How The Forward shamed 23andme into “fixing” Jewish history

“Telling the truth is less demanding than telling a lie.” 
 Eraldo Banovac

To those of you who missed the drama, allow me to start from the beginning. In their latest round of updates, 23andme stumbled upon the Behar et al. (2003) or the later Rootsi et al. (2013) co-authored by Behar and many other scientists that associated the Y haplogroups R1a1 or R-M198 with Near Eastern populations and possibly with the Khazars (given that this was one of the largest conversion events in the history of the region). Here is what Behar et al. (2003) wrote:

One attractive source would be the Khazarian Kingdom, whose ruling class is thought to have converted to Judaism in the 8th or 9th century (Dunlop 1967). This kingdom flourished between the years 700 C.E. and 1016 C.E. It extended from northern Georgia in the south to Bulgar on the Volga River in the north and from the Aral Sea in the east to the Dnieper River in the west—an area that falls within a region in which haplogroup R1a1 NRYs are found at high frequency (Rosser et al. 2000). records migration of Khazars into the Hungarian Duchy of Taskony in the 10th century. The break-up of the Khazar Empire following their defeat by invading Rus led to the flight of some Khazars to central and northern Europe. Although neither the NRY haplogroup composition of the majority of Ashkenazi Jews nor the microsatellite haplotype composition of the R1a1 haplogroup within Ashkenazi Levites is consistent with a major Khazar or other European origin, as has been speculated by some authors (Baron 1957; Dunlop 1967; Ben-Sasson 1976; Keys 1999), one cannot rule out the important contribution of a single or a few founders among contemporary Ashkenazi Levites.

David Goldstein, one of the authors of the Behar et al. (2003) study wrote in his book (Goldstein, 2008, pp. 73–74):

Could Khazaria, I wonder to this day, be the source of Ashkenazi Levite R-M17 Y chromosome?

As with much else of genetic history, there is no way to be sure. […]

I was initially quite dismissive of Koestler’s identification of the Khazars as the “thirteenth tribe” and the origin of the Ashkenazi Jewry. Was this not just another self-aggrandizing Lost Tribe narrative bereft of evidence?

I am no longer so sure. The Khazar connection seems no more far-fetched than the spectacular continuity of the Cohen line or the apparent presence of Jewish genetic signatures in a South African Bantu people. […] I cannot claim the evidence proves a Khazari connection. But it does raise the possibility, and I confess that, although I cannot prove it yet the idea does now seem to me plausible, if not likely.

This was mentioned by Falk (2014).

Rootsi et al. (2013) confirmed the general region:

This line of evidence supports a Near Eastern origin for haplogroup R1a-M582, and hence the Ashkenazi Levite lineage as well.

This is the scientific evidence used by 23andme to make their report. Obviously, no clear cut evidence for a Khazarian origin (no surprise, provided that the Khazars are gone). Instead, there is a strong association with the Near East region (the Caucasus), which allows to make a reasonable speculation of a Khazarian origin, simply because of the historical evidence of their Judaization.

In 2013 I published a study that supported the Khazars theory (Elhaik 2013) over a Judaean ancestry, suggesting that Ashkenazic Jews derived from Near Eastern populations around the first millennium. The conclusions were made based on similar observation to those made by Behar, that is, high similarity of AJs to Caucasus populations.

The Rootsi study was published later that year, only that now its authors struggled to explain why their results CANNOT possibly indicate a Khazarian origin although the haplogroup is from the Near East using their usual logical trickery (if AJs are from the Near East and Israel is in the Near East – then AJs are from Israel).

After the publication of my study, many of the authors of the Behar and Rootsi paper, attempted to refute it, and of course, their own conclusions (Behar et al. 2013). It worked pretty poorly for them as they also mapped Jews to Eastern Turkey, in support of the theory they tried to dispute (see below), but they nonetheless concluded that their findings support a “Middle Eastern origins” (if AJs are from Turkey and Turkey and Israel are in the Middle East – then AJs are from Israel). Discussed here and in (Das et al. 2017).

