Modelling Admixture with D-stats

Fire Haired14

Banned
Messages
2,185
Reaction score
582
Points
0
Y-DNA haplogroup
R1b DF27*
mtDNA haplogroup
U5b2a2b1
I'm creating this thread to post admixture results using D-stats, instead of posting these results at Post Formal States Here!!. D-stat results can be used to get ancestry percentages for populations. What you do, is model a population's D-stat results in many differnt tests a mixture of another population's D-stat results in the same stats.
 
These results are so accurate it's scary. I didn't manipulate the data :). Modelling Europeans as Yamnaya+HungaryNeolithic+CHG+WHG.

South EuropeEN_HungaryYamnayaWHGCHG@ D
Sardinian89%1%6%4%0.0045
Tuscan71%21%2%6%0.0056
Bulgarian62%30%2%6%0.0055
BasqueSpain61%18%18%3%0.002
WestSicily77%12%0%11%0.0148
North EuropeEN_HungaryYamnayaWHGCHG@ D
Lithuanian36%44%20%0%0.0031
BeloRussian43%44%13%0%0.0064
Cornwall_England45%44%11%0%0.002
Hungarian53%39%8%0%0.0056
French56%34%9%1%0.0033
Ancient EuropeEN_HungaryYamnayaWHGCHG@ D
Spain_MN76%0%24%0%0.0157
Sintashta36%55%5%4%0.0075
HungaryBA46%28%22%4%0.0049
 
Fwiw, these numbers make general sense to me, mainly because they track with what the academic papers found, which is close to 50% Yamnaya replacement in northern and central Europe and quite a bit less in southern Europe. The ENF numbers also track with prior work. As for WHG that also makes sense to me, as they were a very small population, and I think their impact was less.

I do think that the admixing population with Yamnaya might have been slightly different depending on the area. Perhaps for Italy it would be Remedello? I don't know how much of a difference that would make, however. Perhaps not a large one.

If you have the time, could you run your program for Northern Italians? Thanks.

Btw, what outgroups did you use?
 
I think that it is somewhat misleading as Yamnaya is a composite population and already includes EN farmers, WHG and CHG. Perhaps we should only compare Yamnaya to other contemporary composites like Late Neolithic European? This way we should truly see impact of Yamnaya in Europe.
 
I think that it is somewhat misleading as Yamnaya is a composite population and already includes EN farmers, WHG and CHG. Perhaps we should only compare Yamnaya to other contemporary composites like Late Neolithic European? This way we should truly see impact of Yamnaya in Europe.

I'm not trying to get distant Paleolithic ancestry percentages. I'm modelling with the best proxies we have. EHG and CHG didn't migrate into Europe seperatly. They came together in populations like Yamnaya.
 
Modelling Jews

I modeled three differnt Jewish populations with the outgroups: MA1, Loschbour, LBK, and Kotais. All fit very well, as a mixture of their host populations, and Druze.

Host populations are: Yemen, Georgian, and BeloRussian.


DruzeHost Population
YemanJew67%32%0.0058
Ashkenazi67%32%0.0096
Georgian Jew54%46%0.0054

The weird thing to me, is Ashkenazi Jews pass so easily pass as white. But DNA doesn't lie. If they were mostly converts from Central/East Europe, there's a 0.00000000...1% chance they'd be getting the results they're getting.
 
The weird thing to me, is Ashkenazi Jews pass so easily pass as white. But DNA doesn't lie. If they were mostly converts from Central/East Europe, there's a 0.00000000...1% chance they'd be getting the results they're getting.

Gross had a hypothesis, that mostly "Aryan-looking" Jews survived the Holocaust (they had a greater chance to pass as Non-Jews). However, the vast majority of Jewish-Americans descend from Eastern & Central European Jews who emigrated to America between 1800 and 1939, before the Holocaust. So this hypothesis is baseless, at least as far as American Jews are concerned.

