly nothing impedes some Eastern lineages like this European C clade or the Siberian pre-NO found in Ust Ishim to have arrived to the West with other now more characteristic lineages like R1, Q, IJ and G. I think in terms of complex populations with complex DNA pools, not of single lineages. Of course, in the long run, some lineages have more success while others tend to be displaced to negligible or even zero presence.

    "And J-L222.2 is arab, trust me there's no doubt about that".

    I was asking for more information precisely so I don't have to "trust you" but rather make up my mind on good reasons.

  15. But the problem with U is that it's comletely western, so it looks to have developped in the west. What F(xG) lineage developped in the west (note that I've excluded G because it broke off from F very early somewehere in the west, probably around iran, so it can't have been originally associated with U since this one is from south Asian (could even be southeast asian for all we know)? That's IJ of course. Some I and even some IJ lineages surely were dwelling in Central Asia only to become extinct sometime after P1's arrival. This is maybe why we find mtDNA U subclades all over Central Asia but not east of that zone. Mal'ta boy, with his U mtDNA, could even be the proof for those P1 guys taking local wives from the previous inhabitants.

    But what could explain this time discrepancy we see between Y-DNA I's apparent TMRCA and U's is the fact that, as we know it, the severe competion that males have to go through drove some of the lineages to extinction, thus making the TMRCA more recent than the date of formation of the clade in question (in this case, I separated from J some 45kya). For the estimations, I suggest Yfull as a reliable source as it is regularly updated, is base upon high-coverage SNP sequencing and least but no least, it takes into account ancient DNA. The estimates for the best studied clades like M269 are of course going to be much closer to the actual age, but because it is subject to much updating, the progress is very notable.

    As for C-V20, yes it could have enterred Europe with a more complex set of haplogroups, but they probably all became extinct. I say that because any other european line is far too young compared to C-V20's 45kya of TMRCA! This is very ancient considering V20 is an extremely reduced lineIts separation from japanese C-M8 should thus be of AT LEAST 50kya! So ya, it pretty much beat everyone else to Europe. It also couldn't have come with mtDNA U I guess, this is why I propose X as a good counterpart.

    And if you're still not convinced by the arab origin of J1-L222.2, I hope the following infos would make you change your mind :
    1- This particular subclade is absent from Ethiopia while very high in Arabia. Thus, it is clear that maghrebi J1 is distinct from ethiopian one.
    2- Among the most common haplogroups of the Maghreb, it is the least diverse and has very close matches with many arabian haplotypes.
    3-It can't be linked with E-M78, as E-V65 (maghrebi subclade) is far more diverse. Plus it has no east-west cline within the Maghreb, which is in complete opposition with J1.
    4-I have never seen it in around 5,000 samples from Iberia! If it was ancient, at least some of it would have crossed the strait. But of course hilalian Arabs never made it to Iberia. Plus, if we were to compare it again to E-V65, this one actually made it to Iberia and even in other areas of Europe way up to the north.

    If you're still not willing to accept it, then I don't know how I'll get you convinced...

    1. U is less completely "Western" than R0. There are several subclades of U in South Asia, while there's virtually no R0 in South Asia. There is also no identifiable unique North African R0 subhaplogroup, while U has one: U6. Of course that both are "Western" (West Eurasian) lineages, as are X, JT or N1 but that's about it. Both have sublineages that are most common in Europe (U5, U4, U2e, U8a but also HV0, H (or most H subclades), some other HV subclades) and are therefore probably coalesced in this subcontinent or almost upon arrival to it, but they also have relatives centered elsewhere (U1, U6, U7, K, R0a'b, etc.), which must be considered along with other rather non-European lineages (in origin) like JT, X, W or N1.

      "I suggest Yfull as a reliable source"...

      I don't see any info on material and methods, references or even authors. I can't judge their estimates on whatever merits they may have (it's possible that various estimates come from different sources for all I know: nothing is explained, it's an act of faith). Therefore I pass of Y-full all I can because faith and science do not make good couple.

      I'd say that both U and R0 were the main (if not the only) matrilineal contributors to Paleoeuropean genetics but that they were so within a wider West Asian or West Eurasian pool: founder effects therefore, involving branches of lineages that stem from West Asia.

    2. Erratum: last paragraph ("I'd say... West Asia") should be before the quote, as it continues the first paragraph.

    3. On J1-L222.2:

      "This particular subclade is absent from Ethiopia while very high in Arabia".

      What about Sudan and Egypt? I believe those are the key areas of Asian immigration in the Paleolithic (and later). Ethiopia or in general the Horn region does not seem so important in comparison, although the tribal province of the south does hold a very high diversity of all kind of stuff, the mainstream Semitic and Cushitic populations seem a bit less interesting.

      Capsian people did not originate in Ethiopia in any case but in Nubia.

      "It can't be linked with E-M78, as E-V65 (maghrebi subclade) is far more diverse".

      That's indeed intriguing because it's much rarer than J1.

      "I have never seen it in around 5,000 samples from Iberia! If it was ancient, at least some of it would have crossed the strait".

      Not really: E1b-M78 in Iberia is almost only the "Greek" variant E1b-V13. The only North African patrilineage to have crossed the strait in large numbers was E1b-M81. J1 does exist in low frequencies anyhow (Alonso 2008), varying from 0% to 3% (Eastern Andalusia, Southern Portugal, Valencia) and it should be the North African variant, Arab or not.

      A big problem when attributing J1 in NW Africa to "Arabs" is that Arabs should have brought other lineages, notably J2. And J2 is negligible in North Africa (3-4%). Per Abu-Amero 2009, Saudi Arabs have 40% J1 and 16% J2 (mostly J2a), so, in the neutral model, for each 8 J1 they should have brought 3 J2, almost half. What do we see in NW Africa? 17% J1 vs 3% J2, much lower J2 than expected. This departure from the neutrality model is even more dramatic in Tunisia, the area with the greatest J1 frequencies: 28% J1 vs 3% J2!

      Additionally J1 is also a very important lineage among various Berbers such as Kabyles (16% per Wikipedia), so it is clearly not at all an "Arab" marker but, at least partly, a Berber one too.

      "But of course hilalian Arabs never made it to Iberia".

