April 30, 2009

Tishkoff et al. on genetic structure of Africans and African Americans

(
(Updated May 1)

The importance of this new paper from the Tishkoff Lab cannot be emphasized enough. It is probably the most comprehensive study of African genetic variation to date. The supplementary material (pdf) is itself 102 pages long and should keep you busy reading for a while (free for non-subscribers).

What this study has found in a nutshell is that "black" Africans belong to 14 distinct clusters. Black Americans belong overwhelmingly to the Niger-Kordofanian cluster, consistent with their origin largely from Western Africa.

The paper covers the levels of diversity in different African populations, finding that:
Within Africa, genetic diversity estimated from expected heterozygosity significantly correlates with estimates from microsatellite variance (fig. S4) (4) and varies by linguistic, geographic, and subsistence classifications (fig. S5). Three hunter-gatherer populations (Baka and Bakola Pygmies and San) were among the five populations with the highest levels of genetic diversity based on variance estimates (fig. S2A) (4). In addition, more private alleles exist in Africa than other regions (fig. S6A). Consistent with bi-directional gene flow (14), African and Middle Eastern populations shared the greatest number of alleles absent from all other populations(fig. S6B). Within Africa, the most private alleles were in southern Africa, reflecting those in southern African Khoesan (SAK) San and !Xun/Khwe populations (fig. S6C) (12). Eastern and Saharan Africans shared the most alleles absent from other African populations examined (fig. S6D).


As I have stated many times before, Bantu speakers have recently expanded from their cradle and contributed genetically to almost all other Africans, while remaining relatively pure in their own homeland:
High levels of heterogeneous ancestry (i.e. multiple cluster assignments) were observed in nearly all African individuals, with the exception of western and central African Niger-Kordofanian speakers (medium orange), who are relatively homogeneous at large K values (Fig. 5C and fig. S13). Considerable Niger-Kordofanian ancestry (shades of orange) was observed in nearly all populations, reflecting the recent spread of Bantu-speakers across equatorial, eastern, and southern Africa (26) and subsequent admixture with local populations (27).

Similarly, the high levels of diversity in East Africa are attributable in part to the multiple waves of migration into the region, which occurred in the last 5,000 years, long after the ancestral Eurasians left the continent:

East Africa, the hypothesized origin of the migration of modern humans out of Africa, has a remarkable degree of ethnic and linguistic diversity, as reflected by the greatest level of regional substructure in Africa (figs. S13, S14, and S17 to S19). The diversity among populations from this region reflects the proposed long-term presence of click-speaking Hadza and Sandawe hunter-gatherers and successive waves of immigration of Cushitic, Nilotic, and Bantu populations within the past 5,000 years (4, 28, 31, 37, 38).

The 14 clusters are: Mbugu, Chadic, Saharan Cushitic, Eastern Bantu, NiloSaharan, Saharan/Dogon, Fulani, Western Bantu, S.African Khoesan/Mbuti, Niger Kordofanian, Sandawe, Central Sudanic, Hadza, W.Pygmy.

Table S8 in the supplementary material also allows us to measure the extra-African influences in African populations (and vice versa). Mozabites, for example, from the Sahara are 60.2% "European" reflecting their substantial Caucasoid influence. The Beja possess about a third of this "European" element, the Dogon about 45%. Conversely, low-level African admixture in the Near East consists primarily of the "Cushitic" cluster.

UPDATE I (May 1):

The results of the STRUCTURE analysis on a global level (Figure S10 in the supplementary material provide a visual display of global diversity when seen in an African context. At K=14, there are 10 African and 4 non-African clusters (European, Indian, East Asian, Native American). Contrast between Central-Western and Eastern Africa is the most salient feature of the broad picture, punctuated by interesting mini-clusters formed by Pygmies, Khoi-San and Hadza.

An interesting quote from the paper (AAC=Associated Ancestral Clusters):

The Fulani and Cushitic (an eastern Afroasiatic subfamily) AACs, which likely reflect Saharan African and East African ancestry, respectively, are closest to the non-African AACs, consistent with an East African migration of modern humans out of Africa or a back-migration of non-Africans into Saharan and Eastern Africa.
I had pretty much argued as much in the old Dodona forum, that the intermediacy of certain African people between Sub-Saharan Africans and Eurasians was due both to the fact that (a) Eurasians did not originate from Africa in general, but from a specific subset of Africans that was already differentiated from the rest of Africans, and (b) there were back-migrations of Eurasians into Africa.

