October 10, 2010

ADMIXTURE on African HapMap populations

Here is the result of running ADMIXTURE on the three African HapMap-3 populations, using about 440K SNPs, including Tuscans as a non-African group.

The Tuscans are in purple and show no trace of African admixture. All the other populations are separated: red: Luhya (Bantu); green: Maasai (Nilotes); Yoruba (Niger-Congo).

The two east African groups show asymmetrical affinities: the Maasai have some Luhya red, while the Luhya have little Maasai green, while they have substantial West African turqoise, consistent with the origin of their Bantu language.

6 comments:

onur said...

Dieneke, the proportion values of the purple and turquoise clusters are in the wrong places in the table.

The Tuscan proportion value is accurate in showing no Negroid admixture. As to the Negroid genetic differences, despite some long-range migrational distortions, they seem to show a cline in accordance with geography that attests to them being parts of the same race (Negroid). Though more Negroid populations would be needed to see that cline more clearly. The Maasai seem to show some Caucasoid admixture, this isn't unexpected given their location (East Africa), but more ADMIXTURE analyses should be implemented to see whether it is real admixture, or an artifact of wrong population choice for analysis, or something else.

DocG said...

Please explain exactly what you mean by "negroid" and "caucasoid." Are you referring to place of origin, skin color, morphology, genetics? Other?

aargiedude said...

The Tuscan proportion value is accurate in showing no Negroid admixture.

Wait a minute. What about their 2% (or more) mtdna L from several hundred samples? That results in 1% sub-Saharan ancestry. It may seem small, but in the scale of such a vast test, with 440K SNPs and almost 100 samples, it should have shown up. Previously we would have abscribed such a difference to the possibility that Tuscans' mtdna L came from east Africa, but Dienekes' test includes east and west Africa.

Marnie said...

"the proportion values of the purple and turquoise clusters are in the wrong places in the table."

The table is fine. Maybe it's the color map on your computer, Onur.

What is clear, even with this limited population set, is that Africans are composed of combinations of genetically distinct groups. Simply lumping Africans into "Negroid" belies the natural complexity.

Onur, you've stated your position multiply times. It's getting old.

Could we instead look at what the data is telling us, without trying to force it into the concept of "race", "pure"?

I agree that it would be helpful to see some other African groups in these runs. In particular, it would be good to see one or two other Bantu groups.

Thanks.

onur said...

Please explain exactly what you mean by "negroid" and "caucasoid." Are you referring to place of origin, skin color, morphology, genetics? Other?

It is a combination of all of them (though place of origin is less important today due to the long-range migrations in the modern era), as in all subspecies.

Wait a minute. What about their 2% (or more) mtdna L from several hundred samples? That results in 1% sub-Saharan ancestry. It may seem small, but in the scale of such a vast test, with 440K SNPs and almost 100 samples, it should have shown up. Previously we would have abscribed such a difference to the possibility that Tuscans' mtdna L came from east Africa, but Dienekes' test includes east and west Africa.

As I wrote earlier, L doesn't necessarily mean Negroid as its origin can be pre-Negroid or some of it may be non-Negroid.

The table is fine. Maybe it's the color map on your computer, Onur.

It is fine because Dieneke corrected it after my warning. It was thus before my warning (I had copied the link to the original non-corrected state of the table to record it, foreseeing objections like yours):

http://4.bp.blogspot.com/_Ish7688voT0/TLCzzDNd3kI/AAAAAAAACtE/uVAE_6h9UTU/s1600/LUHYA_MAASAI_TUSCAN_YORUBANproportions.jpeg

As you see, the proportion values of the purple and turquoise clusters are in the wrong places here.

Could we instead look at what the data is telling us, without trying to force it into the concept of "race", "pure"?

The results of such genetic analyses should be evaluated critically and should be tested multiple times with different populations and markers. Only then we can arrive at more definite conclusions. My conclusions are based on many previous similar genetic analyses, so I am not trying to force anything.

alfio said...

Thanks for explanations Onur, always very accurate.

Marnie:

I don't get why everytime someone shows no admixture we have to shut up or get called "racists" or "purists".

This is genetic, not a racial pride and have nothing to with racism.