      Why not? "Arabs" crossed North Africa in a whiz and landed in Iberia soon after. A few years later they had breached Basque defense lines and threatened France. It was all very fast: the conquest of Africa (Tunisia) culminated in 698 and the rest of North Africa (save Ceuta) was under Muslim control in 709. Two years later, in 711, the Muslims landed in Hispania and the defeat of the Goths was complete by 717. Then they attacked Vasconia-Aquitania for a over a decade, war that culminated in two decisive battles in 732 (Arab victory at the Garonne first and, later, defeat at the Loire, near Tours or Poitiers). My point is that between 698 (conquest of Tunis) and 717 (consolidation of Al Andalus) there are barely 20 years, therefore there should be no significant difference between the Arab genetic impact in NW Africa and Iberia: they are not two different campaigns but just an extended one.

      In any case I'd appreciate if you'd provide the sources of your info re. this lineage, so I can double-check them. Else, I'm always relying on your opinion and I have to make an act of faith.

  16. Update: a Bronze Age ritual burial of weapons not too far from Atapuerca has very clear signature of British origin (one sword is identical to a locally manufactured British one, the ritual also resembles similar British ones). Although the date may be centuries more recent than ATP9, it does reinforce the notion of more or less persistent interactions with the islands in the Bronze Age also in the Upper Ebro area and reinforces the possibility of ATP9 being partly of British ancestry.

  17. I am intesresting on E-v65 subclade of E-M78 in Tunisia, it seems to be a relique of ancient firts hamitics spread in the region , before e-m81?
    is there some clus about that?

    1. Can't say but rather leaning to the opposite direction: E1b-M81 would appear to be the oldest E1b in NW Africa of all, because of its distribution (concentrated to the West, farthest from the Nile) mostly. However it's possible that both lineages arrived simultaneously, what would correspond (my best guess) to the Capsian culture expansion, which surely corresponds to Afroasiatic (ex-Hamitic/Semitic) expansion too (ultimate seed of Berber languages, even if these appear to derive directly from a more recent source in the proto-historical period).

      Ultimately, I'd think that only archaeological and archaeogenetic research can lead us to more solid clues. For example archaeogenetics of ancient Guanches (Canarian aborigines of Berber affinity) show that they carried the following lineages:

      E1b1b1b (M81) - 8 - 26.7%
      E1b1b1a (M78) - 7 - 23.3%
      J1 (M267) - 5 - 16.7%
      R1b1b2 (M269) - 3 - 10%
      K(xP) (M9) - 3 - 10%
      I (M170) - 2 - 6.7%
      E1a (M33) - 1 - 3.3%
      P(xR1) - 1 - 3.3%


      This not only indicates that one of the most isolated offshoots of the Berber peoples had both M81 and M78, but also J1 and relatively high frequencies of presumably ancient "European" lineages such as R1b and I, which are quite rare in modern NW Africa.

      However they are not enough, nor old enough (Canary Islands were only settled some 3000 years ago) to give a reasonably complete picture of the genetic or population prehistory of Northwest Africa.

    2. PS- Obviously the presence of J1 among Guanches proves that J1 in NW Africa is not Arabic but much older, surely arrived with the Capsian culture as well some 10 Ka BP., from an ultimate origin in Egypt and Nubia.

  18. Thank you for your answer,
    i agree for the very west of North Africa(Morocco and west Algeria), the farthest of the Nile, but for the oriental Maghreb like east costal Tunisia , i think E-M78 subclades like E-V65 are much older in these places , like costal libya to Alexandria too ,
    the TMRca of E-M81 is about 2100 years actually with quite severals samples ..
    For J1 its quite interesting , but it true that Some J1 are ancient in NA as the Guanches samples proove.
    E-V65 is found in Sardinia , ,Sicily,Andalusia ,also in Greece(Libya Cyneraic colonies) ),and in the Basque province of Araba..
    I think most of them are Carthaninians signals.
    What is your opinion?

    I Know very the french Basque country as i was born in Bayonne, but i am originate from Tunisia.
    Best regards.


    1. "E-V65 is found in Sardinia , ,Sicily,Andalusia ,also in Greece(Libya Cyneraic colonies) ),and in the Basque province of Araba..
      I think most of them are Carthaninians signals."

      Actually it is, at least largely, Cardium Pottery Neolithic signal via the Aegean. A Phoenician signal should be rich in both J1 and J2 lineages (not one or the other separately: those are older signals, early Holocene ones), and not so exceedingly rich in E1b. Additionally Phoenicians only occupied a few coastal outposts and never really settled the interior, except maybe in Tunisia. As we know from history, the interior of NW Africa, and even more of Iberia, was native land, and, if you ask me, it is almost impossible to detect any clear Phoenician signature in Iberia and only a very weak Moorish one in East Andalusia (only place in Iberia where J1 is significant: 3%). Similarly it's difficult to detect a strong Phoenician or Arabic signature in NW Africa: although there is of course some lesser influence, the bulk of the ancestry is by large Berber.

      "I Know very the french Basque country as i was born in Bayonne, but i am originate from Tunisia."

      If you were born in Baiona, you cannot be "from Tunisia", your parents maybe but you are by right of birth as Basque as I am: no kidding. Another thing is that you may live elsewhere and have been uprooted but otherwise it is as I say, Jamel.

  19. There are proof of E-M78 and its subclades holders in the Capsian culture?
    If yes have you some references?
    Thank's again.

    1. AFAIK there is no ancient North African Y-DNA (and almost no mtDNA either, let alone autosomal DNA) researched to this date. So the answer is that there is no DIRECT proof, only rational inference based on the fact that NW Africa had no significant immigration after Capsian other than the Phoenician coastal colonies and the Arab-Muslim conquest, neither of which fit the profile (they should carry much higher J2 and other West Asian lineages that are almost absent).

      So E-M78 and J1, whose spread pattern overlaps A LOT should be associated with prehistorical migrations and the one that fits best is the Afroasiatic expansion(s) of the early Holocene, which in NW Africa correspond to Capsian culture (first Epipaleolithic, then Neolithic). There may be subtle details to this (for example I wonder if Eastern Sudanic/Nubian languages were also implicated, or was there a secondary Neolithic wave too subtle to identify as distinct in the archaeological record?) but it should fit in the general picture.