Note, that these findings do NOT necessitate a Khazarian origin, as we showed in Das et al. (2016) see here. It is more likely that the mass conversion occurred much earlier and south to Khazaria in ancient Ashkenaz, a region we identified in 2016 in northwestern Turkey. The region was under Khazarian control, which creates historical dilemmas that biology alone cannot resolve. In other words, the genetic results for Khazars or Irano-Greco-Turks who converted to Judaism would be indistinguishable. It may be that the latter is more likely, however if you look close, there are about 7% of AJs that are not in ancient Ashkenaz and very close to Ukraine, and again a historical Khazar land. Could they be the descendants of the Khazars or maybe locals who converted to Judaism? With these data, it is not possible to tell. The answer is in ancient DNA, but before we will cover that, let us get back to the 23andme case.

Understandably, the 23andme update was received with confusion. Why? Well, AJs ARE from the Caucasus-Turkish region, as any biogeography analyses done so far indicated (Das et al. 2017) (and see above). GPS Origin results are here. To hide the ancestry of Ashkenazi Jews (AJs), 23andme and almost every company, except GPS Origins, which I developed, simply report “Ashkenazic Jewish” as ancestry. In other words, they tell AJs (and many other non-Jews) that they are AJs, and thus overcome the origin/geography question. So WHERE ARE ASHKENAZIC JEWS FROM? Well, it’s a sort of don’t ask don’t tell type of situation. And it works extremely well thanks to the PR people that these companies retain, or at least worked so well until a 23andme employee stumbled upon Behar’s papers and made the mistake of telling R1a-M582 males that they are Near Easterners AJs. This would make sense for any population, except AJs, where the ancestry is a mask for the geographical origins and created a dilemma… how can a Jew be anything else other than a Jew? There is an ugly word for that. GPS Origins, by contrast, reports geographical origins equally, indiscriminately, evenly, and accurately for all people of all genders and all kinds without fixing the results and certainly without religion affiliation.

Angry 23andme customers quickly wrote The Forward who shamed 23andme and blamed them in no less than:

wading into a political-genetic debate with far-reaching implications for Jewish identity and the Palestinian-Israeli conflict. The Forward

This is not something you want to be accused of when your company is worth 1 billion dollars (on paper) and you are in a lot of debts since your primary strategy of luring people into give you their genetic data so you can discover disease-causing genes and make a lot of money is not working THAT well…

23andme pedaled backward and offered an apology to all the baffled R1a-M582 carriers assuring them that they are not of Caucasus/Khazar origin anymore as Behar and colleagues reported/speculated and The Forward got to hang another skull on its belt. So is this the end of story? Of course not.

Not realizing (or just ignoring the fact) that the 23andme mutation has been reported by Behar and colleagues and not our group, The Forward article linked our work with this 23andme embarrassment. The Forward are huge fans of my/our work. It was hardly a surprise that they butchered the science along the way. In fact, this is their 6th(!) report aimed to discredit our scientific findings concerning the origin of AJs. Unlike 23andme, I choose to stick with the science. Truth does not need publicity, lies do (Abhijit Naskar).

I would also like to remind The Forward of our recent study where Ashkenazic Jews showed practically no Levantine ancestry when compared with ancient DNA from the Levant and the Middle East and only ancient Irano-Turkish ancestry as can be expected of a Caucasus population (Das et al. 2017). It is the #3 most-read article in Frontiers in Genetics (of all times), but maybe they missed it.


Do you want to learn more about the ancient Israelites? I recently developed a DNA test, that compares your DNA to the DNA of the ancient Israelites and many other ancient populations using ancient DNA from real people who lived in the past. The tests also include a very detailed background on each culture. Check it out here  Just upload your DNA file and order a test or order a DNA test kit, if you never took a test before.



Das R, et al. 2017. The Origins of Ashkenaz, Ashkenazic Jews, and Yiddish. Front. Genet. 8.

Elhaik E. 2013. The missing link of Jewish European ancestry: Contrasting the Rhineland and the Khazarian hypotheses. Genome Biol. Evol. 5:61-74.

Rootsi S, et al. 2013. Phylogenetic applications of whole Y-chromosome sequences and the Near Eastern origin of Ashkenazi Levites. Nat. Commun. 4:2928-2937.

13 thoughts on “How The Forward shamed 23andme into “fixing” Jewish history

  1. Interestingly, Oster and the Forward still speak of the ‘Khazar hypothesis’, while your results more precisely point to an ‘Irano-Turko-Slavic hypothesis’ or ‘İşkenaz hypothesis’, which is not necessarily the same thing, right?

