August 8, 2012

Ample analysis of the genetics of European (Ashkenazi) Jews

A new paper on European (Ashkenazi) Jewish genetics is available:

Ehran Elhaik, The Missing Link of Jewish European Ancestry: Contrasting the Rhineland and the Khazarian Hypotheses. arXiv 2012. Open access. [arXiv:1208.1092v1 [q-bio.PE]]

(Supplemental material available HERE). 

The author prefers to use the term European Jews to that of Ashkenazi, because Ashkenaz is traditionally Germany and this may imply certain assumptions about the origins of this population, i.e. the Rhineland model, which he intends to contrast with the Khazar model and what he calls the Judean model

No intention exists in all the paper of investigating the Anatolian model, by which Western Jews would be descendants, in essence, not so much of Iron Age Jews from Palestine (Judean model) but neither from Europeans, be them Western (Rhineland model) or Eastern (Khazar model), but rather of Hellenistic converts from areas like Cilicia or Cyprus.

This alternative model is clearly supported by autosomal genetics, in this and previous papers. However haploid genetics seems to point to different sources.

In the paper, European Jews were roughly subdivided into Eastern (Belorussia, Latvia, Poland, and Romania) and Central (Germany, Netherland, and Austria) European Jews: EEJ and CEJ respectively in the graphs. They show some but not well defined structure however.

I'd say that the core of the paper's interest is in the following graphs:

1. Autosomal PCA:

Fig. 3 PCA scatterplot of West Eurasians (click to enlarge)

Again, as in previous analysis, including mine, Western Jews, in this case the European or Ashkenazi subgroup appears to cluster near Cypriots, suggesting a Southern Anatolian main origin (Tarsus area) very likely. However this region (ancient Cilicia) is still to be directly compared with Jews (I know it's just it's fine tuning but still when you have a theory, you'd like to see it reality-checked).

Another possibility would be hat this clustering is an artifact, assuming diverse origins homogenized by Jewish endogamy since the Middle Ages. But the fact that all three Western European subpopulations (Ashkenazi, Sephardi and Moroccan) tend to cluster over there (when they do not behave as fully distinct, as happened to me with all but Sephardites earlier this year).

Not shown here is fig. 7 that lists proportions of shared IBDs with West Asian and Caucasus populations (only): Caucasians are highest (roughly 9%), followed by Palestinians (c. 7%), while most groups are at c. 5% (except Lebanese who are at c. 3%).

In general this data supports some but limited Palestinian connection and a much more northernly origin of Western (or at least Ashkenazi) Jews. This is consistent with previous data (see below).

2. Haploid or uniparental relations:

Fig. 7 Relative affinity of (A) mtDNA and (B) Y-DNA of European Jews

Note please that the map is awfully drawn in some key details, for example the dark areas in Kurdistan correspond in fact to Lezgian (from Daghestan, in the NE Caucasus) and another one I don't discern well but that begins with I and hence must be an Iranian or Indian group, not Kurdish. The center of the "Kurdistan" affinity hotspot must be hence much farther NE, in Daghestan or Azerbaijan.

So by uniparental lineages, European Jews do appear as a mixture of Western (strongest in Y-DNA) and Caucasian lineages (strongest in mtDNA), supporting both the Rhineland and the Khazar hypothesis in a combined way. However no comparison with Cypriots nor Southern Turks was made, and these are the populations to whom Western Jews show most direct affinity by autosomal DNA, being Cilicia one, if not the largest, hotspot of the Jewish Diaspora in the Roman Era, a most likely realistic origin for

See also previous analysis:


  1. If European Jews are substantially Caucasian as the paper claims, why do they cluster so closely with Sephardic, Moroccan, and Syrian Jews, who presumably have little to no such admixture?

    1. It is a very good question. I limit myself here to reflect the facts found by this paper but I know that the haploid and autosomal data are contradictory.

      Personally I think that major clues may be found in Cilicia and Cyprus, where all Western Jews seem to have their autosomal closest relatives (not sure about Syrian Jews, who are less studied). It is very possible that the single "Turkish" sample used once and again (from Ankara?) is not representative of the Anatolian diversity and that other Anatolian groups are closer to the Caucasus in genetic terms, specially haplogroups.