Related:


Science doi:10.1126/science.1172257

The Genetic Structure and History of Africans and African Americans

Sarah A. Tishkoff et al.

Abstract

Africa is the source of all modern humans, but characterization of genetic variation and of relationships among populations across the continent has been enigmatic. We studied 121 African populations, 4 African American populations, and 60 non-African populations for patterns of variation at 1327 nuclear microsatellite and insertion/deletion markers. We identified 14 ancestral population clusters in Africa that correlate with self-described ethnicity and shared cultural and/or linguistic properties. We observe high levels of mixed ancestry in most populations, reflecting historic migration events across the continent. Our data also provide evidence for shared ancestry among geographically diverse hunter-gatherer populations (Khoesan-speakers and Pygmies). The ancestry of African Americans is predominantly from Niger-Kordofanian (~71%), European (~13%), and other African (~8%) populations, although admixture levels varied considerably among individuals. This study helps tease apart the complex evolutionary history of Africans and African Americans, aiding both anthropological and genetic epidemiologic studies.

Link

31 comments:

skincolour said...

As I've stated in my latest post on my blog - it is clear that Arab traders began a gradual introduction of tools which facilitated population growth, before a veritable expolosion due to Portugese farming methods. I think the new research is proof, that

1) Bantu speakers were basically active in a process of slave-cultivation, which led to an absolutely unappreciated and ignored population explosion in West Africa following 1500 CE/AD.

2) This is contrary to modern claims that Slavery and the Atlantic and Eastern Slave Trade led dragged Sub-Saharan population figures down.

AK said...

Wouldn't you expect a population explosion to top out with an excess population that could be disposed of through slavery, massacre, or both?

eurologist said...

I haven't had a chance to read the paper, but BBC reports that one conclusion is that modern humans may have originated near the South Africa - Namibia border, before being separated for 50,000 to 100,000 years from a splinter group. This scenario is the exact opposite to that of Spencer Wells, who (at least last year), based on mtDNA argued that the split occurred in East Africa, with the ancestors of the Khoisan moving South and then being separated from a northern/central population by the great East-African droughts.

To me, the new scenario makes more sense, (1) based on very early human locations (South Africa), and (2) also based on the idea that no matter what drought, there would have always been a habitable zone adjacent to the retreating rain forests (which were probably inhabited by earlier non-modern humans, if at all), and in and around the Great Lakes area. This combined productive area would have been far larger than that of the dry South (limited to a very few river valleys and their immediate coastal exits) - explaining the larger effective population size and higher diversity of that (i.e., nowadays dominating African) group.

Dienekes said...

"This analysis indicates modern human migration
originated in south-western Africa, at 12.5° E and 17.5° S,
near the coastal border of Namibia and Angola,
corresponding to the current San homeland, with the
waypoint in northeast Africa at 37.5° E, 22.5° N near the
midpoint of the Red Sea (figs. S2C, S28, and S29). However,
the geographic distribution of genetic diversity in modern
populations may not reflect the distribution of those
populations in the past, although our waypoint analysis is
consistent with other studies suggesting a northeast African
origin of migration of modern humans out of Africa (1, 53)."

ren said...

I will say one thing: The "logic" of Afrocentrics (like Dienekes's friend Charlie Bass) defending the sub-Saharaness of "Ethiopians" makes as much sense as Chinese defending the East Asianess of Somalis.

Deirdre said...

I am merely an economic historian, so please be patient with my ignorance. What intrigues me is that genetic diversity would be accompanied by a diversity of talents, that is, a high variance in intellectual and physical abilities, to the extend that they are inheritable. Africa for example is already known to have the tallest and the shortest people. Therefore (I say optimistically) when in the next 50-100 years Africa attains modern riches (routinely sending a high percentage of young people to university, for instance) we should see a cultural explosion. Twenty Einsteins. Ten Mozarts. Does such an argument (believe me on the economics!) make sense to you genetic folk?

n/a said...

"Does such an argument (believe me on the economics!) make sense to you genetic folk?"

No.

pconroy said...

Deirdre,

The tallest people in the world are from the Dinaric Alps - in the Balkans - average male height 6' 1 1/4 or 186 cms".

The tallest nation in the world is Montenegro, the second tallest is Holland - both of which are European.

It is often claimed that Dinka, Maasai or Tutsi are taller, but I have never seen reliable evidence of this.

pconroy said...