    Ostrer states: “Genetic samples from known Khazar descendants would be needed as control samples to confirm that some Ashkenazi Jews have Khazar ancestry.”

    Is he right, given the ancient İşkenaz-DNA you already have? What about the prospects of finding ancient Khazar DNA (North of the Caucasus/Black Sea), given the recent excavations in the region?

    At any rate, thanks for your intellectual honesty and courage. Keep up your great work!

    1. Right, the mass-conversion event likely took place in ancient Ashkenaz in Turkey. It is still unclear to me what happened in Khazaria.

      Ostrer is right about the Khazar samples needed to address the Khazar question, but his logic is very inconsistent, because he vouched for a Judaean origin without any samples from Judea.

      Ashkenazic Jews are far more similar to Turks from the same region than to anyone else in support of their origin from the area. Is this the only place of origin? That remained to be found.

      1. This is blatantly false. Ashkenazi Jews do not genetically resemble Turks, they most closely resemble Sicilians autosomally. In terms of haplogroups, their Y-DNA best resembles Levantine populations and is extremely distinct from Turks or populations from the Caucuses (who incidentally have nothing to do with Khazaria). Ashkenazi Jews, like all other Jews as well as most Levantine populations–especially those with less Arabian influence–have a substantial amount of ancestry from Asia Minor (or Anatolia), which has nothing to do with Turkic peoples and the Turkic ancestry of modern day Turks.

  2. Do you plan on issuing a reply to Harry Ostrer’s attack on your work in The Forward? Would that magazine have the integrity to allow you to publish such reply in their website?

      1. They are certainly dogpiling on your work (now there is a new article in the Forward from linguist Alexandre Beider) without anyone really challenging them.

        People don’t really understand the intricacies of genetics, and rather follow digested articles like these in the Forward, constantly painting Ostrer, Behar etc as the “credible” geneticists and your work as “junk science”.

        At some point this will become “conventional wisdom” and too ingrained in collective imaginary for facts to matter anymore.

        Sad as it is, PR matters, and baseless attacks have to be exposed with similarly public rebuttals.

  3. Thenotoriousyid says:

    “This is blatantly false. Ashkenazi Jews do not genetically resemble Turks, they most closely resemble Sicilians autosomally. In terms of haplogroups, their Y-DNA best resembles Levantine populations and is extremely distinct from Turks or populations from the Caucuses (who incidentally have nothing to do with Khazaria). Ashkenazi Jews, like all other Jews as well as most Levantine populations–especially those with less Arabian influ toyour post.ence–have a substantial amount of ancestry from Asia Minor (or Anatolia), which has nothing to do with Turkic peoples and the Turkic ancestry of modern day Turks.”

    “notoriousyid ” is a very insulting and racist screenname, BTW.
    As to your post:

    What are the genetics of the people that you call “Sicilians?”
    Most people would say that they’re highly ancient Greek admixed.
    Ancient Greek populations were all over ancient Anatolia in the region of NE Turkey that Dr Elhaik
    has identified as ancient Ashkenaz….

    How can you arrogantly talk about the genetics of the populations of ancient Khazaria?
    Go you have Y-dna, mtDNA or autosomal dna data of the multi-ethnic and extinct inhabitants of ancient Khazaria?
    Do you have ANCIENT Israelite dna to compare with MODERN populations?

    Many studies have opined on the closeness genetically of MODERN Jews to NON Semitic peoples of the Near East.
    Ariella Oppenheim, et al:
    “In comparison with data available from other relevant populations in the region, Jews were found to be more closely related to groups in the north of the Fertile Crescent (Kurds, Turks, and Armenians) than to their Arab neighbors.”

    1. How is “notoriousyid” offensive? “Yid” means Jew in Yiddish, the language that my Ashkenazi Jewish ancestors spoke for hundreds of years. I am a Jew. It’s clearly a play on the name of rapper The Notorious BIG. Anyway, it’s an old username and that’s beside the point.