      E-M81 may also have originate there and have experienced a particular founder effect or, alternatively, it could represent some unknown older wave.

    2. Of course a part of me is very proud of its Basque origins.
      My mother grew near Saint Jean pied de Port (Garazi)..
      Obviously, Sousse (Founded by Phoenicians) Shares 9% of J2 with 5 % of E-V65, it could be a remanence of the Liby-Phenicians who were particulary known in this country as part of the Punic Army..
      But as you said sadly we don't have any ancient local Ydna.
      Milechker Maju.

    3. Effectively, the 9% J2 in Sousse (Arabic and Berber Susa, ancient Hadrumetum) should be interpreted as a strong Phoenician ancestry signal, no doubt.

      Ondo pasa, Jamel.

  20. And what about the E-v65s in Sousse,
    Older than the J2s??


    1. Not too sure what difference you want to make underlining V65, I've generally treated all M78 as one unit, excepted sometimes V13. So IMO it's Afroasiatic or Capsian stuff arrived from Nubia and Egypt mostly c. 12 Ka BP.

      If you search for V65 you get results for M78 and V68 (its upstream precursors). This is the kind of map you find:


      And it is the kind of scheme I have in mind grosso modo, so V65 would just be in appearance the main Western (NW African) branch of M78.

      In this sense I need to correct my previous assessment for V65 in Iberia, because it's probably a North African marker and not a Neolithic one from the Aegean. I was thinking in V68 surely: the Aegean-derived lineages like V13 are another branch of the same early Holocene expansion from the Nile, which went to NW Africa (crossing the Western Med at some points) but also to the Levant and, from there, to the Aegean and Europe (this leg within the Neolithic).

  21. Very right Maju the Arabic Susah and the ancient Hadrumetum (Romans), Hadrim (Punic), all matche with actual Sousse.
    So How interpret the E-V65 in Sousse in your opinion ?

    1. "Hadrim (Punic)"...

      I didn't know that one: sounds pretty cool! We get to learn Latin versions of Phoenician names but I'm sure that in most cases the original words are much better (or at least shorter).

  22. This comment has been removed by a blog administrator.

    1. ^^ This character was banned from commenting. The fact that he insists after the ban underlines his trollish nature.

  23. I want to underline V65 because many peoples(Not the most) tend to assimilate it to E-M81 as a "Magrebi subclade", i think it's not exact and as a E-m78 subclade it has a different origine and history in North Africa.
    With all the respect for E-m81 people , another great lineage
    in the area..

    1. Yes, it seems so. You are one step ahead of me in being well aware of that.

      However it is also true that V65, as part of M78, it is more "neatly" linked to other areas and related branches via Egypt and Nubia, while M81 is more isolated, although it is clearly a relative but only "ultimately so", and therefore it is a bit more mysterious. IMO V65 or in general M78 in NW Africa should be associated to (most) J1 arriving from the same origin at the Nile, on the other hand M81 could be also involved in that (being a "frontier" founder effect) or belong to an entirely different, surely older, migration (AFAIK not documented archaeologically as of today).

      An issue of M81 is that it appears very closely related to mtDNA U6, notably in Iberia, where they have almost exactly the same distribution and percentages in a very striking way. While E1b-M81 is certainly an African lineage, mtDNA U6 is an Eurasian derived lineage and the only clear "africanness" it has takes off from Morocco and not the Nile at all -- and it may even have originated in Europe with Iberia playing a role as "trampoline" in the Solutrean/Oranian, see:

      The best solution to this puzzle is I guess to consider the association of M81 and mtDNA U6 as a mirage but is it?

  24. Thank you Maju, very interesting.
    It's true that i have a special interest on V65 E-m78 subclade as it is my subclade, and i want to precise his history the town of my father Sousse,
    Some people associate it (E-v65) to the Banu Hilal Arab inavsion, like in Morocco (Is it exact?), Others specialists see it as a marker of ancient Libyan tribes , we have a great Libyco/punic history in Sousse (Hadrumetum),
    Me ,i would like to relie that past (Libyco/punic) to E-V65..
    Perhaps i take the false way..
    Who knows?

    1. I don't see any sense in relating V65 with the Arab invasion, it is very apparent that it is an early Afroasiatic (proto-Berber, Capsian) marker. Why? Because there is no V65 in Arabia, as simple as that, and also because we can track the arrival directly from the Nile at dates comparable to the arrival of its M78 relatives to the Levant (and later to Europe with the Neolithic peoples). This M78 expansion is very clearly part of the Northern Afroasiatic expansion (regardless that it might also have included other ethno-linguistic groups originally, such as Eastern Sudanic), expansion that in NW Africa is defined by Capsian culture.

  25. thank's Maju ,
    your knowledge impose the respect.


    1. Thanks to you. In the end everyone must make up his/her mind, I just feel the need to argue in favor of what I find most compelling.

  26. Hi Maju,
    I have a question
    How do you explain that in Morocco where the Capsian culture dind't enter physically as all specialist confirm , there is one of the higher level of V65 ?Best wishes. Jamel.

    1. I thought Capsian replaced Iberomaurusian in all the region, in any case Iberomaurusian (Oranian) does not extend beyond the younger Dryas, so something else must have replaced it, and I've always assumed that Capsian and Neolithic of Capsian tradition were the replacements. However I reckon I may have been making some wrong assumptions.

      If you look online, there is no other referenced culture for this period and region, or at least I cannot find it. Wikipedia for example says of the "Mesolithic" in Morocco:

      In Mesolithic times, between 20,000 and 5000 years ago, the geography of Morocco resembled a savanna more than the present arid landscape.[13] While little is known of settlements in Morocco during that period, excavations elsewhere in the Maghreb region have suggested an abundance of game and forests that would have been hospitable to Mesolithic hunters and gatherers, such as those of the Capsian culture.[14]

      In the Neolithic they mention Cardium Pottery (South European Neolithic of Thessalian-Epirote roots) BUT I know from other sources than only Northernmost Morocco was affected by it, possibly implying a "bounce" of the migrants in their way to Portugal and SE Spain, where they might have acquired NW African lineages.