      We have all the time this kind of problems when a large and likely diverse population like French, Turks, Iranians, Russians or Egyptians are simplified in a single sample from Lyon, Ankara, Teheran, Moscow or Alexandria. And then instead we micro-sample other areas like the Caucasus and oversample rare and highly inbred populations like the Druze.

      It may be that kind of artifact.

  2. Western Jews have little affinity to Turks and watsoever. They are closest to Cypriots , who are virtually Levantines , just European shifted genetically due to Hellenization they've been through. By contrast , modern Levant Arabs are also Arabian shifted , and the Druzes are the closest to pre-Arab Levantine population , to which the Jews are the closest out of all Middle-Easterners. Now that we have plenty of Mediterranean Jewish samples (Syrian , Turkish , Greek , Bulgarian , Italian , Algerian , Moroccan , Libyan , Tunisian , Djerban and Moroccan). We can see clearly the pattern of the Jewish diaspora.

    1. Not at all. As you can see in other studies (check the 'see also' section at bottom), Cypriots are closest to Turks. I also thought at first (in 2010, when the Atzmon and Behar paper was released) that Cypriots would be closest to Lebanese or Syrians but nope.

      In fact in my own Levant-and-Anatolia-only mini-study, Cypriots, Turks and Sephardites clustered together. very tightly even at K=9, more than enough depth for just seven populations. (I could not work with the too-bottlenecked Ashkenazi or Moroccan Jewish populations but previous studies always place them very close too Sephartdites, so it should be valid).

      Cypriots are also not really "Western-shifted", at least not more than Anatolian Turks: they have very little genetic relation with European Greeks in spite of language and history. However they may have got closer relation with Hellenized Anatolians from past time, most of which were deported to Greece after WWI and the subsequent Anatolian wars. But modern Turks should also essentially descend from those Hellenistic Anatolians (and only minimally from Central Asian or other ancient Turkic immigrants), so it is, I believe a matter of comparing with the correct modern Turkish population from the Mediterranean region of Turkey, specifically from the Adana, Mersin and Hatay provinces, which are very close to Cyprus.

      Meanwhile Cyprus will remain the closest reference but not because of "shift" because in the West Asia only analysis, a single persistent component (cyan at K=6) is shared by the three populations: Cypriots 78%, Sephardites 76% and Turks 78% (although Cypriots and Sephardites are more similar in the other components' distribution). This "Highlander" component is also important (to lesser extent) among Lebanese (46%), Jordanians (43%) and Syrians (42%) but weak among Palestinians (only 17%).

      Sephardic Jews do have some rather subtle but still interesting "excess" presence of one of the two components I labeled as "Palestine", the less common one (Palestine 1), which is found at 15% among Palestinians and 5% among Sephardites, however the main Palestine-only component (Palestine 2, at 34% among Palestinians) is absolutely not outstanding among Sephardites.

      With my means and limited knowledge, I cannot discern if these or other Palestinian components (Palestinians have high genetic diversity and also the richest regional genome in terms of different constitutive components) have some geographical structure (like ancient Judea vs. Samaria or Galilee, or whatever else) but it is indeed a possibility. But what is clear is that the "Highlander" component seems a quite homogeneous thing and not any artifact (much like NW Europeans are very homogeneous among them, from Ireland to Poland) and in this bloc of high "Highlander" genetics are Jews, at least Sephardic Jews, along Cypriots and Turks but not genuine Levant peoples (i.e. Syrians, Lebanese, Palestinians or Jordanians).

    2. PS- If you don't trust me, you can do it yourself, using the "1000 Genomes" data and freeware, just like I did. The isntructions I followed were posted by Razib at this address:

      If you have any doubt, feel free to ask.

    3. PS- Notice also that the concept of "Levant" in Behar 2010 was extended from the usual meaning (Syria, Lebanon, Palestine/Israel, Jordan) to include also Cyprus for ideological reasons. So when Behar says that Western Jews look Levantine, he just means they look Cypriot, not Palestinian, Lebanese or Syrian.

      And genetically speaking Cypriots are closest to Anatolian Turks.

  3. Luis, which Turkish set of samples are you referring to? Cinnioglu (Y chromosomal) or Behar (autosomal)?

    1. Haven't checked but I think that this paper is using the Behar sample.

    2. So, in your first post you are referring to the autosomal, and not to the uniparental, analyses of Turks in the Elhaik paper.

