Pop finder / Geno2 admixture differences

[khazaria]

Update: there is now a scientific paper about the discovery of the East Asian mtDNA haplogroup M33c in Ashkenazim, based in part on Family Tree DNA samples:

"A genetic contribution from the Far East into Ashkenazi Jews via the ancient Silk Road"
by Jiao-Yang Tian, Hua-Wei Wang, Yu-Chun Li, Wen Zhang, Yong-Gang Yao, Jits van Straten, Martin B. Richards, and Qing-Peng Kong
Published 11 February 2015 in Scientific Reports vol. 5

http://www.nature.com/srep/2015/150211/srep08377/full/srep08377.html

People from multiple ethnic groups (Han, Zhuang, Kam-Tai, Yao, Miao, Tibetan) in China, mostly southern China, have M33c. It is also found in a few non-Chinese people from Vietnam and Thailand, but its greatest genetic diversity is found in China. The table of M33c samples is at http://www.nature.com/srep/2015/1502...p08377_T1.html
and they have a geographical map at

http://www.nature.com/srep/2015/150211/srep08377/fig_tab/srep08377_F1.html

Ted Kandell was right: outside of the Ashkenazic population, which got it from an intermarriage with a non-Jewish woman, M33c isn't found in peoples native to India or anywhere else in the world. China is the place of origin. M33a, M33b, and M33d are different haplogroups.

[Mars]

My Geno 2.0 results are 51% mediterranean, 30% northern european, 18% southwest asian. According to MyOrigins I am 85% european and 16% middle eastern. The 85% euro is 69% southern Europe and 16% Scandinavia, the middle eastern is all Asia Minor. I'm a northern italian.

[Táltos]

At the present moment I don't think such a paper exists. This is still a relatively new discovery. And as I wrote on my page, we need Costa's research team to separate out the names of the haplogroups they grouped together as "M/A". I saw that the East Asian haplogroups have been an afterthought to the more exact delineations of the European and Middle Eastern haplogroups, due to the very small frequencies of M and N haplogroups in Ashkenazim.

Thank you for reminding me that in FTDNA's "M" mtDNA project some other Ashkenazim belong to M33c as opposed to M33c1, but some of those are the same kits I already saw in FTDNA's "Jewish Ukraine West" project.

I no longer think Y-DNA Q1b1a could come from Khazars. The evidence is convincing now that ancestors of the Ashkenazim acquired it in the ancient Middle East. It also seems doubtful that the East Asian mtDNA haplogroups and autosomal relationships have something to do with Khazars.

Thank you for the clarification. And yes the Ashkenazi are primarily from the Middle East, and admixed with Europe. Though I think it is to be expected that some will show small amount of some kind of Asian, as other European populations such as Russians, Ukrainians, and so forth will too.

Also in your link you mentioned that Ashkenazi were not getting any Central Asian in the autosomal. But some fully Ashkenazi that I'm sharing with over on 23andme will get some small bits of South Asian. In the South Asian category they do lump Central Asian such as Uygur and Hazaras. Though I think 23andme is the best of the lot in these admixture tests, sometimes they have odd groupings.

The best information that can be currently found on the various subclades of Q1b is this paper as it takes a serious approach studying the phylogenetic structure of of it. It involves full Y chromosome sequencing. https://www.academia.edu/7117246/The...ome_sequencing

[khazaria]

Can you please provide an actual paper that shows the M33c1 is in fact the same exact subclade of M33c for both the Ashkenazi and South Chinese

At the present moment I don't think such a paper exists. This is still a relatively new discovery. And as I wrote on my page, we need Costa's research team to separate out the names of the haplogroups they grouped together as "M/A". I saw that the East Asian haplogroups have been an afterthought to the more exact delineations of the European and Middle Eastern haplogroups, due to the very small frequencies of M and N haplogroups in Ashkenazim.