      So you tell me if you know something I don't, please.

  27. It's true that it 's not because we din't find Typical Capsian sites in Morocco ,means that Capsians didn't reach the place;
    For exampel we didn't find any signifiant Capsian site in west Libya , only in Cyrenaic(east Libya), but we know that the region was occuped by the Capsians groups .
    In any case The Guanche Mummies of Canary islands attest the presence of E-M78 in Morocco a long time before the medieval period..?


    1. "In any case The Guanche Mummies of Canary islands attest the presence of E-M78 in Morocco a long time before the medieval period..?"

      Yes. And not just this lineage but also J1, which has been sometimes wrongly attributed to Arabs. Canary Islands were settled c. 3000 years ago (maybe a bit earlier but unattested), so they should reflect to some extent the genetic makeup of Western Berbers of proto-historical times. We must consider that there would be founder effects involved in the settlement but, in any case, it's clear that their lineages, particularly those of NW African affinity like E1b and J1, must have come from pre-Phoenician, pre-Roman and pre-Muslim NW Africa. The fact that we don't see any J2 (common not only among Semites but also in Italy, Spain, etc.) excludes settlement by the aforementioned civilizations.

  28. Maju, have you heard about Himilco voyage, he was a Carthaginian sailor who reach Ireland and England 2500 years ago?
    His travel reach from carthage, Cadix passed and stoped near actual Galicia and Cantabria ..

  29. We know that Galicia and Cantabria has the higher level of E-M78 (V65) in North Europe..

    1. Just because some guy traveled along the coast to sell and buy it's not going to change the genetic landscape. Some of the areas with greatest North African signature in Iberia are inland, for example the Maragatos of the mountains of León, there's no correlation with the coast or with anything that suggests Phoenicians.

      Ancient Phoenicians should be identified primarily by the presence of both J1 and J2, which are in relative equilibrium in the Levant (and most of West Asia). Where we see a clear imbalance between both lineages (lots of J1 and almost no J2 or vice versa) the signature must be much much older, from a time when those two lineages were more strictly segregated geographically, for example the Mesolithic and Neolithic. So in Africa Semitic signature in general should be associated with Y-DNA J2, while in Europe, where J2 is common, it should be associated with J1 instead (although it can also be Moorish but from a more recent time than the E1b Atlantic signature, for example in Eastern Andalusia, which roughly corresponds to the Kingdom of Granada and where J1 reaches an unusual 3%, not a lot but much more than elsewhere).

      BTW, Hanno reached Cameroon and other Phoenicians seemingly circumnavigated Africa for the Egyptian pharaoh. However their genetic legacy in all those areas is very apparently zero.

  30. Exact Maju, but don't forget that Carthaginians were different from the primary Phoenicians, indeed since the begining Carthage population known as the Punic civilisation was a very mixed population especially with the Libycs autochtonus people, wich at my sense ruled with e-v65 and e-m81 and J1? clades, the True Phoenicians were a minority, but about Galicia and Cantabria you're probably right, althought in England ,in Cornwall department ,a carthaginian was their first king so it seems that they settled in place for along time, not only for Selling..

    1. I agree re. admixture with locals but do not forget that:

      1. Gadir and the South Iberian colonies (Malaka, Abdera, etc.) were nearly as old as Carthage itself, Gadir particularly is extremely old, probably the second Phoenician colony after Carthage itself (those guys were aiming at Iberian and Atlantic metal wealth, previously traded by Mycenaean Greeks probably, and even earlier by ancient Cypriots, they knew well where they aimed at). So Iberian Phoenicians should have their own specific admixture, distinct from that of Lybo-Phoenicians.

      2. In any case we should see the already mentioned "balance" of J1 and J2, which is hard to find. You have already mentioned the case of Susa or Hadrim, where the Phoenician signature is strikingly apparent in high frequencies of J2, otherwise unusual in the region. This is even truer for Europe, where we should find a J1 legacy, if not from Lebanon, at least from North Africa.

      3. You're kind of obsessed with E1b-V65, I can only imagine that you have been tested for Y-DNA and produced that lineage yourself, and that, for your own personal reasons, maybe of "prestige", you seem to WANT it to be Phoenician rather than Berber. I have to insist that you are wrong on that interpretation for all the reasons mentioned in this discussion.

      "in England ,in Cornwall department ,a carthaginian was their first king"

      That's mythology and one I've never heard of, BTW. It is true that Phoenicians (at least as likely to be from Gadir as from Carthage) traded with Cornwall, but they never seem to have settled anywhere beyond Gadir, at least not in substantial numbers (there's some speculation about some towns in Portugal but nothing solid).

      Instead there is one county in Northern Wales where E1b-M81 is concentrated at frequencies much higher than anywhere I know in Iberia, something like 30% or 35%. It's probable it represents some sort of miner colony, be it slaves or freemen, with a Moroccan origin, and also because it is an isolated enclave, but it's not likely to be Phoenician but rather something older, surely from the Atlantic Bronze Age period.

  31. Maju I have already mentioned you that E-V65 was my subclade,I want it rather Carthaginian than phoenician, it's rather you that speak about Phoenicians , i always use the term of Carthaginian wich is different as you recognise before, I am very proud about V65 lineage as it is autochton , i read lot of books about Carthage and punic civilisation , and i can assure you that Hannibal himself Was surely more african than phoenician as he was Carthaginian.. The Libycopunics especially in Byzacene (Sousse)were known as mixed between LiBycs and Phoenicians since the beginning, its not a question of prestige it's simply a part of the history of the region, Berber is a new term not very adapted , i rather like the term of Libycs people wich is used by historians.

    1. "I have already mentioned you that E-V65 was my subclade"

      Sorry I forgot altogether or did not pay enough attention, my bad. However it became quite apparent as you seem to try to force on me your favorite interpretations of its origins.

      "I want it rather Carthaginian than phoenician, it's rather you that speak about Phoenicians"

      Well, everybody speaks of Phoenicians at least until the Carthaginian "empire" coalesces. Originally the various cities were autonomous and, if anything, dependent on Tyre.

      In Iberia at least, the cultural and trader talk is always about "Phoenicians", with "Carthage" only becoming relevant after the First Punic War (although it probably held sway over the Phoenician colonies earlier too), when the Barcids actively expand inland and found a couple of cities of their own (Cartagena and Mahón).