It's fascinating that at k=13, the Dogon of the Sahel look about 85% European and 15% San in this study - images of them, also bear this out, except that they are phenotypically darker than either European or San.

I wonder what this says about the peopling of the Sahel?

pconroy said...

It's also strange why they have a group Tutsi/Hutu - when these people are ethnically different, the former tall pastoralists of possible Nilotic origin, the latter shorter agriculturalists of Bantu origin.

vesid said...

it is interesting that the "european" element exist almost only in the areas of afro-asiatic peoples. Compare this map with the map of afro-asiatic languages in wikipedia. This maybe tells something about the origin of these languages.

pconroy said...

It would be interesting to see Egypt and Morocco on this map, and also all of South Central Africa (Angola, Zambia, Zimbabwe and Mozambique), but I'd especially like to see samples from Madagascar, including highlands, West and East coasts...

vesid said...

Compare the map of the E3b distribution in africa with the maps i mentioned before. It seems to be a conection between "european" element, E3b and afro-asiatic languages.

WP said...

Why is it that whenever there is any mention of Africa, certain individuals manage to weave in the subject of slavery. I find it quite narrow-minded and simplistic to associate the origin of mankind with the recent practices of slavery, which has existed literally everywhere.

Most historical sources indicate the Bantu expansion was fueled by agriculture and metalworking technology starting around 3000 BCE, in the Nigeria-Cameroon area, long before the existence of Arabs in North Africa.

wolcupitol said...

According to Tishkoff, one of the main conclusions of her study is that the world population samples produced 14 genetic clusters, and that 9 of them were in Africa. She said the clusters represent "genetically divergent ancestral population clusters". "You're seeing more diversity in one continent than across the globe". Clearly, according to Tishkoff, the 14 clusters she identified represent 14 ancestral populations, each with its own individual migratory history, existing separately from the other 13 ancestral populations for a long enough time to develop their own genetic profile, discernable today in Tishkoff's study. [Another important conclusion is that "the new findings will help medical researchers tailor drug treatments for different groups of Africans rather than treating them as homogenous."]

But identifying a genetic cluster doesn't automatically mean it represents an ancient population cluster. In her previous study from 2004, "Implications of biogeography of human populations for ‘race’ and medicine", she analyzed Rosenberg's genetic study of the 52 global populations of the HGDP samples, which resulted in 6 clusters. The clusters consisted of entire continents, except one which was made up exclusively by the Kalash in the mountains of north Pakistan. Tishkoff said that this 6th cluster "probably reflects high levels of inbreeding and genetic drift in that group". That's very likely and I completely agree. So in this case we have an example of a genetic cluster that doesn't represent an ancestral population, and in fact, the exact opposite, an extremely recent population cluster.

In this study, Tishkoff hasn't looked into this, because 3 of the 14 genetic clusters are almost certainly the result of inbreeding. In graph S10 (page 34) of the supplementary file of the study, they are the Dogon (gray), Hadza (yellow), and Mbugu (white). The latter 2 have less then 10,000 people, like the Kalash, and are clear candidates for inbreeding. The Dogon are instead almost a million people, but their samples were treated differently than the rest. "Because the inclusion of closely related individuals can impact population genetic inferences, we took the conservative approach of excluding individuals inferred to be third degree or more closely related. An exception was made in the case of the Dogon as it is difficult to reliably infer relative pairs in a small sample." The Dogon consisted of 9 samples (the average was 20). They were exempted from a test of relatedness, which means inbreeding can't be ruled out. And they form a cluster to which they belong 100% and which isn't found in any other population in any amount. This is identical to the Kalash or the Hadza, and it's to be expected from inbreeding. All the Dogon samples came from the same small city. I would easily consider the Dogon, Hadza, and Mbugu clusters to be the artifact of inbreeding, like the Kalash in Rosenberg's study, and discard them. That leaves 11 global genetic clusters, 6 of which are found in Africa.

http://img212.imageshack.us/img212/2829/dogonhadzambugu.gif

The remaining clusters are authentic, but they still don't necessarily represent ancestral population clusters. According to Tishkoff in her 2004 study: "individuals living in the same geographic region are more likely to mate with each other than with individuals from more distant regions. Therefore, due to the process of genetic drift, populations have differentiated over time." As regions become differentiated, they develop structure that can then be identified as a genetic cluster.