      Yes, Sicilians do partly descend from Aegean Greeks. While Greek-speaking peoples did inhabit Anatolia, mostly along the coastal regions, those that lived and continue to live in Northern Turkey, Pontic Greeks, are genetically completely different from Aegean Greeks, Sicilians, and Western Jews. Other peoples from NE Turkey also don’t cluster anywhere near Ashkenazi Jews or Sicilians.

      We do have some Khazar remains and in no way can they be directly connected to Ashkenazi or other Jews.

      We don’t yet have [publicly-available] ancient Israelite DNA, but we have modern Jewish populations that lived in all ends of the world. The fact that the vast majority cluster together and in only TWO distinct groups (East and West) that correspond to distinct periods in Jewish history (Babylonian Exile and Roman diaspora/exile) is more than enough evidence that these groups are all related and descend from a common population. The fact that these groups all cluster near the Levant reinforces this. Ashkenazi Jews, Sephardi Jews, Italki Jews, Romaniote Jews, and North African Jews ALL cluster near each other. On the other side, Iraqi Jews, Iranian Jews, Georgian Jews, Kurdish Jews, and Uzbeki Jews ALL cluster near each other. Syrian Jews, incidentally, seem to cluster in between the two, though closer to Romaniote and other Western Jews.
      And, with the exception of some neighboring non-Jewish groups that are already near the Levant, all of these groups cluster closer to each other than to neighboring populations.

      The Oppenheim study, aside from being 17 years old, in no way supports any of Elhaik’s claims. For one, it very clearly notes the close relationship between ALL Jewish population, which Elhaik consistently ignores or dismisses.

      And to quote a few points in the study itself:

      “It is worth mentioning that, on the basis of protein polymorphisms, most Jewish populations cluster very closely with Iraqis and that the latter, in turn, cluster very closely with Kurds. These findings are consistent with known cultural links that existed among populations in the Fertile Crescent in early history.”

      “We propose that the Y chromosomes in Palestinian Arabs and Bedouin represent, to a large extent, early lineages derived from the Neolithic inhabitants of the area and additional lineages from more-recent population movements. The early lineages are part of the common chromosome pool shared with Jews.”

      “These haplotypes and their one-step microsatellite neighbors constitute a substantial portion of the total Palestinian (29%) and Bedouin (37.5%) Y chromosome pools and were not found in any of the non-Arab populations in the present study. The peripheral position of the modal haplotypes…suggests that the Arab-specific chromosomes are a result of recent gene flow. Historical records describe tribal migrations from Arabia to the southern Levant in the Byzantine period, migrations that reached their climax with the Muslim conquest 633–640 a.d.”

      Anyway, there are more recent studies that completely contradict every single one of Elhaik’s theories and claims.

      “Running RFMix on the AJ genomes with our EU and ME reference panels and summing up the lengths of all tracts assigned to each ancestry, the genome-wide ancestry was ≈53% EU and ≈47% ME…”

      “For AJ, we found that Southern Europe was the most likely EU source for the largest proportion of the AJ chromosomes. Specifically, 43.2% of the AJ chromosomes had Southern EU as their most likely source, 35.4% had Western EU, and 18.8% had Eastern EU…”

      “We observed that in simulations of admixed genomes, the Middle-Eastern regional source could have also been recovered by running the same localization pipeline. Applying that pipeline to the AJ genomes, we identified Levant as the most likely ME source: the proportions of chromosomes classified as Levantine was 51.6%, compared to 21.7% and 22.2% classified as Druze and Southern ME, respectively.”

      “The best match to the AJ data was obtained (in both cases) when the Levant ancestry was almost entirely exclusive (45% out of the total 50% ME ancestry; the magnitude of the minor components was close to zero also when we simulated 50% Southern EU ancestry). This result supports a predominantly Levantine origin for the ME ancestry in AJ, and justifies using the Levantine genomes for the ME ancestry in our simulations.”

      “The availability of dense genotypes for hundreds of AJ individuals, along with the development of new analysis tools, demonstrated genetic relatedness between AJ and other Jewish groups, and suggested Europe and the Middle-East as putative ancestral sources.”

      “Hypotheses such as a wholly Khazar, Turkish, or Middle-Eastern origin have been disqualified, but even a model of a single Middle-Eastern and European admixture event cannot account for all of our observations.”

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 )

Facebook photo

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

Connecting to %s

%d bloggers like this: