November 02, 2010

ADMIXTURE analysis of Spencer Wells

One of the people who've entrusted me with their DNA for analysis in the Dodecad Project is none other than the Genographic Project's Spencer Wells. His project ID is DOD162, and he is the very last individual to be included in the project's pilot phase.

I've enjoyed watching the Genographic Project's various documentaries and reading their published articles, often commenting on them in this blog, so it's a nice opportunity to give something back to the leader of one of the few organizations that is really helping advance our knowledge of human origins.


Here are the results of the admixture analysis: Spencer Wells is in the first bar, and admixture proportions of the 10 components I am using are color-coded for both himself and 36 other populations.

His results are uneventful: his bar is very similar to the one next to it, which summarizes the admixture proportions of 25 individuals from the HapMap-3 CEU population. His top component is North European (60.6%) as his appearance and Northwest European ancestry would suggest. Next is South European (24.9%) and West Asian (13.5%). Rounding up his results is a small slice (1%) of Southwest Asian.

Below you can see Spencer followed by the 25 CEU individuals; these are American Whites from Utah, and as you can see, while there is some small variation in proportions and minor components, he "fits right in" this population.
I have also compared him against the 16 Dodecad Project participants who belong to my "American White" category. This is a rather fuzzy category, consisting of European-descended Americans and Canadians whose ancestry was not entirely from one of my other categories. In that population, the average components are: 11.3% West Asian, 0.2% Northwest African, 26.8% South European, 0.1% Northeast Asian, 1.2% SW Asian, 60.4% N European, 0.1% S Asian, also quite close to Spencer's results.

26 comments:

apostateimpressions123 said...

How comes Lithuanians show no Asian ancestry when they have high levels of haplogroup N? Does the analysis distinguish between north European and north Asian properly? I see a NE Asian category but no NW Asian.

And how can Wales for instance, which is 82% R1b be the same "northern European" category as Lithuania, which is only 5% R1b? There has been minimal mixing between the Welsh and the Lithianians for many thousands of years. So how comes they have been homogenised into the same genetic ancestral component?

Dienekes said...

Re: Lithuanians

http://dienekes.ifreepages.com/blog/archives/000209.html

"The results obtained by the Baltic Expedition, in which the author participated, showed that the Lithuanians, for example, do not disclose any difference from the Caucasians in regard to eyelid structure. In all probability, the distinctions of Russians from Caucasians should not be explained by the fact that the eyelid structure of Northern Europeans is different from that of Southern Europeans, but by the fact that the Russians absorbed elements which are characterized by some Mongoloid traits."

Re: Northern European component

Y-chromosomes tell only part of the story. The interesting question is why the three main West Eurasian components form, in terms of Fst distances, an equilateral triangle, all of them being about Fst=0.05 from each other. I think I have the answer, but it should be best left for another post.

Marnie said...

"The interesting question is why the three main West Eurasian components form, in terms of Fst distances, an equilateral triangle, all of them being about Fst=0.05 from each other."

Fertile Crescent: Still With US?

http://linearpopulationmodel.blogspot.com/2010/11/fertile-crescent-still-with-us_03.html

eurologist said...

I think I have the answer...

And I think many will place the answer 10,000 to 30,000 years before your time frame...

Just a humorous guess - and not to be taken too seriously.

Dienekes said...

And I think many will place the answer 10,000 to 30,000 years before your time frame...

Hint: East Africans, Northeast Asians, and West Eurasians form another equilateral triangle with Fst about 3 times the smaller Eurasian one

onur said...

East Africans, Northeast Asians, and West Eurasians form another equilateral triangle with Fst about 3 times the smaller Eurasian one

I don't see such an equilateral triangle.

aargiedude said...

Speaking of riddles, how about posting a guessing game like you did with facial composites, in which we had to guess their country of origin, but instead do it with some of the samples that were sent to you by the people that tested with 23andme. Of course, they would have to be ok with it.

onur said...

Marnie, your bar chart results neglect the importance of the North European component in the northern parts of West Asia. The North European component is probably very old in Asia Minor, the historical Armenian highlands, the Caucasus (including Transcaucasia), northern Mesopotamia, many parts of Persia and the Eurasian corridor, considering its notable presence in Uyghurs, in what is now Pakistan and, to a lesser extent, in Yakuts. Expanding Neolithic waves coming to Europe from Asia Minor probably had the North European component; they might even be the originators of that component if the demic diffusion hypothesis is true (in a demic diffusion scenario, it is likely to have increased in amount as the Neolithic populations from Asia Minor advanced in Europe, as the Neolithic populations would lose genetic diversity as they advance).

Marnie said...