      Probably the Carthaginian "empire" evolved in the context of conflict with Greeks, but again we don't talk of Syracusans or Massilians most of the time, rather just Greeks. Surely Sycilian and Hesperian Greeks were surely not exactly the same as those from the motherland after a time of dealing and marrying with the locals, but still their ethnic identity was quite strictly and distinctly Greek. Same for the Phoenicians, who never became Numidians or Turdetani or anything like that.

      Anyhow, the issue is if Carthage as such had any colonial influence in Western Iberia other than Gadir and AFAIK the answer is nope, barring a few doubtful outposts maybe. Instead the area was until the 2nd Punic War held by Celtic and pre-Celtic tribes. Phoenician legacy in the archaeological record is AFAIK quite straightforward and recognizable, just as Greek one is. Phoenicians had a significant and growing influence inland but primarily not in the areas we are considering here, rather in the South and East of Iberia, much as Muslims would have in a later time. They were a Mediterranean culture after all.

      There was the famed Silver Route along the Eastern Interior but it went in any case through Celtic and Lusitanian lands. At no time those areas were under direct Phoenician control. In fact Phoenicians only controlled their own coastal colonies until the Barcid expansion and is even doubtful that they managed to hold a minimal hinterland around them.

      It's just like (Massilian) Greeks: they had a number of settlements along the Occitan and Catalan coasts and they did make ethno-political impacts surely, but not directly, instead they probably just tipped the scales of pre-existent inter-ethnic conflicts in favor of Iberians and against Celts.

      So, regardless of the depth of the Lybo-Phoenician integration in Africa (Tunisia), the impact of that phenomenon in Iberia should be almost nil and, if anything, concentrated in the towns founded by the Barcas: Cartagena and Mahón, which again are in the SE of the Peninsula (the former) and in a small island also to the East (the latter).

      The Ibizan genetic signature (which might be attributed to Carthaginians) is indeed somewhat peculiar but not at all what you imagine:

      Ibiza has no detected J1 although it has a local peak for E1b (M-78 and M-123, no M81 though). But the striking features are highest frecuencies in Iberia for K(xP), surely T, and G. These are surely founder effects and, while they may (or not) have been mediated by the Carthaginian foundation, they do not say "Africa" at all, nor "Lebanon" either.

      As for using the terms Berber, Libu/Lybian/Lybics, Amazigh, Numidian, Maurus/Mauretanian/Moor(-ish), I think they are all mutually interchangeable, which at most a very fine detail of shades: they all refer to the native peoples of NW Africa.

  32. It's somewhat like Aderahman III who became surely more Basque than Arab.

    1. Abderhaman III was genetically 3/4 Basque for all we know but culturally, linguistically, ethnically, identitarily... he was 99% or even 100% Andalusi Arab. And, unless some unknown cuckolding took place in his Umayyad lineage, his Y-DNA must have been Arab, probably J1, J2 or maybe some other and certainly not R1b nor I, the Y-DNA lineages that make up 99% of the Basque genetic pool.

      It's not so easy to have a purely neutral melting pot. Even in Latin America, say Mexico, people may be half or more native in terms autosomal, but in Y-DNA they are almost invariably Spaniards, and also they speak the Spanish language and their main cultural element is Spanish, religion included. So in the Carthaginian state, I would expect the citizens and very especially the aristocrats, to have very high frequency of "true Phoenician" Y-DNA, what means J2 as most visible signature in Africa and J1 in Europe. It'd be striking to imagine a Phoenician colony that would not have a sizable fraction of those markers, which together make up at least 50% of the Levant's genetic pool.

      So in the example of Ibiza that I was just talking about: we just do not see the Phoenicians, because we do not see the J1. It probably was there back in the day but the colony maybe was too small to compete with the aboriginal genetics, because it was spared from destruction by the Romans.

      (Note: its supposed to have been resettled from Catalonia upon "reconquest" but the genetic signature is very different from Catalan one, so probably not quite what they say, as so often with "repopulation" stories).

  33. This comment has been removed by the author.

  34. This comment has been removed by the author.

  35. Sorry:
    Tite-Live décrit l' infanterie libyphénicienne comme « mixtum Punicum Afris genus », prenant probablement Polybe comme source. Elle constituait la force la plus loyale et capable de l'armée carthaginoise. Par cette expression de Tite-Live, il faut voir « une race punique mêlée d'africains». Les Libyphéniciens sont décrits comme une race mixte, à moitié punique et à moitié africain. Un mélange de sang africain et colonial, qu'ils occupaient dans les colonies de Carthage en Afrique, tels que Utique ou Hadrumentum.

    1. That's probably much like Mexican "mestizos" mutatis mutandi. If they were such thing, they probably had most usually Phoenician (Levantine) Y-DNA. After all, in colonial processes, which take place within patriarchal societies, admixture between native men and colonist women is very rare, while the opposite is common. It gets more complicated as generations pile up but the primary trend should be that.

      In any case, J1 is a very common lineage in Tunisia, it was there certainly before Phoenicians and it should be apparent in any Phoenician/Carthaginian colony in Europe (where J1 is at best very rare).

      What's the frequency of J1 (J(xJ2) in Iberia? Per Adams 2008 it peaks (so to say) in East Andalusia, South Portugal and Valencia with 3%, followed by Castile-La Mancha, North Portugal and Majorca with 2%. The rest have 0-1%. This is a very dim legacy attributable to either Phoenicians/Carthaginians or Muslim Arabs/Moors. If we imagine it as Levantine (pure Phoenician) it'd represent about 1/3 or 1/4 of their overall genetic legacy, so at best they could have made an impact of 12% Y-DNA in certain areas, if we imagine it as Tunisian (Lybophoenician) it's similar (a bit lower probably, so greater impact in theory but still small). But most commonly it's considered to be "Moorish" (very little would be "true Arab"). If we use the Moroccan or Algerian frequencies (lower than Tunisian ones: 8% and 15% respectively), it could be multiplied by as much as x10 maybe so up to 30% in the peak regions. But in all the African origin scenarios we must consider also the E1b frequencies, which usually don't make a good fit at all. So it is more complicated and therefore historical North African legacy must be smaller, while pre-historical one (without J1) should be understood as much more important, particularly in areas like Asturias, with 15% E1b (10% E1b-M35*!) and zero J1.