Check this thread - Eran Elhaik (who used to be a supporter of the so called Khazar hypothesis), now has a new theory, about Eastern Anatolian origins of Ashkenazi Jews:
http://www.eupedia.com/forum/threads/32082-Guess-the-ethnicity-of-this-lady?p=476359#post476359

Proposed homeland:

Aszkenaz.jpg


According to a Polish Jew, Henryk Szpidbaum, the most common phenotype among Polish Jews was Armenoid (or Hither Asiatic or Anatolid). This hypothetical Eastern Anatolian homeland corresponds well to prevalence of Armenoid phenotype.

But Armenoid / Anatolid / Hither Asiatic phenotype is common in all of West Asia, not just in Eastern Anatolia.

A population of Ancient Levantine origin could also have a high percentage of this phenotype.

========================

As for light skin of Ashkenazi Jews:

Maybe in Ancient times, before Arabic expansion and Muslim Sub-Saharan slave trade, Middle Easterners were "whiter"? Some Non-Muslim populations of the Middle East still tend to be light, even if they never moved out of the region (e.g. Samaritans).

Or, alternatively, it could be that Ancient Jews had swarthy skin pigmentation until relatively late in time, but after that the same selective sweep favouring light skin tone made them light, which had made everybody else in Europe white before.
 
Maybe already in Early Medieval pogroms, swarthy Jews were more likely to die than white-skinned ones ???

==================

As for the Khazar hypothesis - we actually do have Khazar ancient DNA already:

Saltovo-Mayaki culture (800-900 AD; associated with the Khazar Kaganate):

Burial site, individual(s) - haplogroups:

Podgorovsky, sample A80301 - R1a1a1b2a (Y-DNA), I4a (mtDNA)
Podgorovsky, sample A80302 - D4m2 (mtDNA)
Podgorovsky, sample A80410 - G (Y-DNA)
Podgorovsky, sample A80411 - J2a (Y-DNA)
Aksenov & Demetrius, six samples - 6 x G2a (Y-DNA)

Sources:

- Afanasiev, G.E., M. V. Dobrovolskaya et al. (2015)
- Afanasiev, G.E., S. Wen et al. (2015)
- http://www.ancestraljourneys.org/medievaldna.shtml

Compare that with modern Ashkenazi Jewish Y-DNA lineages:

http://jewishdna.net

As for those ten Saltovo-Mayaki samples:

The problem is that we don't know if those were ethnic Khazars, or other inhabitants of the Khazar realm:

https://en.wikipedia.org/wiki/Saltovo-Mayaki

"Saltovo-Mayaki is the name given by archaeologists to the early medieval culture of the Pontic steppe region roughly between the Don and the Dnieper Rivers. Their culture was a melting pot of Onogur, Khazar, Pecheneg, Magyar, Alan, and Slavic influences. During the ninth century the Saltovo-Mayaki culture was closely associated with the Khazar Khaganate, and archaeological sites from this period are one way that historians track the geographic scope of Khazar influence."

==================

It would be nice to obtain aDNA samples from these Medieval graves at Worms, Germany:

http://www.eupedia.com/forum/thread...azi-Jewish-DNA?p=475572&viewfull=1#post475572
 
I modeled three differnt Jewish populations with the outgroups: MA1, Loschbour, LBK, and Kotais. All fit very well, as a mixture of their host populations, and Druze.

Host populations are: Yemen, Georgian, and BeloRussian.


DruzeHost Population
YemanJew67%32%0.0058
Ashkenazi67%32%0.0096
Georgian Jew54%46%0.0054

The weird thing to me, is Ashkenazi Jews pass so easily pass as white. But DNA doesn't lie. If they were mostly converts from Central/East Europe, there's a 0.00000000...1% chance they'd be getting the results they're getting.
Yemeni Jews are actually quite similar to non Jewish Yemenis, Saudis and Bedouin.
 
