No Levantine ancestry for Ashkenazic Jews

Last year, we applied GPS to the genome of ~400 Ashkenazic Jews (Das et al. 2016). GPS localized most AJs along major ancient trade routes in northeastern Turkey adjacent to primeval villages with names that resemble the word “Ashkenaz” (covered here, here, and here)

Figure from Das et al. (2016)

This is the first and only place in the world named Ashkenaz and it was compatible with biogeographical evidence. It was also compatible with the hypothesis of an Irano-Turko-Slavic origin for AJs and a Slavic origin for Yiddish, proposed by Paul Wexler, and at odds with the Rhineland hypothesis advocating a Levantine origin for AJs and German origins for Yiddish – supported by many linguists.

These findings were very coherent and made sense, except of course, that they flew in the face of years of genetic and linguistic research. No wonder the dormant value on Yiddish made the list of most controversial values in Wikipedia after our study was published.  At best, geneticists and linguists could have at least argue that their results have some internal consistency (i.e., I will agree with you if you will agree with me, but neither of us is right). To maintain their self-conviction, authors oftentimes contradicted themselves or clung to very sketchy arguments. For example, Behar et al. (Behar et al. 2013) has published a paper titled “No evidence from genome-wide data of a Khazar origin for the Ashkenazi Jews“ where they wrote:

We confirm the notion that the Ashkenazi, North African, and Sephardi Jews share substantial genetic ancestry and that they derive it from Middle Eastern and European populations, with no indication of a detectable Khazar contribution to their genetic origins.

This is despite of the fact that their own biogeographical analysis traced Jews to Western Turkey, not too far from ancient Ashkenaz in Eastern Turkey (See figure below in red). Doron Behar refused to provide the data from their figure, as if they couldn’t be obtained from their blurry figure with a little bit of effort. Behar and colleagues must have hoped that considering Turkey as part of the Middle East and ignoring the common Turkish origin of the Khazars and Ashkenazic Jews would be sufficient to convince their readers.

3 plots.jpg

Figure from Das et al. (2017)

Next comes Flegontov’s study of the admixed Ket people (Flegontov et al. 2016a) where the authors employed GPS using the GenoChip . Why is this of relevance? Because in challenging the study of Das et al. (Das et al. 2016) Flegontov et al. (Flegontov et al. 2016b) had to criticize a paper they published only six months earlier.

Here is Flegontov a)

We genotyped approximately 130,000 autosomal SNPs and determined mitochondrial and Y-chromosomal haplogroups with the GenoChip arrayThen we applied GPS (Elhaik et al., 2014)… For that purpose we compared the GenoChip SNP array data for the Ket, Selkup, Nganasan, and Enets populations to the worldwide collection of populations (Elhaik et al., 2014) based on 130K ancestry-informative markers (Elhaik et al. 2013)…. Kets derive roughly 30–40% of their ancestry from ancient North Eurasians… In addition to the proposed population and geographic location, GPS also reports prediction uncertainty (the smallest distance to the nearest reference population)..,

And Flegontov b)

Pitfalls of the geographic population structure (GPS) approach applied to human genetic history.

[GPS] is hardly suitable for admixed populations.

Another fundamental problem lies in data reduction inherent in the GPS approachgenotypes at about 100,000 sites (Das et al. 2016) are not analyzed directly, but collapsed to just few variables, i.e. admixture coefficients for nine hypothetical ancestral populations.

Flegontov c), if you must know, was one of the three reviewers that approved Das et al. (2017) that invalidated the arguments of Flegontov b), but supported the methodology of Flegontov a).


Next comes linguist Aptroot (Aptroot 2016). There is little doubt that Aptroot can pile up linguistic evidence produced by herself and colleagues but this would be useless since the self-consistency of the Rhineland hypothesis is no longer the issue. The challenges to the German origin of Yiddish are now based on historical, demographic, and genetic perspectives wrapped by Weinreich’s truism that the history of Yiddish mirrors the history of its speakers. If Yiddish is a German language then history and genetic evidence must be in agreement with that – or Yiddish is not a German language. Aptroot writes:

The idea that Ashkenazic Jews in Eastern Europe were not necessarily descendants of Jews who lived in Germany in the High Middle Ages could explain the purported “exceptional growth” in population numbers among this group in Eastern Europe…. Linguistic evidence, however, does not support the theory that Yiddish is a Slavic language, and textual sources belie the thesis that the name Ashkenaz was brought to Eastern Europe directly from a region in the Near East.