      Otherwise you'd have to imagine very specific founder effects (one or very few settlers) with a surprisingly massive impact. And that doesn't make sense.

  36. Thanks for these explananation but Maju it seems that it's you that are quite obssessed by J1.
    i just wanted you to you iamgine that E-M78 could have been a consequent lineage in the Punic population(Carthage), cf my reference about the libo-Phonicians.. The tradition want to attibute the berber marker on E-M81, E-M78 i think has another prehsitory and history, more related to nile valley a ithink more than E-M81, i don't definite it as a berber marker even V65, more libyc , berber denomination is quite recent.. as you said V65 is juts the West branch of E-M78 subclades.. and it is probably linked to the Capsian culture..
    Best wishes.

    1. I'm sorta "obsessed" with J1 because it is no doubt a key marker of recent North African or West Asian ancestry in Europe. For example, Jews, who have "recent" West Asian ancestry do carry this lineage in significant amounts (as expected although lower than if they would come from Palestine, they rather seem "Cypriot" based on that and other data). That's at least 1500 years ago, probably more, so almost overlaps (with lesser leniency) with the Phoenician colonization phenomenon. So we should expect roughly the same kind of pattern, even if it is from Carthage (J1 is a major lineage in Tunisia and, per the Guanche mummies and common sense, is clearly pre-Arab and even pre-Phoenician in the region).

      "i just wanted you to you iamgine that E-M78 could have been a consequent lineage in the Punic population(Carthage)"

      Sure, it was. I do not question that. What I say is that where E-M78 is found without significant J1, it is almost certainly from a much older origin - maybe Neolithic, maybe Chalcolithic or maybe Paleolithic but for sure not Phoenician or Carthaginian.

      I fully agree that E1b-M78 is today very tightly related to J1 in NW Africa and that's why I think that they probably arrived with the same wave, surely the Capsian early Afroasiatic peoples, leading to Berbers later on.

      "berber denomination is quite recent.. "

      Well, yes, but it indicates the ethno-linguistic descendants of those ancient North African peoples sometimes known as Libyans but also by other names. Arab North Africans are also mostly their descendants, with just a stroke of greater cosmopolitanism in most cases, but ethno-linguistically they represent another layer, mostly cultural of course, not so important genetically.

      North of the Strait there is also reluctance to understand that modern Iberians (and others) are in essence descendants of ancient Iberians and therefore akin to Basques. Because ethno-cultural and linguistic "brainwashing" they tend to think instead in terms of Romans, whose genetic legacy in the peninsula is surely very small however.

      The same happens, mutatis mutandi, south of the Strait: the Arab ethno-linguistic, cultural and religious legacy blocks most from perceiving themselves as Berbers (Libyans, whatever) and there is a resistance to that identification. However from the rational and genetic point of view, there is no doubt: they are essentially the same people and "arabness" is just a cultural veil that hides the deep reality.

      "i don't definite it as a berber marker"

      There's no such thing as a Berber or Libic marker: there are just North African markers, which correspond to Berber or Libic ancient roots.

  37. Ok Maju, I understand
    Iberians are a very old substrat, it's quite interesting, sorry to make a fixation on Carthage, but i know that iberians were a substential part of Cathaginans infantry and mercenairs.. and lots of Cathaginians chefs married iberians princess, Sophonisbe fort exemple was from an iberian tribe and was Hannibal wife, as the tradition tell..It's quite strange that it isn't very related on Spain History, severals from the oldest town in Spain come from a Carthaginian settlement,, Cathagene,Malaga.. As in lots of others part of the ancient world Rome passed out others older civilisations, as the Iberian civilisation itselve.
    If I understand Basques peoples are related to ancient Iberians ?

    1. Of course that Hannibal the Great used Iberian, Celtic and also Numidian troops. In fact one of the differential aspects of the Roman and the Carthaginian armies were that the first one was a "national" army with very limited use of mercenaries and allies, while the Carthaginian one was primarily mercenary, maybe owing to the mercantile "essence" of the Phoenician republics -- the Romans were definitely only good at one thing: war, they were extremely militaristic and Mars, god of war, was one of their main patrons, unlike in ancient Greece where it was rather shunned.

      But if you study the history of the Barcas or Barcids, yo should realize that the expansion in Iberia that they championed, only happened after the First Punic War, as they felt they needed a new power base to replace Sicily and Sardinia, and Iberia was obviously ideal for that (large, wealthy in natural resources and very divided in city states and tribes, that were comparatively somewhat less advanced than the Carthaginians, Romans and Greeks).

      So basically, Hannibal and his father Hamilcar, who can be considered the first rulers of Spain (not all but about half maybe), built that colonial empire out of scratch, much like the British built their colonial empire in India after they lost most of North America to the newly independent USA. It did not exist before the end of the First Punic War (excepted a set of colonies in coastal Andalusia) nor, obviously, after the end of the Second one. It was a flower of a day, so to say, although indeed most important in Hannibal's epic campaign.

    2. It is true that some of the most important ancient towns in Iberia were Phoenician colonies (Cádiz, Málaga, Almería, Cartagena, Ibiza...) but I cannot agree that all the continuously inhabited cities from antiquity are Phoenician foundations. Nor that they are "Carthaginian" in most cases: in fact, most of those cities were founded by Tyre, not Carthage. Carthage would only raise to preeminence later and in a gradual form. Only Cartagena and Mahón are Barcid (i.e. Carthaginian) foundations AFAIK. Gadir, Malaka, Abdera, etc. are much much older and in some cases (Gadir) as old as Carthage. Basically the Tyrians founded Carthage and Gadir in the same period. This is part of the proto-history of the Mediterranean, which is very interesting.