Thank you for reminding me that in FTDNA's "M" mtDNA project some other Ashkenazim belong to M33c as opposed to M33c1, but some of those are the same kits I already saw in FTDNA's "Jewish Ukraine West" project.

I no longer think Y-DNA Q1b1a could come from Khazars. The evidence is convincing now that ancestors of the Ashkenazim acquired it in the ancient Middle East. It also seems doubtful that the East Asian mtDNA haplogroups and autosomal relationships have something to do with Khazars.

[Táltos]

The up to 2% of Northeast Asian ancestry in Eastern Ashkenazim isn't noise but a real component. I gathered the evidence we know so far about the relevant East Asian hair gene, mtDNA haplogroups, and autosomal results at http://www.khazaria.com/genetics/aj-...admixture.html in case you want to read it.

khazaria you said in your link:

The mtDNA haplogroup M33c1 is found among Ashkenazim from Belarus as well as among South Chinese people "but from absolutely nowhere in between" those two regions,

Can you please provide an actual paper that shows the M33c1 is in fact the same exact subclade of M33c for both the Ashkenazi and South Chinese that is listed here: http://en.wikipedia.org/wiki/Haplogroup_M_(mtDNA)

Though Wikipedia can be useful, I also take it with a grain of salt as anyone can alter it. Also in the haplogroup M project here I cannot see any M33c1 that is South Chinese. https://www.familytreedna.com/public...tion=mtresults

There are many forms of M in this project, lots from India, some with Arabic sounding names too.

My father line is Q1b1a. Though I am not Jewish, apparently through DNA testing I have found I have some pretty far back. I don't care to push any agenda to prove that Ashkenazi people came from the Middle East or that they descend from Khazars. I think they are a mix of Middle Eastern, European, and I am sure a small bit of Khazars did get in there.

My father and his father had told us that we are Lithuanian, Polish, and Mongolian in this line. Very interesting that they said we had Mongolian ancestry over Jewish. And yes I agree that other populations will score some form of Asian, such as the Russians, Ukrainians. Seeing some of their myOrigins results they have East Asian, Central/South Asian, and even New World in their results.

Here is something that I found to be interesting too concerning the Asian admixture scores in Levantine people. http://www.eupedia.com/europe/Haplogroup_Q_Y-DNA.shtml Scroll down to the section: Central Asian & Indo-Iranian Q1b

[khazaria]

Was wondering how Asian got in.

The up to 2% of Northeast Asian ancestry in Eastern Ashkenazim isn't noise but a real component. I gathered the evidence we know so far about the relevant East Asian hair gene, mtDNA haplogroups, and autosomal results at http://www.khazaria.com/genetics/aj-...admixture.html in case you want to read it.

[josh w.]

Thanks for the discussion everyone.
Now to try to understand it all.
Brian

When Gedmatch is running again, chose the Oracle option with the program. The Oracle results should be similar to PF but with more information

[Jdna]

Thanks for the discussion everyone.
Now to try to understand it all.
Brian

[grahcom]

Admixture analyses is pretty useless IMHO as most database reference groups are either too small or non existent.

[josh w.]

I have played with Gedmatch previously(the site has been down for a few days).
Their results from various parameters are similar to Geno2.
If I can pick and choose which Gedmatch results I like,
Jtest was the only one that had Ashkenazi 31.3%, Eastern Med 20.6%, but a low Mid Eastern 7.1%

Admixture analyses vary in terms of the nomenclature for different regions and how they carve up geographical regions. The Eastern Mediterranean covers Greece, Cyprus and the Levant. Jews show similarity to reference groups from these areas. A study by Haber for National Geographic found that Jews were closest to Lebanese particularly Lebanese Druze. The study did not include a southern Italian sample and I suspect that Jews would be just as close to that group. Regarding the Near East, most programs do not include the Near East region specifically and there may not be an Eastern Mediterranean region. In that case, for Jews the highest Asian component would be West Asian (northern Middle East) with a smaller Southwest Asian component.

