May 05, 2010

European admixture major source of structure in African Americans

The inset barplot (left) shows the proportions of European (blue) and African ancestry in the studied sample of African Americans. As can be seen, African Americans are quite similar in their admixture of the various "African" clusters (at least the ones included here), but differ from each other primarily in their level of European ancestry.

Figure 4 shows quite well how the various African components in African Americans are uniformly distributed.

From the paper:
The largest African ancestral contribution comes from the Yoruba, with an average of 47.1% ± 8.7% (range, 18% to 64%), followed by the Bantu at 14.8% ± 5.0% (range, 3% to 28%) and Mandenka at 13.8% ± 4.5% (range, 3% to 29%). The contributions from the other three African groups were quite modest, with an average of 1.7% from the Biaka, 0.5% from the Mbuti, and 0.3% from the San.

Table 1 has a breakdown of the ancestral components of African Americans from different geographical regions.

Genome Biology 2009, 10:R141doi:10.1186/gb-2009-10-12-r141

Characterizing the admixed African ancestry of African Americans

Fouad Zakharia et al.

Abstract

Background
Accurate, high-throughput genotyping allows the fine characterization of genetic ancestry. Here we applied recently developed statistical and computational techniques to the question of African ancestry in African Americans by using data on more than 450,000 single-nucleotide polymorphisms (SNPs) genotyped in 94 Africans of diverse geographic origins included in the HGDP, as well as 136 African Americans and 38 European Americans participating in the Atherosclerotic Disease Vascular Function and Genetic Epidemiology (ADVANCE) study. To focus on African ancestry, we reduced the data to include only those genotypes in each African American determined statistically to be African in origin.

Results
From cluster analysis, we found that all the African Americans are admixed in their African components of ancestry, with the majority contributions being from West and West-Central Africa, and only modest variation in these African-ancestry proportions among individuals. Furthermore, by principal components analysis, we found little evidence of genetic structure within the African component of ancestry in African Americans.

Conclusions
These results are consistent with historic mating patterns among African Americans that are largely uncorrelated to African ancestral origins, and they cast doubt on the general utility of mtDNA or Y-chromosome markers alone to delineate the full African ancestry of African Americans. Our results also indicate that the genetic architecture of African Americans is distinct from that of Africans, and that the greatest source of potential genetic stratification bias in case-control studies of African Americans derives from the proportion of European ancestry.

Link

8 comments:

Andrew Oh-Willeke said...

One of the interesting things about the European admixture percentages is how small the extremes of the distribution are in this study. Only one in 136 have no European ancestry. Eight in 136 have more than 45% with only four or five having more than 50% European ancestry and none having more than 72%. Part-African populations in Jamicia and most of Latin America have much more variation in this distribution, I suspect, with significantly more part-African individuals who have predominantly non-African origins. Presumably this has roots in miscegenation laws (which the genetic evidence would suggest were passed following a period of widespread use of slaves by white male owners for sex), Jim Crow laws, and differences in the nature of Anglo-American slavery and slavery elsewhere.

The study is too small to capture any statistical differences from Louisiana, but if one were to look for statistical outliers, that would be the place to look, because the way the French handled slavery and race was quite different than the Anglo-American model and there were significant "free colored" populations there that were largely absent in the American South. The French pattern prevailed for a few decades after the Louisiana purchase and there was little incentive for free colored populations to migrate to other places in the U.S. (those who could leave often went to France or Canada).

The lack of structure in the African component of the population is also notable. The study found some West African, Nigerian and Pygmy/Khoisan component in the ancestry of every single one of the 136 individuals tested, in quite stable proportions, in a population that came to the New World mostly in the period from 320-160 years ago and wasn't very geographically mobile once placed on plantations (the American South had a very anemic rail network and didn't have very good roads as recently at the American Civil War). This isn't that many generations of bring about such complete admixture.

The study is not designed to catch Native American ancestry, estimated in other studies at 1.85%-3% nationally and much more in some areas (e.g. many Puerto Ricans have both Mestizo and African descent), or low frequency African populations (e.g. from Southeast Africa and Madagascar, or Ethiopia).

Maju said...

Andrew: "and none having more than 72%".

Logically: they would have been assimilated as whites. Octoroons are normally invisible: they almost only display traits of the major component.

Socio-psychologically it would work this way: neither blacks nor whites would anymore recognize that 75% European person as "black" anymore.

...

Overall: I find interesting the high apportion of "Yoruba" blood among AAs. It does seem to say something about where the bulk of the ancestry originated: coastal West Africa. I'd like of course a more comprehensive set of African references (Mozambicans, Malagasy...) but guess this is informative anyhow.

terryt said...

"These results ... cast doubt on the general utility of mtDNA or Y-chromosome markers alone to delineate the full African ancestry of African Americans".

Many seem to forget that problem.

"I find interesting the high apportion of 'Yoruba' blood among AAs. It does seem to say something about where the bulk of the ancestry originated: coastal West Africa".

That's interesting because it's usually assumed the earliest slaves came from around Senegal-Gambia. The Mandenka are at 13.8% ± 4.5% (range, 3% to 29%) (and Wolofs don't appear at all, perhaps not included in the study). Presumably later arrivals from further south came to dominate, or the people of Senegambia were trading slaves they had sourced far inland.

Maju said...

No, Terry, there's no Wolof sample in the study. It's rather limited in this aspect.

However, the bulk of 18th and 19th century slave trade did concentrate in Nigeria (Slave Coast) and earlier too, most slave forts were placed in coastal West Africa but further west, in the Gold Coast area.

While in the 19th century there was a true "breeding program" in the USA to face the loss of imports (illegal under the constitution, also British growing blockade), we know little of the survival rate of slaves before those dates. If we have to compare with other plantation economies of America, it was surely very low, demanding more and more human merchandise from across the ocean.

This may also help explain the origins of the genetic pool towards a late slave trade logic, rather than an early one.

terryt said...

"we know little of the survival rate of slaves before those dates".

I'm reading a book at the moment where the author claims a very low survival rate for the early arrivals, so that supports your assessment.

andrew said...

"Logically: they would have been assimilated as whites. Octoroons are normally invisible: they almost only display traits of the major component.

Socio-psychologically it would work this way: neither blacks nor whites would anymore recognize that 75% European person as "black" anymore."

FWIW, I suspect that the socio-psychological cutoff in the U.S. would perceive someone as "black" at quite a bit less than 25% European ancestry; and the admixture rates of black and white populations in the U.S. from about 1865 (really closer to 1850) to 2000 have been exceedingly low, so while the category transition theory is plausible, I suspect that it isn't empirically accurate.

An African said...

LOL!

Its funny how europeans are so desperate for AAs to be mixed with them they tell themselves fiction to make it true.


Europeans have the most wrinkled, badly aging, heavy skin of any race of people on this planet.


African Americans are just the opposite. Yet poor low self-esteemed whites will tell themselves that AAs are mixed with them even though the inherent contradiction defies common sense.

Maju said...

To "An African": You know what? Your sardonic-moronic racist attitude makes me think that you are not a true Black but a KKK agitator making fun of them.

You are just saying what you say to give bad reputation to Black People: of racist and of stupid. I'm pretty much tired of your meaningless tirades.