PDA

View Full Version : Eurogenes Genetic Ancestry Detective



cyna
04-26-2014, 02:02 PM
I asked Davidski to take a look at my AncestryDNA raw data. Ancestry tests more ancestral SNPs than 23andme, so I figured I would get better coverage with that set of data. Here is what Davidski said:

----------------------------------------------------

- Begin Analysis -

European 39.5%
Amerindian 1.1%
Southeast_Asian 8.7%
Sub-Saharan Africa 50.7%
TOTAL 100%


British 15.9%
Norse 13.9%
European_Jewish 9.7%
Amerindian 1.1%
Southeast_Asian 8.7%
Sub-Saharan Africa 50.7%
TOTAL 100%

Alright, here we go. Normally I send out a lot of PCA and various maps, and I already sent you a few, but I don't think that stuff is very useful in your case.

For instance, if I break up your genome into four presumably unadmixed grandparents, then these four land in western Nigeria, Denmark and eastern Kazakhstan. This is obviously impossible, since Southeast Asian and Amerindian ancestry, which you carry, isn't native to these regions. Indeed, you also appear to be partly of central and southern European Jewish origin, which further makes the initial classification highly unlikely.

So I've put together a couple of graphics showing your ancestry proportions (you can see the details in the relevant text file). The first one tackles continental ancestry, and in my opinion shows very robust results. On the other hand, the second one attempts to characterize in more detail your European ancestry, and is somewhat speculative. That's because most European groups are very closely related, and it's difficult to break them up based on DNA alone.

From memory, you said that 23andMe and Ancestry gave you very little Scandinavian ancestry. In the end I came up with 13.9%, and I think that's a reasonable figure, because I picked up Scandinavian signals in your genome in almost all of the tests I ran. However, many of your Northwest European haplotypes appear to be of Scottish origin, so perhaps Norse-like Scottish ancestry is inflating the figure? In any case, your Northwest European ancestry is in large part from around the North Sea. The Norse is probably not recent, but I couldn't help putting it in there because Norway lights up in all the tests. Actually, it was obvious you were part Scottish after I did the full scan, but I reckon your Scottish ancestor must have come from an area of Scotland with lots of Norse influence.

Your Jewish ancestry seems to be mostly, but not entirely, Ashkenazi. The reason I say that is because 10% of Ashkenazi admixture is easy to pick up in my tests. On the other hand, the signals in your case were a bit muffled, and you were coming out partly Italian and Spanish in some analyses. This might well be true to some extent, but I suspect that you have Sephardic and other southern European Jewish ancestry, in addition to the aforementioned Ashkenazi ancestry. However, it's very difficult to come up with precise figures in this case. The best I can do is give you a highly speculative estimate for your Jewish/southern ancestor; 3/4 Ashkenazi and 1/4 Sephardic. The problem here is that Sephardic Jews aren't as drifted genetically as Ashkenazis, so it's very difficult to tell them apart from southern Europeans. They're very similar to Italians and Iberians, and indeed you can be fitted as part Iberian and Italian.

The Southeast Asian admixture in your genome is also difficult to place precisely. You share very few haplotypes with Southeast Asians that I was able to compare you with, but those you did share anything with were from Vietnam and southern China. So I'd say your Southeast Asian ancestors came from around that area, or not too far away. Based on the data alone, northern Vietnamese are your best fit. But South Han Chinese are next. I don't know the history of Vietnam too well, but I do know that there were Chinese communities all over Indo-China. Maybe they came from Hong Kong or nearby?

- End of Analysis -

------------------------------------------------------------------------

I have to say that he nailed my ancestry closer than any other service. I am very pleased with his product. A step up from the venerable McDonald, IMO, because he goes into a great deal more detail. I also suspect the quality control is a bit better, as he worked on my profile for about 5 or 6 days, whereas McDonald tends to churn out the reports quickly because of the volume of requests (you get what you pay for). The only thing McDonald offers in excess is the full Chromosome painting, but I think you could ask Davidski for one by request. Anyway, here is how to order a report from Davidski:

http://bga101.blogspot.com.au/2014/03/eurogenes-genetic-ancestry-detective.html

cyna
04-28-2014, 12:46 AM
For some reason, the full analysis didn't paste over. Here is the rest:

- Continue Analysis -

You mentioned that you're not aware or any Amerindian ancestry and that the 1% or so of Amerindian admixture that shows up occasionally in genetic tests might be noise created by your partly Southeast Asian origins. In my opinion this is unlikely, because Southeast Asians generally don't show any Amerindian noise in genetic tests.