Perhaps one way to be able to tell if the genetic cluster represents an ancestral population cluster or simply regional drift is to measure the genetic distance between populations that belong to different genetic clusters. If the cluster is the result of an ancient population cluster, then we would expect great genetic differentiation. This is what we see between the 5 non-African genetic clusters identified by Tishkoff (Caucasians, India, East Asia, Oceania, America). Their genetic distances average 0,1300 FST, which is pretty high. In the case of regional drift the genetic distances could be just about anything, depending on the severity of the drift. In a study of Sardinia, 8 isolated villages close to each other had an average FST distance of 0,0200, incredibly high for people living so close. This is an example of extreme inbreeding. On the other hand, Bauchet's 2007 study of Europeans produced 6 genetic clusters, 5 of which could be assigned to a geographic region (Spain, Basques, north Europe, southeast Europe, Finns). The average distance between these 5 clusters is about 0,0050 FST. Almost certainly, Bauchet's clusters represent regional genetic drift and not 6 ancient population clusters that mixed together to form modern Europeans.

We can measure the genetic distance between several of the African genetic clusters in the Tishkoff study. Though a lot of the populations belong to a mixture of several clusters, some belong almost exclusively to one, and others belong to a mix that doesn't include any clusters from the former.

http://img522.imageshack.us/img522/9360/senegalnigeriakenyasout.gif

The distances between the 3 populations made up mostly of the orange cluster with respect to the purple/red Kenyan population ranges from 0,0100 to 0,0170 FST. This doesn't look like ancient population clusters, it looks like regional genetic drift. The FST distances amongst the orange populations range from 0,0100 to 0,0160 FST. This means that the orange populations are as distant to each other as they are to the purple/red Kenyan population. This lack of correlation between membership in a genetic cluster and genetic distance was also seen in Bauchet's study of Europe. In it, several of the populations belonged almost exclusively to a single cluster, and despite this, the genetic distance between populations belonging to different genetic clusters was no different than between populations belonging to the same cluster, over the same geographic distance. The only exception was Finland, which made up one of the six genetic clusters.

Realizing that regional genetic drift is a factor that will also produce genetic clusters leads to the following conclusion: the regions most represented in the study will tend to produce the most amount of clusters. In Tishkoff's study, 61% of the samples were from Africa, and 64% of the genetic clusters she found were from Africa. In Bauchet's study, he made an initial analysis in which he included 2 African populations from opposite ends of Africa, together with Indians and Mongolians, and of course the European samples. The African samples made up only 23% of the samples, and only 33% of the genetic clusters (2 out of 6). So it seems that after the genetic clusters are identified, their nature must be determined, and not automatically assumed that they represent ancient population clusters.

Only 2 African genetic clusters in Tishkoff's study could arguably constitute an ancestral population cluster different from the principal population cluster of Africa (ergo 3 total ancestral population clusters). They are the Pygmies (green) and the Khoisan (light green). Their distance with respect to the main of the African population range from 0,0400 to 0,0900 FST, close to the distances between Caucasians, East Asia, Oceania, and America.

Kamama said...

strange why they have a group Tutsi/Hutu - when these people are ethnically differentThese samples are from people who had one Tutsi parent and one Hutu parent.

Kosmo said...

Wolcupital, thanks for that. That throws some new light on the data.

PConroy, you wrote in response to Deidre: "The tallest people in the world are from the Dinaric Alps - average male height 6' 1 1/4."

This is true, certainly-- these people have the tallest average height of any population.

I think Deidre's point is still a good one though. The massive African genetic diversity would produce a disproportionate number of individuals in the long tails, regardless of where the average sits. And height is a good analogy for her "twenty Mozarts" idea. Your population from the Alps might have the tallest average height, but the genetic diversity found in Africa has produced populations so divserse that Manute Bol (a 7'7" basketball player) was not the tallest person in his village. Manute was a Dinka, and while the Dinka's may not have the greatest *average* height, I bet they're the only population in the world that has a handfull of seven footers hanging out at the local water hole.

The bell curves in Africa should be flatter and longer because of their genetic diversity.

pconroy said...

wolcupitol,
Very good points!

Yes, it certainly would seem that a few of these "ancestral" populations are merely the product of inbreeding and genetic drift.

BTW, your links are broken - can you repost - thanks.

Kosmo,
I don't understand your point, as if you agree with wolcupitol points, then you would be agreeing that there are a number of flaws in methodology and analysis in this study, which give an inflated level of diversity in Africa.