      To make a long story short, the first large walled towns (cities) in Iberia seem native (or maybe partly influenced by Cyprus???) They appear in the Chalcolithic, c. 2600 BCE and seem to act as autonomous powers and centers of civilizations, the oldest ones in non-Balcanic Europe. Two are particularly notable:

      1. In the west, north of modern Lisbon, there was a major city, Castro do Zambujal, part of a wider culture known as Vila Nova de Sao Pedro (VSNP for short) which also had several other smaller towns in the Lisbon-Setubal region. It was very influential in the context of Megalithism and later Bell Beaker (both phenomena may have originated in Portugal sequentially) and lasted for a long time until the very end of the Bronze Age, when the silting of the 10km "marine branch" that linked it to the Ocean suddenly silted (tsunami?) and the civilization ended. I suspect it is at the origin of what Plato tells about Atlantis, some details fitting like glove in hand, but the site is not sufficiently researched yet.

      2. In the SE first appears the city of Los Millares, quite well known, and is later replaced by a neighbor: El Argar, which controlled all Almería and Murcia provinces and surely also indirectly the smaller towns of the Valencian Country and the mottle-bailey castles of La Mancha (motillas), so similar to Sardinian nuraghi. This culture is almost certainly the precursor of later Iberians or Eastern Iberians. They traded with Mycenaean Greece (where some Greek legends may be traced to: Hercules and the Hesperides, Hercules and Geryones) and adopted Greek burial practices in the second or B phase (in turn Greeks imported tholos burial from Iberia). This relation was surely based on the trade of tin, which mostly had to be brought from the more barbaric NW Iberia (Galicia) and maybe other Atlantic locations such as Cornwall or Brittany. Tin was in the Bronze Age like oil is for us: the most strategic mineral and it was only abundant in Atlantic Europe, so trade and maybe war with Iberia became suddenly very important and the Mycenaean Greeks seemingly controlled it, although it is unclear what role Sardinia may have played (we see "Shardana" with horned helmets as those of Nuraghic Sardinia acting as mercenaries and pirates in the Eastern Med, I have some theories but I'll spare you that part).


    3. ...

      With the Bronze Age collapse, which happens just after the destruction of Troy and Ugarit (among other places) by the Myceanean Greeks and allies (last "Sea Peoples" coalition), El Argar also collapses as (probably) unified kingdom and becomes a set of independent city states, known as Post-Argaric culture. As I mentioned, these are at the roots of historical Iberians and some of the historical towns of the Punic and Roman period are rooted back then.

      So Iberia was no barbaric land, even if it was clearly a step behind the Eastern Mediterranean in development (too far away): there were cities or towns all around, kingdoms, lords, warriors and surely priests what not. Quite intriguingly the Iberian semi-syllabary script is not derived from the Phoenician or any other alphabet but might be loosely related to Eteocypriot. In any case it seems it began developing locally long before the Phoencians arrived, even if most of the preserved texts are from the few centuries before Roman conquest, when they became popular in monuments, usually slabs (tombstones or legal admonitions?)

      After the Bronze Age collapse the Phoenicians ("evolved Northern Canaanites") took over the Mediterranean trade very quickly, founding in quick succession the colonies of Carthage, Gadir and others. They had the insight and the skill to be interested in the Far West, where there was clearly a power vacuum and lots of riches. In Asia they soon lost their independence to the new empires (Assyria, Babylon, Persia and Macedon) so the colony of Carthage took over the role initially held by Tyre. However that does not mean that automatically Carthage controlled every other Phoenician colony, but it grew gradually into a very clear hegemon.

      As we know from the legend of Hercules, in ancient times the route to the Far West (Hesperides) was via North Africa and that was the route that the Phoenicians adopted. However some time later the Greeks challenged them with the founding of Marssalia (Marseilles) and the opening of an alternative route via Italy. Massalia had a major influence surely in Iberian ethno-politics and seemingly sided with ethnic Iberians versus the Celts that had been pouring since the Bronze Age collapse (Urnfields since c. 1300 BCE). This made Iberian Celts separated from continental ones, reason why they did not partake in later La Tène culture (which also had city states, called oppidae by the Romans) nor adopted druidism.

      So it's a much more complex process than just Phoenicians brought civilization to the West. There was civilization before, there was civilization outside of Phoenician control or influence, the Phoenicians of Gadir actually claimed to have destroyed the semi-mythical city of Tharsis or Tartessos, not yet located, etc.

      The Phoenicians did bring steel (Basque word for iron is "burdin" of likely Semitic origin) and potter's wheel, among other advances and did have a very clear cultural influence in the so called Tartessian-Orientalizing culture of Southern Iberia, which is nevertheless native at the core.

    4. "If I understand Basques peoples are related to ancient Iberians ?"

      Seems so, although it is a complex matter. Linguistically it is clear nowadays, I understand, that Iberian and Basque are related and also to Paleo-Sardinian, Paleo-Corsican, etc. Also there's a vast array of evidence pointing to a much wider "Vasconic" substrate in most of Europe. So my interpretation, combined with genetics, is that Vasconic is the linguistic family of European mainline Neolithic, rooted in the Aegean, which was later replaced (in several waves) by Indoeuropean languages coming ultimately from Russia but with a secondary center in Central Europe.

      It's also possible that, in the context of the ethnic wars against Celts, the more civilized Iberians influenced proto-Basque peoples altering their language. In any case, there is no clear "Basque" ethnic area until the Celtic invasions of the late Bronze and Iron Ages, which cut the pre-IE Vasconic countries in three: the Ligurians around the Western Alps, the Iberian in SE Iberia and the proto-Basques around the Bay of Biscay and the West-Central Pyrenees. Within Massilian (Greek) influence this separation is mostly "healed" but probably there was some distinction forming between Basques and Iberians, which represent two quite different backgrounds within that Neolithic heritage shared context. Historians aren't sure if for example the Ilergetes (the people of Ilerda, modern Lérida or Lleida) were "Basque" or "Iberian", the distinction is not clear and Caesar for instance was unaware of it, saying that the Aquitani were "Iberian", when we know for a fact that they were precursors of modern Basques.

