Last week I received an email from ATAB, a Turkish-born German who took a genetic test with one of the genetic companies and was not very happy with the results. He was informed that he was 25% Sephardic and 1.8% Ashkenazic Jew, and provided that none of his ancestors are Jews, this is quite a substantial and unexplained genetic heritage. Now, a typical counter argument would sound like this: He is a Jew but doesn’t know it, another lost sheep, welcome to the tent’s tribe Jacob! As we shall see soon, this is hardly the case here. I explained that some tests are less reliable than others and that those that report Jewish (Elhaik 2016) and Spanish ancestry are the worst since they ignore the actual origin of these Jews and simply assume that religious practices are the same as genetic differences. None of the other results made sense to him either.
He read about our Das et al. (2016) study and asked for information about this region.
ATAB’s family history in Turkey goes back over 400 years. His great maternal grandfather lived around 1600 and was a “Katip” at the Ottoman Empire, a writer in the office. The family name came from Katipoglu (today Yazicioglu or Yazici). His maternal grandmother was called the “Mariel or Maryeller” – the maryels line. His paternal line came from Maras, Turkey some hundred years ago. The maternal line of that lineage came from Konya. Some relatives still live there. His maternal and paternal family still live in Trabzon – Caykara – Hopsera village. The meaning of this name is “widow”. In terms of language, all the older generations spoke or still speak the pontic “romeyka” language. All the villages and districts in Trabzon-Caykara have old romeykan and modern Turkish village names. He noted that the “Romeyka” language is spoken by ~5,000 people worldwide. The name derived from “Rums” – those were the Greeks living in Ottoman Empire (they were now always called Greeks).
I suggested that he would take the GPS Origins test that we built (COI: I consult DDC). GPS Origins traced his two parental lines to a site only few kilometers from where his entire family was born. The two migration lines starting at Southern Russia and Eastern Turkey meet at the border of Turkey-Armenia with his family’s origin nested between them.
Obtained from ATAB’s GPS Origins results (birthplace annotation was added manually)
The red line was dated 763 AD – 1434 AD and the blue line 1213 AD – 1591 AD. We speculated that Ashkenazic Jews (AJs) left the region by the 7th century with the movement of the Arabs toward the Near East and moved north where they influenced the Judaization of the Khazarian Empire. However, ATAB’s genomic signature was formed shortly afterwards, so GPS Origins interpreted his results in light of the Muslim invasion, not the Jewish chapter that has just ended. How to interpret these results?
In Das et al. (2016) we applied GPS to the genomes of AJs. GPS localized them to major primeval trade routes in northeastern Turkey adjacent to primeval villages with names that may be derived from “Ashkenaz.” We termed this region “ancient Ashkenaz.”
Our results indicated that “Ashkenaz” is not a synonymous with “Jew” but rather a placename that was associated with Armenia and Eastern Turkey well before the emergence of Ashkenazic (and possibly Sephardic) Judaism. The Armenians considered themselves descendants of Ashkenaz and Togarmah. Ashkenazic Jews are Jews from Ashkenaz, similarly to Iranian Jews, Moroccan Jews, and any other group. Had AJs emerged from the local population in “ancient Ashkenaz,” then their neighbors would have the same genetic makeup. Therefore, a strong support to our theory would be finding Ashkenazic non-Jews. That is, people with the same genetic makeup as AJs, only non-Jews. Could ATAB be one of them?
A first glance of his 36 gene pools, identified by GPS Origins, confirmed that the gene pool proportions are remarkably similar to those of AJs. But looks can be deceiving, particularly with 36 data points. To visualize the similarity between ATAB and AJs, I used the graph analysis that we used in several previous papers. In this analysis we calculate the linear distance between the 36 gene pools of every person to everyone else. Arcs connect every two people that are highly similar, provided some threshold.