What you have to realize is that Eurasia has undergone a tremendous number of "selective sweeps", where advantageous alleles sweep East-West or West-East, due primarily to the fact that much of this vast area is at similar latitudes, with few barriers to geneflow - especially after the domestication of the horse. This is not the case for Africa, which has had many barriers to geneflow. The main selective sweeps events in Africa, seem to have been the back-migration of R1b populations to the Sahel region, and then from there the Bantu expansion, which was still taking place in the 1700's into South Africa.

So while certain alleles have fixed all across Eurasia - the picture is not similar in Africa, where you have many relict populations, which have received little exogenous geneflow.

Also, Manute Bol is not a good example, as based on physiology, he almost certainly has Marfan Syndrome - a condition which produced great height combined with a concave chest and heart malformations, and is highly heritable. In an inbred population, there would be many close relatives in a village with Marfan Syndrome.

A friend of mine, a Serb from near Montenegro, who is 6' 8", and has a brother also 6' 8", and a 11 yo sister, who is 6' 0", said his grandfather was 6' 10", and about 40% of the guys in high school class was over 6' 6".

Here's a good site on Height Statistics, which shows what percentages of extremely tall people you will have at various average heights.

Kosmo said...

Kosmo, I don't understand your point, as if you agree with wolcupitol... points, then you would be agreeing that there are a number of flaws in methodology and analysis in this study, which give an inflated level of diversity in Africa.

-- (?) I agree with Wolcupital that a genetic cluster doesn't automatically equal an ancient population cluster. As for the level of genetic diversity, it simply is what it is. And Africans are more diverse. I get what you're saying about selective sweeps, and gene flow, and allele fixation, but it seems like we're talking past each othe a bit, because my point was simply that the genetic diversity in Africa would flatten their bell curves in comparison to the rest of the world-- and the selective sweeps and allele fixation you mentioned would only work to cause that by constricting the bells curves of the rest of the world in comparison.

Kosmo said...

A population with a broad, smeared-out bell curve will produce more individuals on the tails than a population with a constricted bell curve.

pconroy said...

Kosmo,

What I'm trying to say, is that Africans appear more diverse, because African populations have experienced less geneflow across different populations clusters, compared to Eurasians.

So in Eurasia, for trait X, you might have a few beneficial alleles derived from different population clusters - say allele a from Europe, allele b from South Asia and allele c from East Asia - and all 3 alleles fixed to a greater or lesser extent, across the whole area.

While in Africa, selection is mostly local, with much less geneflow across population clusters - with the exception of the Bantu agriculturalists, which the vast bulk of Africans are descended from.

So there is probably less opportunity for advantageous alleles to read fixation in Africa, due to barriers to geneflow, then in Eurasia.

Therefore I'd expect African populations NOT to exhibit a "Normal Distribution" for many traits, but to exhibit a Multimodal Distribution.

The point to remember, which wolcupitol alludes to, is that much of the diversity in Africa is produced from including very small fringe groups. If these groups were excluded, diversity is actually greater in Eurasia.

The size of a population cluster is positively correlated with its chances of developing a beneficial mutation - i.e new beneficial allele.

Kosmo said...

"I'd expect African populations NOT to exhibit a Normal Distribution for many traits, but to exhibit a Multimodal Distribution."

-- I see now what you're trying to say. Still, I'm not sure that a multimodal distribution would be a good argument against the idea that longer statistical tails might occur in Africa, and thus Deidre's idea of "many Mozarts" might not be so strange after all.

I'm sure somebody else must have thought of this, too, but basic premise is this:

Genetic diversity = diversity of trait expression = a broadening of the bell curve for a given quantitative trait = more outliers, regardless of the mean.

Thus it would be possible for an imaginary population X, with high diversity and a low mean, to produce more individuals at the highest quantitative levels than a population Y with low diversity and a high mean.

Broaden the curve enough, and pretty soon, at the far edges, the shape of the curve is far more important than than where you pin the mean.

I've also been thinking about what you said about Manute and Marfan syndrome. I hadn't heard that he might have Marfan, but even if he does, I don't think that disqualifies him as being part of the long tail I'm talking about. Let me explain why. Marfan happens all over the world, and while it causes extreme height, it doesn't usually produce an individual who is 7'7". Then why is Manute 7'7"? If Manute DOES have marfan, then it is apparent that there is something unusal about Manute's genetic background that has caused the Marfan allele to express in a way that is more extreme than typical. This is exactly the kind of outlier that you'd expect to pop up when an allele of large effect is expressed within a genetically diverse population.