In plain words Aptroot claims the following: the only way to explain the vast number of Jews in Germany is via a “Demographic miracle” (as opposed to migration from other regions) – a nonsensical argument made by some authors (Atzmon et al. 2010; Ostrer 2012) – however our linguistic hypothesis exists in a separate universe where no one cares about historical, demographic, and genetic evidence to the contrary. We are quite happy about it, so leave as alone.

In answering our arguments, as this is still a rebuttal, Aptroot limits herself to the origin of the name “Ashkenaz” that, according to her, although associated with the Near East could not possibly have been brought by Near Eastern Jews named Ashkenazic Jews to Germany. No, according Aptroot Jewish immigrants in Europe (but not from the Caucasus) have transferred “Ashkenaz” from the bible onto Germany. Had Aptroot read Hebrew, ancient Hebrew, or even Aramaic writings she would avoid taking this route.

Biblical names were not selected at random and glued to new places Jews encountered. A translocation was made only when the two names had similar sounds. Germany and Ashkenaz do not have the same sound. Moreover, Judaeans already knew Germany as “Germana,” or “Germamja” in the Iranian (“Babylonian”) Talmud – so what possible reason do they have to call it Ashkenaz all of the sudden? Germamia was associated with Gomer, Noah’s Son, whereas Ashkenaz was associated with Noah’s grandson Ashkenaz and the Caucasus region – again, two different entities. Perhaps Aptroot can enlighten us on these problems with her hypothesis – problems that, as a linguist, she SHOULD be well familiar with.

The entire notion that Germany is Ashkenaz is based on a misunderstanding of Rashi’s writing. This can be easily contested by examining the writings of his followers, which we can identify since they used Rashi script. These authors referred to Germany as Almania, not Ashkenaz. Germany was called by many names, usually after the dominant tribes that inhabit it, but none of these tribes was ever called Ashkenaz. Aptroot’s proposal makes no sense. What is far more reasonable is that Ashkenazic Jews from ‘ancient Ashkenaz’ who started settling in Europe from the 10th century onward carried their name with them, just like Persian Jews and Moroccan Jews. The only difference here is that the place name Ashkenaz, disappeared over time and was forgotten, until we found it in 2016. Ashkenaz doesn’t mean Jewish. It is the name of a place. Some Turks are very similar to AJs and can actually be considered Ashkenazic-non-Jews.

No point describing the pitfalls of other authors, these are discussed at length in (Das et al. 2016). Let us now consider a more important question – can we be certain that our results truly reflect the past?

By the time this paper was written ancient DNA data from the Levant and the Near East had surfaced. These data allowed a direct examination of AJs ancient origins. Using two different methods we found that AJs have 0-3% ancient Levantine ancestry. AJ’s ancestry was 88% ancient Iranian with some Anatolian – in agreement with our previous findings. These findings not only reject the German origin. In fact, they rule out any ancient European origins. As you can see below, it is simply not in the cards. Ironically, the only ones boosting the European and Levantine ancestry numbers (AJs on the far right) are the half-Jews and not because of their Jewish descent!


Figure from Das et al. (2017)

The results were replicated in a separate analysis. The evidence for the non-Levantine origin of AJs is very strong. Interestingly, these findings also explain the results of previous genetic studies that relied on the similarity between AJs and Palestinians to claim Levantine origins for AJs. This similarity, however, is likely based on the Iranian and Turkish components of Palestinians (about 40%), not the Levantine one that Jews don’t have. The only ones with a full Levantine ancestry (as far as we know) are about half of the Bedouins (far left). Please note, that Semitic DNA from Israel is still missing, however Lebanese Semitic genomes and Egyptian ones were published. These genomes exhibit high similarity to the Levantine genomes used in our study, so we do not expect major surprises when Israelite Semitic data would be published.

Das et al. 2017 results also emphasize how useless are genetic tests that report “Jewish ancestry,” based of course, on similarity with other Jews, not actual biomarkers (Elhaik 2016). It is baffling that people are willing to pay a lot of money to learn absolutely nothing about themselves. It’s like saying that cats are cats.

From The Smithonian

Fortunately, cats don’t make us buy ancestry tests for them because they do not need that kind of validation. They remained pretty much the same over the past 9000 years and they know it.




 Aptroot M. 2016. Yiddish language and Ashkenazic Jews: A perspective from culture, language and literature. Genome Biol. Evol. 8:1948-1949.

Atzmon G, et al. 2010. Abraham’s children in the genome era: major Jewish diaspora populations comprise distinct genetic clusters with shared Middle Eastern ancestry. Am. J. Hum. Genet. 86:850-859.

Behar DM, et al. 2013. No evidence from genome-wide data of a Khazar origin for the Ashkenazi Jews. Hum. Biol. 85:859-900.