[Jdna]

I have played with Gedmatch previously(the site has been down for a few days).
Their results from various parameters are similar to Geno2.
If I can pick and choose which Gedmatch results I like,
Jtest was the only one that had Ashkenazi 31.3%, Eastern Med 20.6%, but a low Mid Eastern 7.1%

[josh w.]

Hello,
I just received my Family Finder results and would appreciate help in understanding something.
My Population Finder results say I'm "100% +/- 0.01% Middle East (Jewish)"... No surprise, but 100%, no admixture?

My geno2 results report me as:
56% Mediterranean, 21% Northern European, 20% Southwest Asian, 2% Northeast Asian.

Why is there a difference in admixture results?
Why didn't Population finder, find my Geno 2 admixture?

I'd appreciate any help in understanding.
Take care,
Brian

The results actually are consistent. PF gives a single best guess as to the reference group you most resemble. But Ashkenazi genes and SNPs are not limited to the Middle East. I have not used Geno 2 but any admixture test covers any region where the ancestors of the ancient Hebrews lived as well as regions Jews moved to after the Diaspora. In particular Jews lived in Roman areas after the Diaspora and there were many conversions to Judaism in these regions. As a result, Jews have a large Mediterranean component as well as West Asian, Southwest Asian(Arabian) and Northern European components. If you upload your results to Gedmatch (choose Dodecad or Eurogenes---+ Oracle) you should see other analyses similar to PF and Geno 2. The Northeast Asian component could just be random noise, but some ethnic Russians have such a component, e.g. from Tatars.

[1798]

The Asian component does not come from Africa. It comes from Asia.

If the simple answer for all of us is that all of our ancestry comes from Africa we should all stop getting DNA tests and doing genealogical studies because none of what happened between 60,000 years ago and now has any significance. FamilyTreeDNA should just close shop except for a page that says YOUR ANCESTORS CAME FROM AFRICA. STOP LOOKING ANY FURTHER.

It depends on which group of your ancestors you are more interested in. I am only interested in the most recent even 2000 years is too far out. FF is within that range and for me it is a brilliant test for 99 dollars. The "y" test for me has faded in to the mist. This FF test is by far the best test in the world for breaking down the brick walls. I have two cousin matches with names that are in my family tree and I don't believe they are coincidental.

[Armando]

The ancestors of modern humans came out Africa 60,000 ybp and spread all over the world.

The Asian component does not come from Africa. It comes from Asia.

If the simple answer for all of us is that all of our ancestry comes from Africa we should all stop getting DNA tests and doing genealogical studies because none of what happened between 60,000 years ago and now has any significance. FamilyTreeDNA should just close shop except for a page that says YOUR ANCESTORS CAME FROM AFRICA. STOP LOOKING ANY FURTHER.

[Armando]

The difference is that the two programs are trying to
do different things. Both can in this particular
case be exactly right (or not). No one program
can reliably tell, unless the person is really truly
>90% Jewish ... and even then I'm not aware of a publicly
available one that can do that reliably.

The Geno2 analysis is one of the absolute worst. And that's
not just because they are using a too-small set of markers.
But it is in general not actually "wrong" ... it won't say you are English if you are Han Chinese.

Please expand on your claim that Geno2 is the absolute worst. It actually provides the source populations and the components that those populations are broken up into. You can't get more transparent than that. Population Finder does not provide that info and actually gives us incorrect ancestry at times. Population Finder does not tell me how my the people in the countries that my ancestors are from are broken up and which components they have. Geno2 does that.

I assume you are focusing on the portion of Geno2 that tries to match a person's results with a reference population. That portion is definitely inaccurate most of the time and should be ignored, but that is not what the original post was about.

I agree that they are trying to do different things and I disagree with the way Population Finder is doing it. Based on the original poster's Geno 2.0 data I was able to predict he has Eastern European ancestry. I would never have been able to do that with Population Finder.