    5. PS- You may like to learn that a curious Basque expresion for "very far away" is "Utikan", i.e. "in Utica", one of the oldest Phoenician colonies in what is now Tunisia. :)

  38. Interesting.
    Sidon(Alyssa ) when she escaped from Tyr and the violence of her brother Pygmalion, stopped firstly in Cyprus with her float, the legend talk about an virgin girls rapt on Cyprus, before they accost in the bay of Carthage to settle the new city.. Cyprus has certainly a cultural and genetical influence on Phoenicians and Carthaginians colonies.;
    Maju, overall, we couln't underestimate the level and the influence of the new city of Carthage wich surpassed the ancients cities of Canaan(Byblos, Tyr ..) in severals domains, she became indeed the Capital of a new empire in the west, even if there were others old colonies , even older (Utique, Cadix, Hadrim are older than Carthage), the legend place the foundation of Carthage 814 BCE, not very far of the the foundation of Rome..

    1. I don't mean to trivialize the role of Carthage, just that you cannot trivialize Tyre either, which never ceased to exist and was commercially important until well into the Roman period. Also Tyre set up the religious and cultural aspects of Carthage and other cities that it founded. I guess it's a matter of emphasis: I'm used to consider the ethno-culture as "Phoenician" and only the polity that was rival to Rome as "Carthaginian", even if both terms may be often (but not always) interchangeable.

      Anyhow, you mention Dido by the name of "Sidon". Why? Is that the Arab name? Because Sidon was actually another quite active Phoenician city, which also founded a number of colonies (I've read that Tyrians worshiped primarily Melkart, while Sidonians favored Astarte/Ishtar instead) BUT Sidon is not the mother city of Carthage, it was Tyre and Dido (Elissa) herself was from Tyre, according to the legend, not from Sidon.

  39. Maju, Do you know the history of Arich, a 2500 years old man from a tombal sarcophage taken out the coline of Byrsa Carthage, it seems that he was from the nobless of Carthage, his Mitochondrial Haplogroup has been tested and it is from an old European ancestry, especially from ancient iberian lineage..

    1. No, that's new for me. Do you have a link?

      How do they know his name? Is it "-ch" as in French (sh) or as in Spanish and (usually) English? I ask because Aritz or Haritz is a common Basque name that means "the oak", cf. Eneko Aritza (Íñigo Arista in Spanish) - speculating on almost nothing here.

    2. *Erratum: "haritz" means "oak", "haritza" would be "the oak".

    3. Aniway, I guess the name could well be Phoenician Or Libic (Berber). There's a lot of NW African mtDNA that seems original from SW Europe, notably all the H haplogroups (H1, H3, H4 and H7) and plausibly also V (but this one could also be Neolithic from an unknown source). This issue was settled by a couple of papers back in 2009, as discussed HERE, the most relevant papers for North Africa are Cherni's and Ennafaa's.

  40. Maju here is a link :


    1. Thank you very much. U5b2c1 s a relatively rare sublineage that, according to Eupedia, is "found in the British Isles, France, Germany and Mesolithic Spain", by which they mean La Braña 1. This is surely the strongest reason to claim it as "old European ancestry", which is probably true. Another thing is when the lineage migrated to Africa, if in the Paleolithic (Oranian), in the Chalcolithic (Megalithism) or with that particular Phoenician individual, whose mother might be from Iberia herself (he might be from Gadir or some other Iberian colony or even an assimilated native).

      It's maybe of interest to mention that U5b2 is not a modern Basque lineage (U5b1 is instead) but it does exist in nearby areas (Gascony, Central France, Pasiegos), see (among other anecdotal online info):

    2. Notice anyhow that U5b2 (in many cases never tested for downstream markers) is found at low but significant frequencies in NW Africa (2,5-5% range). So it may well be a local lineage, although more research would be needed to confirm it.


      Also, according to FTDNA, "There are five U5b1c1 test results with ancestry in Italy, Spain, Scotland and the UK".


      So it should be a West European (and Italian?) Paleolithic lineage and, as such, it could have jumped to NW Africa at any time since the Last Glacial Maximum. I wouldn't be surprised if it is a local North African lineage, really, although a more immediate Iberian origin can't be discarded either.

  41. :

  42. Ariche is a new name given to him , I think..

  43. Arich means "the well loved" in ancient punic ..

  44. This comment has been removed by the author.

  45. This comment has been removed by the author.

  46. This comment has been removed by the author.

  47. Hi Maju, something interesting the discovery of a prehistoric site at Nefta south Tunisia, it is estimate around 100000 YBP, wich could make one of the oldest site of modern Humans, and help us to understand first movements of modern humans..

  48. Replies
    1. No big deal, I got one course of French long ago and anyhow it's a Vulgar Latin dialect, just slightly more diverged from Spanish or Italian as would be Moroccan Arabic from Iraqi Arabic, for instance, so I can read it with some difficulty (also almost half of English vocabulary is French-derived).

      The question is that the article is not too informative, the findings are limited to stone tools and the date is not so old (c. 92 Ka BP). While it's not specified I bet it is an Aterian site or, in any case, a site belonging to the wider "out of Africa" wave of expansion, which happened (based on archaeology) between c. 125 and 90 Ka BP (a pluvial period in which the Saharan and Arabian deserts were "green") and also sent people to NW Africa (this based both in archaeology and population genetics). We see Aterian type tools (the characteristic back-tipped "spear point") not just in NW Africa but also in India, we see the same kind of seashell usage in NW Africa and in Palestine, that Aterian wave was a part of the wider "out of Africa" expansion, even if in this case they remained in Africa (but not in the core area towards Sudan, Ethiopia, Eritrea, Chad and maybe Uganda).

      They do mention, in case you're not aware, the presence of Homo sapiens in much older times however toward Morocco. They call this site Témara, although I'm more familiar with the name Jebel Irhoud. From a genetic viewpoint, I see no reason to think that those first settlers of NW Africa (Jebel Irhoud) survived but I am reasonably persuaded that the Aterian ones did and left a legacy, which is still present across the board at very low frequencies (around 1%) but much more important in South Morocco (around 14% of the local genome). This I already mentioned before with links and such, I believe.

  49. Yes Maju, it is considered from Aterian period , onlyn the first Thermoluminescence test in England gave the date of 92000 Y BCE, but they date the site around 100000 years, there are not only stone tools but severals bones of savanah animals , and also some human bones , the site seems to have been inhabited by severals generations, they are other articles on it, and a Tunisian specialist gave an interview. very interesting, thank you for your ineterest.


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... OFF (keep it that way, please)