Onur, I'm not neglecting the northeast component, just trying to break out the Fertile Crescent from the data.

Yes, the North European component is surely very old, but it's center of gravity origin is likely in the North, on the other side of the Caucasus, and likely on the Dnieper River. (Jean's website: http://www.buildinghistory.org/distantpast/peoplingeurope.shtml#copper)

Cypriots don't have the North European component. Neither do Egyptians.

So although Turks and Armenians have the North European component, I'm asserting that that represents an admixture event between North Europeans and peoples of the Fertile Crescent.

Marnie said...

Sorry, the first sentence in the above post should read:

Onur, I'm not neglecting the North European component, just trying to break out the Fertile Crescent from the data.

onur said...

Yes, the North European component is surely very old.

I was talking about its oldness in West Asia.

So although Turks and Armenians have the North European component, I'm asserting that that represents an admixture event between North Europeans and peoples of the Fertile Crescent.

You can't make such a conclusion based on these data. Components don't come with ages. Also, as I said, the North European component is very widespread across Eurasia (Yakuts live in eastern Siberia), so it is probably very old in West Asia.

onur said...

Marnie, and lastly, everything would be clearer if Dienekes included Iranians (preferably also native West Asian Jews except far southerners like Yemeni Jews, Uzbeks, Druze and Samaritans) in his K=10 analysis.

clusteredmaps said...

I wonder why Dienkes is intent on leaving out the eastern European.
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005472#pone.0005472.s003

eurologist said...

y-DNA N likely entered the Baltics from between Ukraine and the Urals, where both Balto-Slavic and Finnish likely originated - but y-DNA is a horrible indicator of autosomal DNA. As I sometimes say, if you take three non-related people from neighboring villages in the German Saale region, you have a good chance of getting one R1b, one R1a, and one I2b1. However, their autosomal DNA would be almost indistinguishable.

The Northern European component in my view is clearly pre-neolithic and exactly not colored by neolithic elements. However, it is still possible that it attained its widespread distribution largely in the Neolithic.

eurologist said...

Hint: East Africans, Northeast Asians, and West Eurasians form another equilateral triangle with Fst about 3 times the smaller Eurasian one

Well, if I divide 60,000 to 120,000 by three, I get 20,000 to 40,000... ;)

Dienekes said...

I wonder why Dienkes is intent on leaving out the eastern European.
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005472#pone.0005472.s003


Find me the data, and I'll include anyone you want.

Fanty said...

"As I sometimes say, if you take three non-related people from neighboring villages in the German Saale region, you have a good chance of getting one R1b, one R1a, and one I2b1. However, their autosomal DNA would be almost indistinguishable."

Exactly. It becomes even more drastic if one sees it like this:

Say, you have R1b and your mothers father has J.

Your own aDNA will be much closer to that J Grandfather than to any R1b in the world except the father and grandfather where this R1b came from.

Also one does spread his aDNA for sure. Wich is not the case with Y-DNA or mtDNA.

If a man has only doughters, his Y-DNA is lost for the world. But his aDNA survives.

If a woman has only sons, her mtDNA is lost for the world but her aDNA survives.

Drift (the magic word to explain if the distribution of a haplogroup does not make sense or does not fit into a theory) could totaly eliminate several Y-DNA and/or mtDNA haplogroups, but their aDNA would still be part of the genepool.

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

"And how can Wales for instance, which is 82% R1b be the same "northern European" category as Lithuania, which is only 5% R1b? There has been minimal mixing between the Welsh and the Lithianians for many thousands of years. So how comes they have been homogenised into the same genetic ancestral component?"

The magic drift word. *snickers*
And not to forget mtDNA H.

People tend to always forget mtDNA.

I once made me some mtDNA maps, based on a chart (I am a visual thinker somehow. I hate charts ;) )

Some of the major mtDNA in Europe:
http://img266.imageshack.us/f/mtdna2.jpg/

Someone could suggest a connection between the "Northern" European component with mtDNA H.

At least in the extented version of the K12 of Polako where the Northern cluster assimiliated the "Atlantic" cluster.

Just a thought.

onur said...

http://img266.imageshack.us/f/mtdna2.jpg/

Fanty, thanks for the map. Is there a West Asian version of it?

clusteredmaps said...

Haplogroup H mtdna shows no association with the North European Component.

clusteredmaps said...

The Baltic-Slavs should have their own cluster as should the Eastern Europeans.
"At a more detailed level, however, several distinct regions can be distinguished within Europe: 1) Finland, 2) the Baltic region (Estonia, Latvia and Lithuania), Eastern Russia and Poland, 3) Central and Western Europe, and 4) Italy, with the southern Italians being more “distant” (Figure 2)
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005472#pone.0005472.s003"

onur said...