Indeed, you actually share small haplotypes with Mexicans and Caribbeans in parts of your genome where you also show inflated affinity to the Americas and Siberia. What this suggests to me is that you have a very small amount of ancestry from Central America. I've attached a chromosome painting showing the areas of your genome affected by this putative Central American admixture. Admittedly, some of these fragments are tiny, so we might well be looking at noise, but certainly not in all of the cases.

In regards to your African ancestry, most of it probably comes from Nigeria and surrounds. That's because you share lots of sizeable haplotypes with Yorubans from Nigeria. In part this might have something to do with sampling bias (Yorubans are the most numerous African samples in the various public genetic databases), but only in part because the signals are very strong.

It's also interesting to note that you regularly share haplotypes with east Africans, specifically those from the Luhya tribe in Kenya. So either you actually have some ancestry from east Africa, or perhaps what we're seeing here are the effects of the massive and rapid Bantu expansion across much of Africa?

Please let me know if you have any questions, or if I missed something.

Cheers

- End analysis -

cyna
05-28-2014, 03:13 AM
I asked Davidski to also take a look at my 23andme raw data as well. He had tweaked his test and added a few additional (African) samples, and here is what he come out with:

-------------------------------------------------------------------------------------------

Ok, I've redesigned the original test so that there weren't any markers missing from either file, and this is what I got...

POP File1 (AncestryDNA)

Ashkenazi - 4.1%
North Sea - 29.8%
Amerindian - 1.4%
Other Mediterranean - 5.5%
Sephardi - 0.4%
Southeast_Asian - 8.1%
Sub-Saharan Africa - 50.7%

POP File 2 (23andme)

Ashkenazi - 3.1%
North Sea - 29.4%
Amerindian - 1.3%
Other Mediterranean - 4.2%
Sephardi - 2.5%
Southeast_Asian - 8.2%
Sub-Saharan Africa - 51.3%

cyna
05-28-2014, 03:46 PM
Compare against my McDonald custom results for my 23andme raw data:

Irish 0.2360 Jewish 0.1648 Yoruba 0.5144 Yi 0.0848 or
English 0.2492 Jewish 0.1479 Yoruba 0.5147 Yi 0.0882 or
Irish 0.2344 Jewish 0.1662 Yoruba 0.5142 Naxi 0.0852 or
English 0.2474 Jewish 0.1496 Yoruba 0.5145 Naxi 0.0885 or
Irish 0.2325 Jewish 0.1614 Yoruba 0.5128 Malaysian 0.0932 or
English 0.2438 Jewish 0.1464 Yoruba 0.5130 Malaysian 0.0967 or
Irish 0.2346 Jewish 0.1604 Yoruba 0.5129 Cambodian 0.0921 or
Irish 0.2211 Sephardic 0.1807 Yoruba 0.5115 Yi 0.0867 or
Irish 0.2190 Sephardic 0.1826 Yoruba 0.5113 Naxi 0.0871 or
English 0.2463 Jewish 0.1449 Yoruba 0.5131 Cambodian 0.0956 or
English 0.2383 Sephardic 0.1596 Yoruba 0.5123 Yi 0.0897

The most noticeable difference is Davidski breaking down the Asian into Asian + Native, and providing a breakdown of the Jewish into Jewish + Mediterranean. Differences exist on the margins but big picture, the two sets of results tend to agree (especially if you consider Asian + Native American to be one super-group). Of course, I still maintain that the Native that Davidski found could be some sort of shared Native/Asian ancestry, but he (Davidski) seemed pretty sure it was not.

CelticGerman
09-15-2014, 07:04 PM
I don't have any known ancestry from the British Isles, but .... (it isn't the whole text David sent me) 2598

Similar results from Doug McDonald (the second pushed the dot eastwards) 25992600

cyna
09-18-2014, 04:05 AM
I don't have any known ancestry from the British Isles, but .... (it isn't the whole text David sent me) 2598

Similar results from Doug McDonald (the second pushed the dot eastwards) 25992600

The difficulty, of course, is that given only a part of the picture, it is impossible to really tell what Wesolowski is saying. I have a colleague that utilized his services and matches some Russian folks on one segment on a particular chromosome. He is bothered that it does not show up in any of his admixture tests, including Wesolowski. IMHO, I don't put a ton of stock in any particular result these days. I think Wesolowski is better than most for sure, but the science overall has a very long way to go.