      For your info, the Behar Turks (19 individuals in total) were exclusively sampled from the part of Turkey that was the Cappadocia region historically (Mait Metspalu, one of the authors of the Behar study, told me). Unfortunately, he does not know which provinces of the Cappadoian part of Turkey they were sampled from. BTW, Cappadocia did not include Ankara (Ancyra or Ankyra, historically), so I do not think any of them were sampled from Ankara.

      Here is a map showing the location of Cappadocia in Anatolia:

    3. In the uniparental analysis maps there's only one dot, what does that represent? I don't have time to double check everything.

      Whatever the case it's clear that you can't compare a large area like Turkey, including some 70 million people, and also densely populated in the past, with a point (and pointless) mega-sample like some 50 Galilean Druzes (again haven't checked the exact number but I know that they are many, they are inbred and non-native, and tend to distort things). There must be a sense of proportion and Turkey, I think, deserves some more carefully detailed study on its own right but also in this case of comparison with Western Jews, who look Cypriot or Turk by autosomal genetics.

    4. And I know where Cappadocia is, I was there as tourist some 20 years ago. Nice place that I certainly enjoyed, specially hosted at motel made out of one of those "fairy chimney" houses - so fresh! :)

    5. I checked the supplements of the Eran paper, which give detailed information about the sources of the samples used for either uniparental or autosomal studies. These are the sources of the Turkish samples used in the Eran paper: the Y chromosomal analysis samples are from the Cinnioglu et al paper 2004 (from all over Turkey, 523 individuals intotal), the mitochondrial analysis samples are from the Behar et al. 2010 paper (383 individuals in total, from historic Cappadocia), but most of them are from the Tambets et al. paper 2000 ultimately, and the autosomal analysis samples are again from the Behar et al. 2010 paper (from historic Cappadocia, 19 individuals in total).

    6. Alright, thanks for clarifying, Onur. I wonder why Eran pools all Turks/Kurds (which Cinnioglu reports regionally) as a single dot. And I also wonder if improved info can be achieved by comparing with the regional samples. I guess I'll check myself when I have time... no promises though.

    7. Every population is represented by a single dot on those haplogroup affinity heat maps. So those dots are nothing but abstract representatives of the geographical locations of the populations. Such single dots are needed in heat maps for ease of understanding.

    8. What I mean is that Turks should better (at least to test my "Cilician hypothesis) not be treated as one single population but several.

      There's no reason I know why 70 million Turks should be treated as one single homogeneous population but then a dozen of Caucasus groups numbering often just in the thousands be treated as separate ones.

      In any case, I am interested in regionally-divided comparisons of large populations like said, Turks or French or Egyptians. They should be informative because the people of Cilicia, for example may have some different affinities from those from Troad or Pontus. Or whatever else... even if only to better understand Turks themselves regional analysis is interesting.

    9. Cinnioglu et al. divided Turkey into 8 regions (9 if Istanbul is counted as a region) and examined each region separately. Of those regions, the one you would most want for the Eran paper is the region 6. Eran Elhaik could do the same instead of lumping them into a single population. I certainly agree with you that Turkey or Turks should be investigated regionally, just as Italians are frequently investigated regionally. BTW, Greeks too should be investigated regionally, as, like Turks and Italians, they show significant genetic variation between regions.

    10. As you insist and you don't come with the issue well chewed, I have been comparing and the results are that there is no apparent strong similitude that I can see with my limited means and time between the Mediterranean region of Turkey and Ashkenazi Jews in raw percentage comparison.

      Left Euro-Jews, right Mediterranean Turks:

      E-M35 > 18% : 12%
      G ··· > 07% : 09%
      I ··· > 04% : 03%
      J1 ·· > 22% : 15%
      J2 ·· > 19% : 24%
      K* ·· > 02% : 03%
      N ··· > 00% : 03%
      R1b ··> 09% : 22%
      R1a ··> 13% : 09%

      They don't look much the same, diverging specially in the ratio of J and R1 subclades (Ashkenazim have more R1a and J1, what may be caused by the Khazar introgression and maybe also by residual Palestinian origins in the J1 case). The rest are not that different but these are the major lineages.