Genetic distances between the gene pools of ATAB and reference populations
The genetic similarity between ATAB (red dot) and AJs (green) with a reasonable threshold of 0.08 is shown above. As expected, AJs formed a cloud, but it is separated from that of Turks. ATAB and several other Turks from Trebizond and the surrounding are the closest to the AJ cloud. I was wondering which gene pools were off, so I compared how ATAB’s gene pools fall within the AJ range. Here is the list:
- Southeastern India – ATAB has 14%, AJs have 3-10.5%. This probably was the major factor for the observed difference.
- North western Africa – all AJs have a <1%. ATAB has nothing.
- Arabian – ATAB is 1%. AJs are less than 0.7%.
- Eastern Mediterranean – ATAB has 0.7%, AJs have 0-0.5%.
- East China – ATAB has 9.8% and AJs have less than 0.82%
Indeed, ATAB’s Southeastern Indian component is the most dominant one.
Obtained from ATAB’s GPS Origins results
This and other components are of Indian origin and exist in all Indo-Iranian speakers to various degrees. This component ranges from 0 to 17% in Turks with an average of 8%, so ATAB is in the highest percentile with other Turks from Trabzon and Aydin (both have a strong Greco-Roman history). This is in agreement with the nearly extinct language his family speaks. AJs have 0.3% – 10.6% of this component. Did they emerge from this population and halved that gene pool following mixtures with other populations? Had the local population mixed with other populations after the departure of AJs? Provided that Greco-Italian populations have 0-10% (mean of 7%) of this component, the second explanation is more likely. In Armenians this component is 7-30%! So let us add some Armenians into the mix.
Genetic distances between the gene pools of ATAB and reference populations
The black cloud is the Armenians. Zooming in, these results show the mixed Armenian – Turkish ancestry that GPS Origins identified since ATAB sits right in the midst of this Armeno-Turkish cloud.
Genetic distances between the gene pools of ATAB, Armenians, and Turks
ATAB asked if he is AJ, not an atypical question. My findings and opinion on the matter were summarized in (Elhaik 2016). Genetic Jewishness tests are fabrications, alternative facts if you will. They are not real. It is possible, however, to test ATAB’s clustering with other Jewish communities. Using the same exact threshold as before, ATAB clusters nicely with Caucasus and Near Eastern Jews (who cluster with Turks non-Jews). You can see that Turkish Jews are so diverse that they embody AJs and connect them with the Caucasus – Near Eastern cluster.
Genetic distances between the gene pools of ATAB and Eurasian Jews
Let us add some Turks from Trebizond to the mix.
Genetic distances between the gene pools of ATAB, Turks, and Eurasian Jews
The plot is now flipped, but don’t let it confuse you. Interestingly, the Trebizond Turks cluster with both the left (Caucasus\Near Eastern) and right clusters (Morocco, Turkish, and Ashkenazic Jews). ATAB is positioned on the left cluster at the point that reaches to the right one, again, probably due to some Armenian heritage that seem to have affected some, but not all of the Turks from that region.
Can ATAB be considered an Ashkenazic non-Jew? That depends on your definition. He is clearly very close to the cloud dominated by AJs, but 1000 years of uneven mixing may have shifted the gene pools just enough to create some differences. Nonetheless, these results do support our hypothesis on the origin of Ashkenazic Jews in Ancient Ashkenaz. If results from people like ATAB would become available, the gap between the two clusters would no doubt be filled with Northeastern Turks whose stories would help us better understand the event that took place in the region past the 6th century. I would be very curious to see the results of his family, particularly those along the red line.
Finally, a word of caution. These data are from different sources. As such, they were subjected to different biases, which can easily account for the observed differences.
Happy DNA day.
Watch a neat GPS Origins video
Das, R., P. Wexler, M. Pirooznia, and E. Elhaik. 2016. Localizing Ashkenazic Jews to primeval villages in the ancient Iranian lands of Ashkenaz. Genome Biology and Evolution. 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”. Frontiers in genetics. 7.