Erfurt paper revisited in G25

[Rodoorn]

Yesterday I deleted a posting because Michal pointed me to a correct version of DA161. My bad. Hereby corrected.

Erfurt paper and G25 modeling
In the Erfurt paper they took contemporary Russians and Northern Italians in the modeling. I take a different road. With the starting point of not taking contemporary samples but acute samples. And I take original samples and no proxies. I have further maintained the division of the paper into Erfurt ME- Rhineland Ashkenazim- and Erfurt EU- Central-East Euro Ashkenazim.

As a basic model I have chosen:
- Israel IA: as the cradle of the Jews;
- Krakauer Berg: East German samples from the fourteenth century, as well as Erfurt;
- Khazar: DA161, as a possible input into the Ashkenazim admixture.*

This has the following result:

*Erfurt ME (cor with thanks to De Paris 4*)


*Erfurt EU 


As an addition I have included two samples from the Roman period, one from the NW from Germany and one from the SE from Greece.

This has the following result:

*Erfurt ME 

*Erfurt EU 


My preliminary conclusions
In short my conclusions are based on these G25 models:
- Although the mutual distance between Erfurt ME and Erfurt EU is small in G25 (0.050) they show mutual differences in admixture, they cannot be completely equated to each other.
- In Erfurt EU at least the local Slavic component and the Khazar (DA161) component are striking.
- This indicates a certain independent development of Erfurt EU, in which Jews from the Byzantium and Abbasid Caliphate could also have occurred directly without the intervention c.q. intermediation of the Rhenish Ashkenazim.


*PS the DA161 was a sample from the Khazarian empire time and place, it most probably belonged to the Alan subpopulation.

[DeParis]

Can you share the coordinates that you used to make these models? They are kind of strange. I tried recreating your three-way models in Vahaduo and the results look very different, the IA Israelite is uniformly higher. Are you mixing scaled and unscaled coordinates? 

Erfurt ME:

   

Erfurt EU:

   

There's a third Erfurt group that is somewhat intermediate between ME and EU:

   

[Rodoorn]

Can you share the coordinates that you used to make these models? They are kind of strange. I tried recreating your three-way models in Vahaduo and the results look very different, the IA Israelite is uniformly higher. Are you mixing scaled and unscaled coordinates? 

Of course:
*Erfurt,
See,
https://www.eupedia.com/forum/threads/ge...ost-649214
(Original Davidski)

*Krakauer Berg,
See:
https://drive.google.com/file/d/1rhUzN7A...MHyBt/view
(Original Davidski)

*Khazar (Alan)
DA161.SG,0.088782,-0.01828,-0.003394,-0.028747,-0.059088,-0.006136,0.002585,-0.007846,-0.041518,-0.000729,0.008607,0.007643,-0.013528,0.002615,0.003664,-0.011005,-0.014342,-0.004814,-0.000503,0.004377,-0.003369,0.006801,-0.007395,-0.021208,-0.0122143721
(via collectioneur Michalis)

*Roman Greek en Roman Germany
Germany_Roman.SG,0.121791,0.152329,0.048271,0.003553,0.05201,0.005857,0.00094,0.000462,0.023929,0.026789,-0.004384,0.010041,-0.015609,-0.012248,-0.000543,-0.001591,0.008084,0.004941,0.001006,-0.002001,0.004243,0.002844,0,-0.003253,-0.003233
Greece_Roman,0.104717,0.145221,-0.027907,-0.064277,0,-0.024821,-0.001645,-0.002308,-0.000409,0.007654,0.012017,0.005395,0.005649,0.001239,-0.008007,0.011403,0.02034,-0.003801,0.006788,0.003126,-0.005366,-0.004328,0.002958,-0.000964,-0.006826
via Davdiski files:
https://drive.google.com/file/d/1F2rKEVt...mA-sl/view

*Israel IA
Israel_IA,0.084229,0.147252,-0.063356,-0.09367,-0.016311,-0.047969,-0.008695,-0.003923,0.01084,0.009294,0.008607,-0.005995,0.022596,0.005367,-0.008279,0.001458,-0.024773,0.006081,0.007416,-0.00988,0.003119,-0.002844,-0.002958,-0.001566,-0.005748
Same file as the Roman ones.

All scaled.

[Rodoorn]

They are kind of strange. 

Thanks!

Agree something must have gone wrong, excuse, replaced it with the better figures.

Doesn't affect the tentative conclusions......