      I'm going to see if I can check Sephardites from Adams 2008, the Iberian Y-DNA paper, because these are less likely to have got odd founder effects as Ashkenazim do.

      Another possibility we cannot exclude is that there is undue high presence of "Judean" (and Khazar) Y-DNA lineages because of male bias in reproduction with converts. So it may need a multivariate analysis.

      Also particular lineages may, I guess be tracked to origin, by means of haplotypes. But that's a colossal task.

    11. I checked Sephardites (cf. Adams 2008: The results are (Med. Turks : Sephardites : Ashkenazim : haplogroup, only reporting >1% freq., rounded up):

      12 : 08 : 18 : E-M35
      09 : 16 : 07 : G
      03 : 01 : 04 : I
      15 : 22 : 22 : J1
      24 : 25 : 19 : J2
      03 : 06 : 02 : K* (T)
      00 : 02 : 04 : P* (Q)
      21 : 13 : 09 : R1b
      09 : 05 : 13 : R1a

      In general Sephardites look a bit closer to Southern Anatolians than Ashkenazim do but this varies by haplogroups:

      Best match Med. Anatolians w/ Ashkenazim: G, I, K* (T)
      Best match Med. Anatolians w/ Sephardites: E-M35, J2
      Best match Sephardites w/ Ashkenazim: J1, R1b

      So I'd say that, Y-DNA-wise, the most salient Western Jewish-specific characteristics are low R1b (c. 11%) and high J1 (c. 22%). These look Palestinian-like, and certainly non-Anatolian.

      But Sephardites specially show Southern-Anatolian affinity in the high levels of J2 (c. 22%) and relatively low ones of E-M35 (c. 10% on average).

      The Ashkenazi-South Anatolian elements of affinity are less important lineages: comparatively low levels of G and T and higher levels of I.

      Inconclusive for my taste. It needs deeper analysis.

    12. I am not insisting on anything. It is you who brought forward the southern Anatolian origin hypothesis and I just tried to help you in your research.

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

  5. This comment has been removed by the author.

  6. This comment has been removed by the author.

  7. Octavià,

    Armenians and Greeks are extremely small minorities in today's Turkey. Their total does not even come close to 1% of the current population of Turkey, so today they are negligible in number.

    Today, ethnic Turks make up a little more than 3/4 of the population of Turkey, while Kurds make up about 15% of Turkey's population, other ethnic groups forming the remainining 8%. (in Turkish; it is the most scientific and comprehensive study of the distribution of the ethnic groups in present-day Turkey I have seen)

    The Behar et al. 2010 Turkish samples (also used in Yunusbayev et al. 2011 and Rasmussen et al. 2011) and the Hodoğlugil et al. 2012 Turkish samples are exclusively ethnic Turks, the Dodecad and Eurogenes Turkish samples are almost exclusively ethnic Turks and the great majority of the Cinnioğlu et al. 2004 samples from Turkey are ethnic Turks.

    1. Please don't feed the troll: Octavià is banned in my blogs because he's a troll and an attention whore.

  8. This comment has been removed by the author.

  9. This comment has been removed by the author.

  10. Here is a rough ethnic groups map of Turkey:

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

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

  13. BTW, there is a minor decline in the percentage of ethnic Turks in Turkey and a minor rise in the percentage of Kurds in Turkey since 1927, the year of the first ethnicity and mother tongue based census in Turkey. Ethnic Turks seem to make up between 80% and 85% of the population of Turkey while Kurds seem to make up barely more than 10% according to the 1927 census.

  14. Note: I always include Zazas in Kurds.

  15. Greetings. I just went through Elhaik's paper, and there are some things I don't understand - maybe you can help me...

    1) Is he identifying some specific "Markers" that are only present in Ashkenazi Jews and Caucasian people - but are absent in other Middle-Eastern/European/Jewish populations (What about Spharadic Jews from North Africa - perhaps they'll turn out to be "Khazars" too by the same Criteria)?

    2) In regard to figure 7 in that paper (Pairwise genetic distances between European Jews and other populations measured across mtDNA and Y-chromosomal haplogroup frequencies). How can someone claim that Ashkenazi Y-DNA is close to the Caucasians' Y-DNA? As far as I understood till today, Y haplogroup E, which is common among Ashkenazi Jews (15%-20%) is very rare among Caucasians; Y haplogroup J1 among Ashkenazi Jews is J1c3(J-P58) which is also very rare among Caucasians; and finally, the dominant form of Y haplogroup J2 among Caucasians is J2a4b* - which is rare among Ashkenazi J2...