I read the height data you linked to. That was some very interesting stuff! Thanks.

pconroy said...

Kosmo,

I guess we'll have to agree to disagree, as you seem to be saying:
1. Africa is the most diverse region in the world, far greater than Eurasia.
2. Great regional diversity leads to more extreme distributions.
3. More extreme distributions - on the positive side - leads to more genius, among other things.

While I'm saying, because of the flaws in methodology and analysis, Eurasia has 3 large population clusters, churning out novel mutations, which have experienced geneflow for millenia; while Africa has only 1 such cluster, with very little exogenous geneflow, therefore:
1. Eurasia is more diverse
2 & 3 are the same.

Yes, those height stats are interesting. Especially when you see that 1 in 5,000 Japanese men is 6' 1 1/4", when that height is the average in the Dinaric Alps.
BTW, my friend Todor said he hung with 5 friends in high school, and at 6' 8", he was the second shortest of them, one guy being only 6' 6"!

Kosmo said...

"1 in 5,000 Japanese men is 6' 1 1/4", when that height is the average in the Dinaric Alps."

Makes me wish I was from the Alps. My Dad was 6'2" (Swede/Pole) but ruined my shot at pro basketball by marrying my 5'1" mother. Drat.

Kosmo said...

Pconroy wrote: "I guess we'll have to agree to disagree, as you seem to be saying:
1. Africa is the most diverse region in the world, far greater than Eurasia.
2. Great regional diversity leads to more extreme distributions.
3. More extreme distributions - on the positive side - leads to more genius, among other things."

Well, to be honest, I'm just holding out number three as a *possibility*, though I admit that the distribution spread would have to be pretty darn extreme. But it seems theoretically possible.

I guess one question that needs to be answered is this: Within a population, to what extent does genetic diversity = cognitive diversity?

n/a said...

"Genetic diversity = diversity of trait expression"

Since this is false, you can stop right here.

argiedude said...

wolcupitol = argiedude

Every time I sign up to a new forum, I just quickly make everything up to get it over with, and then later if I decide to stick around I have to get my stuff straightened out. And of course I'm going to stick around in this forum.

The links I posted aren't broken, I just don't know how to do that html thingy. Copy and paste and they open up. Here's my lame attempt to make them clickable:

Kosmo said...

"Genetic diversity = diversity of trait expression"

Since this is false, you can stop right here.

(?) Why is that false? Obviously, it wouldn't work that way for every nucleotide base-pair substitution, since some mutations are silent, but it would still tend to be true across populations, no?

I think it's safe to assume that a genetically diverse population will tend to be more phenotypically diverse than a genetically homogeneous population.

n/a said...

No:

1. Phenotypic variance is not a measure of genetic variance. If we see a population that has a large measure of phenotypic variability, it does not mean that the population had much genetic variability. Perversely, genetic variability can sometimes be lower in a population that has greater phenotypic variance -- often because genetic drift can cause a loss of epistases that once constrained the phenotype. In some cases environmental variance may actually increase when the additive genetic variance declines, because of a loss of developmental robusticity. In any event, we can't just go from a variable phenotype and infer that there's variation in genotypes.

Kosmo said...

"1. Phenotypic variance is not a measure of genetic variance."

Well, that is certainly counterintuitive and more or less blows my theory to crap. Thanks for telling me something I didn't know, N/A.

Karl said...

"More extreme distributions - on the positive side - leads to more genius, among other things."

It's interesting that you mention this, as it's something that I've thought about for quite a while. Academic chess outliers from this group, when you think about it, make as much sense as athletic outliers.

I know of two chess programs with some rigor introduced into African-American Junior High schools - one in Harlem and the other in Philadelphia - both gaining fantastic results. The Vaux school in Philadelphia won 7 straight national titles('77-'83) and 8 of 9 years. The school from Harlem managed 3 straight national titles and 4 of 5 years.

These achievements are more remarkable than African-American over-representation in basketball, track and football when you consider that they lacked, for the most part, the access that wealthier students had to private coaches in supplement of their school coaches. They were and are quite disadvantaged. http://www.timesx2.com/vaux/history.htm Also, they did not come from familial chess cultural backgrounds.