1) Finland, 2) the Baltic region (Estonia, Latvia and Lithuania), Eastern Russia and Poland, 3) Central and Western Europe, and 4) Italy, with the southern Italians being more “distant” (Figure 2)

It is obvious that they misspelled "western Russia" as "eastern Russia" here, as in another section of the same article they write:

"The resulting genetic map forms a triangular structure with a) Finland, b) the Baltic region, Poland and Western Russia, and c) Italy as its vertexes, and with d) Central- and Western Europe in its centre."

Also no where else in the article they mention eastern Russia, but mention western Russia several times and show it in figures and tables.

Fanty said...

"Haplogroup H mtdna shows no association with the North European Component."

Compare that:
http://img266.imageshack.us/f/mtdna2.jpg/

to that:
http://img28.imageshack.us/img28/381/threebigclusters.jpg

Whoever has more than a certain degree on H, is in that "extented" Northern cluster.

The lowest distribution of H is in Italy and the Southwest of Europe = exactly the places where the northern component is extremely low.

To my eyes, that "extented" Version of the northern cluster does show association to R1 aswell as H.

That result (the extended version) came to life in a worldwide K12.

Northern center = Lithuania
Mediteranian Center: Sardinia
and as special guest apearance:
West Eurasian: Pakistan/Northwest India

Fanty said...

correction:

"and the Southwest of Europe"

SouthEAST of Europe.
SIGH.... I have a West/East bug.
Almost like a right/left one.

clusteredmaps said...

"Whoever has more than a certain degree on H, is in that "extented" Northern cluster."

Why then do the Tunisians and Azerbaijanis have so little Northern component?

Fanty said...

"Why then do the Tunisians and Azerbaijanis have so little Northern component?"

Why have Finn 50% N at 8% North Asian aDNA? Or Lithuanians 1% NOrth Asian aDNA at 50% N?
Why are Brahmins 70% R1a, while they only have like 5% (Wich is already an interesting find! Is this from the same source as the R1a?) northern European aDNA?
Or Poles 60% R1a and having next to no Indian aDNA?

Drift? I dont know. ;)

We should recall that, at the current trend to downscale age estaminates and MRCAs age etc... all Europeans Y-DNA seems to originate in very few (possibly 10?) neolithic males.

Means: Massive Drift. Tens of thousands Y-DNA lineages of the neolithic age are GONE. A handfull remains. Haplogroups that once had 50% are down to 0.1% and so on....

What was that? I1 MRCA? 5K years?
25% of the Germans belong to a R1b type that has a MRCA at 6K years? All Russian R1a men, originate in a single R1a man who lived 6K years ago? Thats 50% of the population.

The madness goes on.

If all this would be truth, all of Germany bases on 4 neolithic males. Can these be representative of the Y-DNA makeup of the neolithic? Most possibly not. They are ranomly chosen 4 guys. Their randomly chosen haplogroups dominate Germany now. The other halogroups of their thousands tribal comrades...are gone.

I lost any trust in Haplogroups.

So, it may be drift, random whatever, but... there are two hotspots of H in Europe: One is the Basques and the Iberian Peninsular as a whole (Wich is equaivalent to the center of the "Atlantic Cluster" (A aDNA cluster that shapes if one uses he Basques as Center). And one all around the Baltic sea (equivalent to the center of the Northern European Cluster that shapes if one uses the Lithuanians as center)

And the Atlantic cluster unites with the Northern European cluster into one Supercluster at given circumstances, what suggests that these two clusters are "closely" (by this I dont mean 1000-2000 years... I rather think about 5.000-10.000 years or something like that) related.

And then it springs to the eyes that this supercluster now has the Hotspots of R1a, R1b and the two European H hotspots inside.

But thinking about the massive drift, this could be coincidence.

Talking about drift... what aDNA did the R1 have? Coulnd it be that the very first R1a and the very first R1b had almost identical aDNA? After all they are both mutations of R1 and are virtually equaly old. Maybe they had been literate brothers. ;)
Well ok, unlikely that such a mutation happens twice in one family. Hehe.

Hmm, as far as I know, a very few R1 did actually survive drift and live from Anatolia to India.

Dienekes said...

Fanty, the latest studies of R1a1 show that South Asian R1a1 diversity cannot be explained as a subset of West Eurasian R1a1 diversity. So, R1a1 in South Asia is not, I believe, fully of European origin, although a subet of it might very well be. The antiquity of R1a1 in Europe is given, since it was discovered in Germany in the Bronze Age, but the full extent of its prehistoric distribution is currently unknown.