    Can you please explain to me how can his claims be true in any way?

  16. I have the details a bit forgotten by now but I understand the following:

    (1) He is doing two analysis of various populations (of which the focus are Jews and specifically European Jews, term that he prefers to Ashkenazim):

    First the comparison (by statistical means, always somewhat simplifying but necessarily so) of autosomal genetics (nuclear genome minus Y-DNA). The results are shown in several graphs: figs. 3, 5 and 6, each one done with different means and showing different visualization means, to which most people who have some interest in population genetics are more or less familiar with.

    Maybe the lest common one and the one that best approaches your notion of "specific markers" is fig. 6, which deals with specific segments known as IBD markers. In that graph we can see how European Jews share the most with the populations labeled "Caucasus". Secondarily Palestinians are also notable but only in the error margin zone because in the median they are just like Turks. Lebanese are clearly the less similar ones but looking at the red (median) mark, after the Caucasus population, it is Turks, Cypriots and Palestinians (all about the same).

    Notice however that other studies (see also section of this entry for example), including other Western Jews, have produced different results placing Ashkenazim actually closer to Sephardim and Moroccan Jews. But the matter is very confused by strong founder effects and such.

    Then they also study mitochondrial DNA and the Y-chromosome, which are uniparental markers transmitted linearly from mother to daughter and father to son respectively. The results are plotted in fig. 7 but some populations are misplaced in Kurdistan (being from further Northeast or East), producing a very confusing impression.


  17. (2) It's very probable that they found that both Jews and East Caucasians are high in both J1 and J2, as well as G, without further detail of analysis. Personally I'd agree with you and based only on Y-DNA I'd compare Western Jews (Sephardic also) most directly with the populations of the arc between Greece to Palestine, which, with varied apportions also have the typical Jewish composite of E1b1b, J1, J2 and G but I would need the raw data to make a decent analysis and I can't find it anywhere in this preliminary paper (remember that it's not any peer-reviewed paper but a pre-pub for specialists). I downloaded the supplements but they seem impossible to use so I can only shrug beyond that, sorry.

    In general I found other relatively recent papers on Jews much more interesting but no one clearly conclusive. In general the autosomal genetics tend to cluster with Cypriots and Turks and I'd say that that's the core of the actual origin of Western or "Hellenistic" Jews (i.e. excluding most Asian and African Jew, of which each one is its own unique population, often genetically associated to their immediate non-Jewish neighbors). To some extent I also found that correlation myself in a mini-research I published in this blog (and linked to in the "see also" section again).

    "Can you please explain to me how can his claims be true in any way?"

    I don't say that their claims are true. In this case I found the proto-paper a bit confusing myself and I just reported it. Personally I feel it's rather unhelpful or at the very least mysterious with some results contradictory with previous ones.

    For example compare the figure 5 (Admixture analysis) with the ones by Behar 2010 or Bauchet 2007 or even my own for Sephardim (which according to Behar are similar to Ashkenazim - but IMO with a much less inbred ancestry: Ashkenazim tend to cluster with themselves too strongly, making comparisons hard to make). In all these studies Western Jews cluster with Greeks, Turks, Cypriots and Armenians primarily.

    I cannot defend it but I can mention it "for the record". That's what I did. Sorry not be able of helping much more.

  18. Thanks! I think you did help me - by confirming my suspicions - but I want to be absolutely sure I understood it correctly...

    If I understood it right, the autosomal and/or(?)IBD genetic tests can only give an overall estimation of how many markers two people/populations have in common, but they won't be able to detect tiny tell-tale signs & differences (which sometimes could make all the difference) like the differences between the Ashkenazi most common J1 (J-P58) and the Caucasians' most common J1 (which is not within the J-P58 sub-clade) for example...

    furthermore, in those autosomal tests two populations might appear much alike - with many different common markers - although historically they might have "composed" their DNA in two different ways... For example, according to Figure 5, if we took a typical Middle-Eastern population like the modern Syrians & Lebanese (which I think genetically look much like what the ancient Israelites must have looked like - though perhaps they got a bit more recent "Near-Eastern" influence), and we mixed them up to a rate of 40%-50% with typical East-European population, eventually we'll get a population that genetically looks much like the populations that were formed by natural mixture of populations along the borders between the Levant and East-Europe - Whether it is in western Turkey, Greece or South Balkans and whether it is in the Caucasus...