Yemeni Jews are actually quite similar to non Jewish Yemenis, Saudis and Bedouin.

Indeed.

The reason the first set of data made some sort of sense is because, I'm assuming, it was based on ancient genomes. As soon as you start using modern populations things start to go haywire, and that's the case whether you're doing Admixture based runs, no matter how carefully done, or formal stats.

Just as an example, I'm not sure about using the Druze for these purposes, as although they may not have admixed in the last 1000 years, they may or may not be a good proxy for the Jews of the Roman Era diaspora who are probably the major source population for Ashkenazim. The Druze have a rather unique ethnogenesis, not all of which derives from the Levant.

On some PCAs, the Syrians are closest.
pc-jews.jpg


It's true that Behar found that the most IBD sharing for Ashkenazim other than with other Jews was with Eastern Europeans, but that study is a bit old, and only was able to track rather large segments, which of course would be the most recent ones from their sojourn of about 1000 years in Eastern Europe. Even then, I don't think Behar sees the admixture as totalling one third of their genome.
IBD sharing with Ashkenazi.jpg

http://digitalcommons.wayne.edu/cgi/viewcontent.cgi?article=1040&context=humbiol_preprints

Other analyses have shown anywhere from about 35% to about 50% "European" admixture.


There are all sorts of problems with this approach. "European" is a cultural and geographical construct. "Europeans" come in various flavors. Some of these studies just willy, nilly decide West Europeans are the "real" Europeans, and model accordingly. That's the major problem with Shai Carmi's paper, in my opinion. I mean, really, Flanders?
figure4.png




If you model the admixture with "Southern Europeans" you're going to get a different breakdown.
European_Gene_Flow_into_Ashkenazi_Jews_Fig_3.png


The only way we're going to really know how much "admixture" there was between Jewish diaspora populations in the Near East with "Europeans" in Europe is to get an ancient genome. Then, one from the Rhineland around the time of the Ashkenazi ethnogenesis would be good.

My hunch is that the Jewish diaspora populations from places like Anatolia, who may have already mixed with Greeks or Hellenized Anatolians, were already pretty similar to Southern Europeans, and so there actually wasn't a whole lot of admixture in Europe proper, but time will tell.
 
I'm aware that some people divide humanity into three or more "races", one of which is "Caucasian". I'm not aware of a scientific term called the "White" race. That's a mainly subjective term often used by Americans who don't know much about genetics.

As for the Druse, they have some rather stereotypical "Near Eastern" looking people, but some of them are not all that far removed from Europeans. Perhaps that's because they have had minimal impact from the Arabic invasions and any slave admixture from either West Africa or East Africa. It may also have something to do with their specific ethnogenesis. I don't know.I also don't know, as I mentioned, whether the Druse are all that good a sample for Roman Era diaspora Jews.

However, it's as well to be clear that a good number of them are hardly a whole different "race" even from Central Europeans.

druze%20girl%20and%20mumsie.jpg


Fattan-300x178.png




SH_Tabou.jpg


There's quite a bit of variety:
477132728-israeli-druzes-take-part-in-a-demonstration-gettyimages.jpg


Here are a Druze leader and Netanyahu. They don't look all that different to me, other than a bit of lightening in Netanyahu.
20141713162069580_20.jpg
 
There's also the Syrian Alawites, who, as farmers in an isolated area holding to a minority religion have also not experienced the full brunt of the Arabic invasions. There's quite a bit of variety in them as well:

155029-aef68396-0ed9-11e3-92bf-1ac22055d80e.jpg


She could fit anywhere in Europe. Usually, he looks pretty Near Eastern, but in this picture he has a bit of a Spanish look to me.


The Al-Assad family. The former president looks European to me, the son a little less so, but not much in this picture. In fact, here he's a dead ringer for my hairdresser, who comes from a Bulgarian Muslim family.
120321082058-assad-family-photo-story-top.jpg



Another Syrian president and his family. There may be a class difference at play here, with the lower classes having a slightly different ethnogenesis.