Das R, et al. 2016. Localizing Ashkenazic Jews to primeval villages in the ancient Iranian lands of Ashkenaz. Genome Biol. Evol. 8:1132–1149.

Elhaik E. 2016. In search of the jüdische Typus: a proposed benchmark to test the genetic basis of Jewishness challenges notions of “Jewish biomarkers”. Front. Genet. 7.

Flegontov P, et al. 2016a. Genomic study of the Ket: a Paleo-Eskimo-related ethnic group with significant ancient North Eurasian ancestry. Sci. Rep. 6.

Flegontov P, et al. 2016b. Pitfalls of the geographic population structure (GPS) approach applied to human genetic history: A case study of Ashkenazi Jews. Genome Biol. Evol. 8:2259-2265.

Ostrer H. 2012. Legacy: a genetic history of the Jewish people. Oxford: Oxford University Press.


This entry was posted in BLOG. Bookmark the permalink.

9 Responses to No Levantine ancestry for Ashkenazic Jews

  1. M says:

    Someone wanted to know How do you explain away all that J and E in European Jews? E/J/G/T add up to a little over 70%. These lineages are not Native European. That 70% Figure is about as similar to the figure that haplogroup A/B/ and E are found in African Americans.

  2. Jim Dean says:

    I have been following this debate for some time and have seen a clear bias for wanting a certain outcome that validates a connection to the “choseness” factor. There is no other explanation as to why so many trained genetic scientists will do back flips, even within their own research, much less others, to end up where they want to go. Ancient genetics has made wonderful strides in the last decade, and as always happens, the apple cart gets tipped over for past theories where those with a vested interest in their continuance seek to pick and choose the data that fits their case. This is a classic fingerprint for when politics enters into science. The entry of Dr. Elhaik’s GPS was a natural evolution, in a similar way that after 9-11 American Intel agencies began sharing each other’s data to produce more accurate analysis for the critical area of national defense, and the public reaction was “why the hell was this not done long ago?”. And the answer was that various parties sought to control the outcome of the analysis (for the purpose of competing for funding) by limiting analysis by other parties that might find holes in it, or have better product. But I must say there was a valid concern that the more that Intel is shared the higher the chance that enemy Intel agencies can penetrate. This battle over AJ origins debate will continue not so much because it is a scientific debate, but because of the internet evolution of so many of us having been profiled as to what our selected beliefs are, and how supporting those beliefs, true or not, gives others influence over us, and yes, I mean politically.

  3. Montoglia says:

    You said in your response to Flegontov that “The current GPS version is unsuitable for analyzing two-ways mixed individuals (e.g., Chinese-British) and will report the middle location of the parental populations (in this case, South Russia) since both parental populations are “pulling” in equal strengths”.

    Wouldn’t that mean that Ashkenazi Jews could have Levantine and European ancestry, and that GPS is just showing the middle location, that happens to be in Turkey?

    • eelhaik says:

      There is an entire section (“The Geographical and Ancestral Origins of AJs”) in Das et al. (2016) that addresses this issues :

      “GPS findings raise two concerns: first that the Turkish “Ashkenaz” region may be the centric location of other regions rather than the place where the Ashkenazic Jewish
      admixture signature was formed; second, in the absence of “Ashkenazic” Turks it is impossible to compare the genetic similarity between the two populations to validate the common origins implied by the GPS results.”

      The answer is no, AJs are native to the region and very similar to the local population.

  4. This reads like satire. You are aware that Ashkenazi Jews are extremely mixed, right? Supposing that your DNA GPS location reflect their unified origin as opposed to the average of their genetic admixtures is ludicrous.

    Gomer was Noah’s grandson, not son, and Ashkenaz was Noah’s great-grandson. Bohemia was called “Canaan” in the Middle Ages.

    The distribution of Y-DNA hapologroups among Ashkenazim is consistent with Levantine populations, not Anatolian or Iranian ones. The mtDNA is more consistent with European populations.

    The last figure looks like a poor Photoshop job. How exactly do you define “Levantine?” Bedouins are of known Arabian origin (hence the overwhelming amount of J1), so positing them as “Levantine” is absurd.

    The “demographic miracle” took place in the 18th and 19th centuries, 600-700 years after the Khazars had vanished.

    Why are you so desperate to “prove” the Khazar hypothesis, which is so clearly ridiculous?

  5. Pingback: Shaming 23andme works | Khazar DNA Project

  6. Ton Nuiten says:

    Reblogged this on Tinthor's Blog and commented:
    Courageous DNA-expert, Dr. Eran Elhaik: Ashkenazi Jews are not from Levantine…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s