    So, therefore, the only reliable way of knowing whether Ashkenazi Jews have most of their ancestry in the Caucasus is to check small differences - like whether they got a similar distribution of E-M215 as the peoples of the Caucasus (they don't) and/or whether their J1 is from the same sub-clade typical to the Caucasus (it's not)...

    Did I get it right?

    1. "... they won't be able to detect tiny tell-tale signs & differences (which sometimes could make all the difference) like the differences between the Ashkenazi most common J1 (J-P58) and the Caucasians' most common J1 (which is not within the J-P58 sub-clade) for example"...

      That depends on the level of detail which the authors uses and in this case I feel it's a bit shallow. They could have worked harder and contrasted at the level you say but for some reason they did not. It's a bit like the small type of loans and contracts, you always want to read very carefully... also with genetic studies. The devil is often in the details, as they say.

      A typical case I have discussed in the past a lot is European R1b. R1b as such is a large haplogroup that spans not just Europe but large areas of Asia and Africa but then there are subclades that are very specific to certain areas of Europe and are almost non-existent elsewhere, yet using only the shallow approach (i.e. not looking at the deeper structure) there have been geneticists claiming that all expanded from Anatolia in the Neolithic - what for me at least is very hard to believe - but they have got some "popularity" in spite of that.

      It's more difficult to grasp but much of the same can happen with autosomal comparisons. In many papers and amateur studies (blogs and such) you see drawing very local "conclusions" from global-level comparisons at a depth of K=8 (eight clusters) or a bit more, when you probably need a more regional focus (scrap all those distant populations or at least make them grab less clusters by reducing their size) and allow ADMIXTURE (or STRUCTURE) to go deeper until something really informative arises. There's also a method to determine mathematically the optimal number of clusters but, even using it, sampling strategy is critical in my opinion (I'd rather scrap all clearly endogamous populations, like the so-typical Druzes, and if I have to oversize a sample I want it to be my focus populations not something obviously unrelated). In brief: it requires some thought and also honesty (sometimes I truly wonder...)


    2. ... "modern Syrians & Lebanese (which I think genetically look much like what the ancient Israelites must have looked like - though perhaps they got a bit more recent "Near-Eastern" influence)"...

      Honestly I do not think that modern Jews have too much "ancient Israelite" ancestry. IMO the main ancestry of Western Jews (Ashkenazim, Sephardim and others in the line like Moroccan and Italian Jews) is fundamentally Hellenistic/Roman Era converts from Anatolia and nearby areas (Cypriot Greeks are the closest think, followed by Turkish samples -Konya?-, sometimes also Greeks, Armenians... but Syrians and Lebanese are not too close in autosomal terms).

      A key issue here is that the Palestinian samples that we have are very very diverse and sport their own specific clusters, suggesting a very old distinctiveness as the one we could expect from ancient Hebrews/Canaanites, probably rooted in the local distinct Neolithic (PPNA) and maybe even the Paleolithic (?). Of course that they also display some admixture with other populations of the region but they are not by any means just "generic Arabs". Measures of diversity give Palestinians one of the highest figures in all Eurasia, just after Africans (although maybe some minor African genetic "infiltration" helps to that).

      Historical references, for example those recollected in the 1906 Jewish Encyclopedia, totally support a widespread Jewish proselytism (of which Christianity is just a major dissident sect) in the Hellenistic, Roman and even early Medieval period. Still, while many other Jewish populations have local (or sometimes a bit mysterious) origins (Yemeni Jews, Ethiopian Jews, etc.), Western Jews seem to share a common origin in what is now roughly Turkey, Cyprus and maybe also Greece and such. Of course it won't be too homogeneous because there were Jewish centers in many places but, from the genetic data, it seems to me that Anatolia and Cyprus are the main sources of modern Western Jews.

      There is however at least one component shared with Palestinians but at low levels. Patrilinerly (Y-DNA), modern Jews may retain a greeter degree of Palestinian (i.e. ancient Hebrew) affinity, as is apparent in the abundance of J1, E1b... but autosomally it is much less.