[Michalis Moriakos]

The point of me telling you DA161 isn't going to be representative of Khazars was to convince you to no longer use it in your model. If you want to model Jews with actual Khazars you're going to have to wait for real ones to show up in the ancient DNA record. Who knows how long that'll be. Hopefully we won't be waiting too long but I will bet my life that they are not going to look like this heavily Caucasian guy. The closest thing you're going to get to an actual ethnic Khazar average (as in, real Turkic Khazars from Central Asia-- not their subjects) is something like the Karakalpak and Nogai averages. Here are two appropriate averages if you want to use them as proxies instead of DA161:

Karakalpak,0.069312368,-0.18268842,0.040927474,-0.005746,-0.034694684,-0.01095,0.011045474,0.011076368,-0.0061464211,-0.0037118421,-0.019990789,-0.0022559474,0.0010248421,-0.0024626842,0.0019071579,0.0012979474,-0.00056973684,-0.00080684211,0.000999,0.0069573158,-0.011795053,-0.0041652105,-0.0086728421,0.0013572632,-0.00013221053
Nogai_Astrakhan,0.0756925,-0.176025,0.051477,-0.004522,-0.025389167,-0.012782667,0.0072068333,0.0142685,-0.0077376667,-0.0034016667,-0.019540667,-0.0026225,0.0035925,-0.0032798333,0.0026918333,0.0007955,0.0019775,-0.0014358333,0.0006495,0.0048773333,-0.0084018333,-0.0015663333,-0.01253,-0.00010033333,0.00081833333

You can't use an idiosyncratic outlier with very high Near Eastern ancestry and expect it to be a meaningful marker of typical Khazar ancestry in Jews. Of course Jews are going to score DA161-like ancestry if that individual is your source-- Jews have a lot of West Asian ancestry after all. Also, DA161 could just as well be from the Alan period anyway-- not necessarily even a subject of the Khazar Khaganate, let alone a Khazar.

[Rodoorn]

The point of me telling you DA161 isn't going to be representative of Khazars was to convince you to no longer use it in your model. If you want to model Jews with actual Khazars you're going to have to wait for real ones to show up in the ancient DNA record. Who knows how long that'll be. Hopefully we won't be waiting too long but I will bet my life that they are not going to look like this heavily Caucasian guy. The closest thing you're going to get to an actual ethnic Khazar average (as in, real Turkic Khazars from Central Asia-- not their subjects) is something like the Karakalpak and Nogai averages. Here are two appropriate averages if you want to use them as proxies instead of DA161:

Karakalpak,0.069312368,-0.18268842,0.040927474,-0.005746,-0.034694684,-0.01095,0.011045474,0.011076368,-0.0061464211,-0.0037118421,-0.019990789,-0.0022559474,0.0010248421,-0.0024626842,0.0019071579,0.0012979474,-0.00056973684,-0.00080684211,0.000999,0.0069573158,-0.011795053,-0.0041652105,-0.0086728421,0.0013572632,-0.00013221053
Nogai_Astrakhan,0.0756925,-0.176025,0.051477,-0.004522,-0.025389167,-0.012782667,0.0072068333,0.0142685,-0.0077376667,-0.0034016667,-0.019540667,-0.0026225,0.0035925,-0.0032798333,0.0026918333,0.0007955,0.0019775,-0.0014358333,0.0006495,0.0048773333,-0.0084018333,-0.0015663333,-0.01253,-0.00010033333,0.00081833333

You can't use an idiosyncratic outlier with very high Near Eastern ancestry and expect it to be a meaningful marker of typical Khazar ancestry in Jews. Of course Jews are going to score DA161-like ancestry if that individual is your source-- Jews have a lot of West Asian ancestry after all. Also, DA161 could just as well be from the Alan period anyway-- not necessarily even a subject of the Khazar Khaganate, let alone a Khazar.

Now I must disagree with you my friend. I use the sample in a proper way, not as a representative of the Khazarians still as originated from the Khazarian empire, in time and place. That is a 100% proper use.

I could use Khazarian (Alan) in stead to show that it most probably a sub population- the Khazarian were not one size fits all- at stake:
https://ojs.bibl.u-szeged.hu/index.php/c...0681/11150

"You can't use an idiosyncratic outlier," it is what is Michal, in the cases of samples you can't use it as it 'ought to be', 'idiosyncratic outlier' is an assumption, that would be absurd imo. The Khazarian empire was a multiethnic state!

Khazar Khaganate, 650–850, DA161 wit a red star.

[Rodoorn]

I added Karakalpak, Nogai_Astrakhan beneath a 'Khazar' label and gathered all Israel IA also those of Ashkelon IA.

Result:

*Erfurt ME


*Erfurt EU



This sharpens the conclusions of the initial posting (thanks DeParis, Michalis!):

- Erfurt ME has a definite higher Israel IA and has a bigger chunk Roman Germany, so South European like;

- Erfurt EU had a definite higher 'Khazar' and Slavic ME.

[Rodoorn]

In fact the results are at ease with following trajectory by Elhaik et al:

This could confirm the right trajectory of Erfurt EU with relative high results for DA161-an Alan- because Alans they are primal Iranian.

[Rodoorn]

In fact the results are at ease with following trajectory by Elhaik et al:

This could confirm the right trajectory of Erfurt EU with relative high results for DA161-an Alan- because Alans they are primal Iranian.

Witt an add to this trajectory derived from Dovid Katz, in which it's not exclusive Sorbian, but more integrated in the broader Central East European context, nevertheless still Judeo Aramaic which is the language of the Babylonian Talmud.