President_Shukri_al-Quwatli_and_his_family_-_Beirut,_1966.jpeg


There are definitely some stereotypically Near Eastern looking Alawites. Still, even these Alawite militiamen look closer to Europeans than other Syrians do, in my opinion.
Shabiha_2244515b.jpg



Carlos Menem, former President of Argentina, is of Syrian origin: In his younger, less weird persona, I don't think I would have known he wasn't just a run of the mill Argentinian.
1579e2939253ba75bd0b7bac6c67a299289d372b.jpg


Steve Jobs' biological father: He actually looks quite Ashkenazi to me.


steve-jobs-jandali-father_thumb800.jpg
 
@John Doe, Angela,

It's very clear Yemen Jews are differnt from Yemens, when ancient West Eurasians are much closer to Yemen Jews than to Yemens. This is because Yemens have a lot more African admixture. Yemen Jews behave in these tests most like SW Asians.
 
@John Doe, Angela,

It's very clear Yemen Jews are differnt from Yemens, when ancient West Eurasians are much closer to Yemen Jews than to Yemens. This is because Yemens have a lot more African admixture. Yemen Jews behave in these tests most like SW Asians.

As I mentioned upthread, imo results using modern populations are far less reliable than the ones you generated using ancient samples.

In addition, no one said or I'm sure meant to imply that the Yemeni Jews aren't somewhat different from Yemenis. There are still a lot of similarities.

They actually seem to cluster with the southernmost Palestinians and just northeast of the Saudis.

This is from Behar:
jewsnat2.png


It's important for all of us to be very precise with language and terms, me included.
 
As I mentioned upthread, imo results using modern populations are far less reliable than the ones you generated using ancient samples.

You're right except, because of the Outgroups I'm using Druze or Cypriot or anyone who is SW Asian but not EEF, is a good proxy for ancient Jews. My outgroups are: EEF, WHG, EHG, CHG, and MA1. All the outgroups have a pretty equal relationship to SW Asians, except MA1 has a much smaller one and EEF a much bigger one.

Intra-SW Asian diversity doesn't make an affect at all because of my outgroups. So, a Jew from 1000 BC and a modern Druze, will behave the same way in the tests with my outgroups. Unless, SW Asians in 1000 BC had lots less African admixture or were somehow super EEF-like, both of which are unlikely.

Ashkenazi Jew's fit as 90% EEF in Laz 2014 because they have a similar relationship to the non-West Eurasian outgroups they used as does EEF. The reason for this is probably shared basal Eurasian. So, CHG could be fit as 90% EEF, if we used only East Asian outgroups. However, me using EEF and WHG as outgroups, reveals Jew's relationship to EEF and WHG is too small for them to be mostly EEF. Instead they have a lot of decent from people with a similar amount of Basal Eurasian, but weren't EEF and had little or no WHG. The only people who fit the bill are SW Asians.

It's the same case with Yemen and Georgian Jews. Ancient West Eurasians are too close to Yemen Jews for them to have the high amount of African admixture Yemens do. CHG is too distant from Georgian Jews for them to have the same high amount of CHG Georgians do.

In addition, no one said or I'm sure meant to imply that the Yemeni Jews aren't somewhat different from Yemenis. There are still a lot of similarities.

They actually seem to cluster with the southernmost Palestinians and just northeast of the Saudis.

The Yemen Jews I have D-stats for are not very similar to Yemens. They're much more similar to Druze.
 
Fire-Haired:
The Yemen Jews I have D-stats for are not very similar to Yemens. They're much more similar to Druze.