      One of the problem with Ashkenazi Jews is that they seem to have got a very strong founder effect (bottleneck) when they moved to Central Europe (more or less). That makes their autosomal pool too self-related for me to be able to analyze, so (based on other studies that show that Ashkenazi and Sephardi are very close, as are Moroccan Jews) I have used the less "bottlenecked" Sephardites for my only amateur study. This is partly because I'm not techy enough to make a "supervised" analysis however, which can also be made (i.e. forcing X population to compare only as if they'd be a simple mix of Y, Z and W, so to say).

      Not in this proto-paper but in the analysis of Tel Aviv University scholars (Behar, Atzmon), the Cypriot/Anatolian affinity is way too obvious (Behar expands "the Levant" to Cyprus, in a very political exercise, so he can still say that Western Jews have "Levantine" ancestry but that is certainly not continental Levant in any case). And for that there's no relevant difference between Ashkenazim and Sephardites (Ashkenazim do show some minor European admixture but not too much). And when I did compare months ago these same Sephardites with West Asian populations (Cypriots, Turks, Syrians, Lebanese, Palestinians and Jordanians) the Turkish-Cypriot affinity held very strongly.

      All the mentioned references are linked in the "see also" section at the bottom of the main article.

  19. Dear Maju,
    Ive never communicated with a Basque before, either in person or other means. I have been led to understand that you have a unique language unknown or related to any other language and your people are unique in that you are a self contained unit without admixtures. Is this true or have I been misinformed? If this is off topic Im sorry but I followed a link over to here and lo and behold there is a blog by a member of a nation that I was told no one really knows the origination of. Again maybe a lack of accurate resources. Is anything I have heard of the Baque people true?

    Sincerely if Ignorantly yours

    1. I would not use your wording but indeed Basque language or euskara is a language isolate. There are many language isolates through the World, or also small language families with very few members, but Basque is the only of that kind in Western Europe (or even all Western Eurasia, although the quite small and localized Caucasic and Uralic language families are close to this category as well).

      ... "your people are unique in that you are a self contained unit without admixtures".

      I don't think that is truly correct, we are a welcoming people towards those who respect us (and even some who do not, I fear), but it is probably certain that Basques have been relatively isolated through the centuries, probably because of determined resistance to Indoeuropean invasions (Celts, Romans and Romanic peoples successively) and also because the Atlantic was not at all as important as the Mediterranean before the High Middle Ages at least, what kept us relatively safe.

      Whatever the case the origin of Basques is confounded with the other Europeans who live around us and who have forgotten the old language. Basques just display less apparent influence from the East, almost lacking for example obvious Neolithic lineages like G2a, J2b or E1b. However 10% of Basque male lineages are I2a, which is surely Neolithic Cardial with origin in the Balcans. Almost all the rest is R1b (SW European variant). Pretty much monolithic with very few exceptions.

      I think it can be said with great certainty that Basques were already what they are now in genetic terms at the very early Neolithic. The site of Paternabidea (Navarre) displays, 7000 years ago, a genetic pool almost exactly identical to that of modern Basques, what can be said of very few peoples in Europe, if any at all. Whether these peoples were newcomers, natives of Paleolithic roots or a mixture of both is debatable but I'd rather think based on the currently available data that the Paleolithic ancestry was at the very least important.

      As for the language there is some good likelihood that it may have (with the natural variation in time and space) the same family spoken by most European Neolithic peoples. A Paleolithic origin has been speculated as well but the Vasconic substrate seems important in Italy, which lacked the Magdalenian culture influence, what only seems to leave the Neolithic spread as plausible vehicle for the spread of Vasconic languages, which once were surely spoken in much of Europe: from Scandinavia to Italy and from Central Europe and maybe the Balcans to the Atlantic coasts.

      That's why I understand that euskara is the language of persistence or resistance (eutsi in Basque, which regularly becomes eusk- in compounds). All other languages and particularly Indoeuropean are erdara (either the half-language or the language of division, partition: erdi=half, erditu=to divide in two, also to give birth). The suffix -era/-ara means more than just language or dialect but actually mode. It might also mean a way of life, a way of being. But rambling a bit in this last paragraph, I'm sure you will excuse me for that.



Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).