From what I can remember, Yemeni Jews are about 5% "African", all of it East African. The other Yemenis have much more "African", over 15%, I think, and roughly half of it is West African. The Druze, on the other hand, are about 1.5 % East African in the same kind of analyses. I think that's what moves the Yemeni Jews away from the other Yemenis, given the divergent nature of "African" alleles compared to Eurasian ones. That places them close to Palestinians, especially Gaza type Palestinians rather than the Druse, in my opinion, who have a lot more northern Near Eastern and even Iranian ancestry is their ethnogenesis stories are to be believed. The Druse, in turn, are closer to Cypriots. It's a mistake to think all "Southwest Asians" are the same genetically either now or in the past.

The differences show up very strongly in phenotype.

Druse:
Druze_Lebanon_pic_1.jpg


640x392_42577_79196.jpg


It's not just pigmentation; it's head structure and facial structure and features.

Fire-Haired. Unless, SW Asians in 1000 BC had lots less African admixture or were somehow super EEF-like, both of which are unlikely.

How can it be unlikely given the known migration of entire Arabic tribes into the Middle East post the Islamic conquests? We're supposed to believe, according to Hellenthal and Busby and company that enough North Africans and Middle Easterners came into France in a forty year period (before being beaten back) to account for 20% of the total genome of Gaul, but whole tribes permanently settling in more northern areas of the Middle East didn't have any impact?

With those invasions came the Arab slave trade, where many East African women, in particular, were imported into the Near East.

I don't know how much of an impact they had, and neither does anybody else, but I do think this is part of what explains the difference in the results of isolated populations like the Druse versus Palestinians, for example, and in particular of Gaza type Palestinians and some Jordanians and certainly many Iraquis etc.

That's why using the Druze is a good idea, although maybe using the Samaritans is an even better idea, since there's no question of any outside genetic flow in them for the last 3000 or so years, absent the best idea, which is waiting for an ancient genome. Using Cypriots is not a good idea, in my opinion. We just can't assume that coastal Levantines were the same as modern Cypriots.

Oh, for what it's worth, I think the flow from the direction of the Arabian peninsula, which itself picked up progressively more SSA genes as time went on, was continuous at least since the time of the Nabataens. It just depends when you draw the line.
 
Let's also not forget that most or many of modern Jewish-Americans are not 100% Jewish but mixed with other ethnicities.

Jews used to be endogamous as long as they were Orthodoxly religious, but this has endeded already some generations ago.

For example on another forum I have read posts claiming that some American "Jewish celebs" look "very Non-Jewish", and they listed people such as Gwyneth Paltrow, Paul Newman, etc. as examples. Fortunately there is a website ethnicelebs.com, which clearly says that these people are not fully Jewish. For example Gwyneth Paltrow is listed as ethnically mixed Jewish-German-Irish-English person, and Paul Newman as mixed Slovak-Jewish. Then for example Leelee Sobieski is also mixed Polish-Jewish-Swiss French. And so on.

IIRC, those were posts by a guy who tried to prove that Eran Elhaik is right in his claims that Jews come from Khazars. So he claimed that Gwyneth Paltrow and Paul Newman look "more Khazar than Jewish". :) Do we even know how Khazars looked like?

Why did he expect to find "typically Jewish looks" in people who are just 1/2, 1/4/, 1/8 or even less Jewish? Another example - current First Lady of Poland is 1/4 Jewish and doesn't look much Jewish. Her daughter, 1/8 Jewish, doesn't look Jewish at all.

http://ethnicelebs.com
 
There's also the Syrian Alawites, who, as farmers in an isolated area holding to a minority religion have also not experienced the full brunt of the Arabic invasions. There's quite a bit of variety in them as well:

155029-aef68396-0ed9-11e3-92bf-1ac22055d80e.jpg


She could fit anywhere in Europe. Usually, he looks pretty Near Eastern, but in this picture he has a bit of a Spanish look to me.


She is half British. I think this picture clearly shows it :)

EDIT: Nope, wrong. She a full-blooded Syrian, just raised in England. Good to see that a proper British upraising shows itself in someones looks! Well done, chaps!
 
These results are so accurate it's scary. I didn't manipulate the data :). Modelling Europeans as Yamnaya+HungaryNeolithic+CHG+WHG.

South EuropeEN_HungaryYamnayaWHGCHG@ D
Sardinian89%1%6%4%0.0045
Tuscan71%21%2%6%0.0056
Bulgarian62%30%2%6%0.0055
BasqueSpain61%18%18%3%0.002
WestSicily77%12%0%11%0.0148
North EuropeEN_HungaryYamnayaWHGCHG@ D
Lithuanian36%44%20%0%0.0031
BeloRussian43%44%13%0%0.0064
Cornwall_England45%44%11%0%0.002
Hungarian53%39%8%0%0.0056
French56%34%9%1%0.0033
Ancient EuropeEN_HungaryYamnayaWHGCHG@ D
Spain_MN76%0%24%0%0.0157
Sintashta36%55%5%4%0.0075
HungaryBA46%28%22%4%0.0049

How does that work using other EEF's for EN_Hungary? I am especially curious at Gokhem.
 
Let's also not forget that most or many of modern Jewish-Americans are not 100% Jewish but mixed with other ethnicities.

Jews used to be endogamous as long as they were Orthodoxly religious, but this has endeded already some generations ago.

For example on another forum I have read posts claiming that some American "Jewish celebs" look "very Non-Jewish", and they listed people such as Gwyneth Paltrow, Paul Newman, etc. as examples. Fortunately there is a website ethnicelebs.com, which clearly says that these people are not fully Jewish. For example Gwyneth Paltrow is listed as ethnically mixed Jewish-German-Irish-English person, and Paul Newman as mixed Slovak-Jewish. Then for example Leelee Sobieski is also mixed Polish-Jewish-Swiss French. And so on.

IIRC, those were posts by a guy who tried to prove that Eran Elhaik is right in his claims that Jews come from Khazars. So he claimed that Gwyneth Paltrow and Paul Newman look "more Khazar than Jewish". :) Do we even know how Khazars looked like?

Why did he expect to find "typically Jewish looks" in people who are just 1/2, 1/4/, 1/8 or even less Jewish? Another example - current First Lady of Poland is 1/4 Jewish and doesn't look much Jewish. Her daughter, 1/8 Jewish, doesn't look Jewish at all.

http://ethnicelebs.com


There's a very funny blog on the internet run by an Ashkenazi Jew, or so I've been told, called "Jew or not Jew". :) I love his sensibility, writing style, and wit.
http://www.jewornotjew.com/

There you can learn, should you be interested, which famous people are Jews, half-Jews, quarter-Jews and barely Jews. :)

Goldie Hawn is only half Jewish. So is Scarlett Johannsen, who in addition has had a lot of facial surgery, at least. Jessica Biel is only 1/8 Jewish, which wouldn't have counted even for the Nazis.

goldie-hawn-net-worth.jpg


On the other hand, both Natalie Portman, who I think is gracile Mediterranean, and Lauren Bacall, who I think looks sort of Central European, are 100% Jewish.

Lauren-Bacall_3004391b.jpg


What a woman she was: a stunning face, a lean figure that had a panther's grace, and a husky, sexy growl of a voice. She was intelligent and witty on top of it all. I think she was probably pretty much irresistible; certainly, Humphrey Bogart thought so, throwing over his wife and his prior life for a girl a fraction of his age. Luckily for him she turned out to be a faithful and devoted wife, even through his long bout with lung cancer.

https://www.youtube.com/watch?v=kv2K62fTXIs

So was Bess Myerson, the first Jewish Miss America.

518584643_c_570_411.jpg


So, it's all a roll of the genetic dice.

@Epoch,
Yes, indeed, thank-goodness, because of course that's what everyone should aspire to...looking British or otherwise Northern European, of course. :)
 

This thread has been viewed 43932 times.

Back
Top