Azores to Atlantis: Africa through the shadows

By Razib Khan | July 27, 2012 9:55 pm

In many ways the image of Africa in the minds of Westerners has become a trope. The “Dark Continent,” eternal, and primal. Like many tropes the realized existence of this Africa is only within the imagination. The real Africa is far different. For there is no real Africa, there Africas. This truth is on my mind this week as two papers of great importance in understanding African genetic history finally saw the light of day. First, Dr. Joseph Pickrell et al. posted their preprint, The genetic prehistory of southern Africa, to arXiv. Second, out of the Tishkoff lab came Evolutionary History and Adaptation from High-Coverage Whole-Genome Sequences of Diverse African Hunter-Gatherers. Let me step aside here and observe a secondary, but non-trivial, detail. The former is an open access preprint. The second is a complete paper published in a relatively high impact journal, Cell, for which the paper itself does not seem typical or appropriate. This is fair enough, most people do not read journals front to back in this day. But unlike Dr. Joseph Pickrell’s paper the paper in Cell is paywalled, and from what I can tell you can not obtain the supporting information without getting beyond the gate! So if you need that paper, email me and I will send it onward (I would just post it on a server, but I’ve gotten nasty emails from the legal departments of publishers, so I am wary of doing that).


From where I stand these two papers have different strengths. The genetic prehistory of southern Africa has a wide and expansive sampling of individuals and to some extent populations, and, utilizes very clever statistical genetic techniques on SNP data (i.e., hundreds of thousands of variants among the ~3 billion base pairs). Evolutionary History and Adaptation from High-Coverage Whole-Genome Sequences of Diverse African Hunter-Gatherers, as the title suggests, is a very deep and relatively complete read of the whole genomes of ~15 hunter-gatherers, from three populations, Western Pygmies, and two non-Bantu populations from Tanzania, the Hadza and Sandawe (there are other whole genome analyses already of groups to which they can compare these). Personally, I would have preferred genomic analyses of the Eastern Pygmies, who seem more genetically distinctive in relation to the Bantu, but there were likely logistic constraints.

Probably the most interesting finding of the Pickrell et al. paper is that the Hadza and Bushmen seem to share deep common ancestry. More precisely, the Hadza can be modeled as a combination of East Africans and Bushmen, with a 3:1 ratio. Please note: can be modeled as. This does not mean that that is the real history, but, this stylized result probably gives us some insight into the more complex picture. The Hadza and Bushmen are both hunter-gatherers who have clicks in their language. That should shape prior expectations. Additionally, in various genetic analyses they cluster with the Pygmies of Central Africa as a distinct clade from the agriculturalists of Africa. There are other suggestive results in the Pickrell et al. paper. For example, there is evidence of back-migration from Eurasia into the Sandawe. This is not surprising to me, as I’ve explored this issue with coarser techniques (thanks to Brenna Henn & company for releasing these genotypes). Additionally, there are suggestive results in this paper that the Khoikhoi, a pastoralist Khoisan group, have some genetic influence from East Africans (detected in part via old Eurasian affinities). There have been arguments that pastoralism came to the Khoikhoi not through the Bantu, but via earlier peoples, so this is clearly something that needs to be followed up. Overall, this paper seems to have made more precise and sharpened vague outlines which were already established. Tenuous connections between the Hadza and the Bushmen have been confirmed. Very deep separation between the distinct Bushmen groups is reinforced. And finally, the big chasm between a few groups of hunter-gatherer, and other Africans, is illustrated again.

As a final product the Cell paper is rather polished in comparison. The media has emphasized the evidence of archaic admixture into hunter-gather lineages. This is not that surprising of a finding, other groups have reported the same. If anything, I suspect that there is some underestimate here, because it seems to me that the statistical methods have a hard going at this age range, and without ancient genomes at that (i.e., the decay of linkage disequilibrium). Dienekes has commented extensively on this, offering up his own theories. I don’t have any grand model to propose. Rather, I would suggest that even a “leaky” “Out of Africa” model somewhat misleads the public, by again relegating Africa into a sort of inert black-box. There is deep ancient structure within Africa (e.g., Pickrell et al. confirm that San lineages have been diverged on the order of tens of thousands of years!). Though the dates are not totally clear, it seems possible that the most diverged African groups (e.g., Bushmen) separated from the lineage that led to other populations only moderately more recently than non-African archaics! (e.g., Neanderthals) This may not be multi-regionalism, but it does suggest a very deep history and structure for humanity, as opposed to the “old” model that modern humans as such were the product of a single discrete speciation event ~50,000 years ago. There may be no “behold man!” moment.

There’s a large section on adaptation within the Cell paper. I’ll quote their conclusion, because I have no interest in recapitulating genomic alphabet soup:

We find evidence of selective constraint near genes, and these patterns are replicated in each hunter-gatherer population. We also observe signatures of local adaptation in Pygmy, Hadza,and Sandawe populations, including high locus-specific branch lengths for genes involved in taste/olfactory perception, pituitary
development, reproduction, and immune function. These genetic differences reflect differences in local diets, pathogen pressures, and environments. Thus, Pygmies, Hadza, and Sandawe have continued to adapt to local conditions while sustaining their own unique cultures of hunting and gathering.

Surprised that “ancient” hunter-gatherers continue to evolve? I hope not. At this point, let me take a step back. You may wonder: what does this post have to do with Atlantis and the Azores? (or, for that matter Shaka Zulu!) Reading this sort of paper is very interesting to me obviously. The relic hunter-gatherers of Africa can tell us a great deal about our human past. But at some point we need to remember that these are relics. A few years ago I ran into someone who told me that an anthropologist friend explained that the Bantu expansion had been debunked. This sort of thing is why I think cultural anthropology is quite often such a joke discipline. They should stick to their picketing the econ department, and leave the description of the human conditon as it is  to those with a genuine interest. Not only are the Bantu languages a dialect continuum, but the genetic evidence seems rather clear that you can find relative unity across the vast swath of Central, Eastern, and Southern Africa. Though Africans have a great deal of genetic variation, the distance between the various Bantu groups is not that great (e.g., using Fst).

And, the genetic distance between Bantus and Khoisan is such that we can conclude that the Bantu expansion into South Africa, for example, was one of demographics, not cultural diffusion. The Xhosa, who are the outer edge of the expansion, and even had integrated clicks into their language, are only minority Khoisan in ancestry. How did this occur? One model might be one of demic diffusion, the sort promoted by L. L. Cavalli-Sforza, and still favored some researchers. In this model you have raw demographic expansion pushing a population outward, diluting their original genetic signal over time on the wave of advance. This is a “thermodynamic” model, not requiring concerted or coordinated state action, as family and clan groups are constantly expanding on the frontier. I do not believe this is how it played out. The Bantu expansion was very thorough. I think Shaka Zulu and his cruel military formations tell us a lot more than small villages of female farmers. No, I’m not suggesting a continent wide confederacy. Rather, the genetic signal from the region of the East-Central Africa lakes is so strong as far south as the Natal that I have difficulty understanding how there could not have been aggressive actions taken against the smaller and less organized hunter-gatherers. Imagine, if you will, the Bantu expansion being driven by men racing forward in Zulu-like military units, clearing territory of hunter-gatherers, killing and starving them out (and taking the occasional girl as a concubine). A Mfecane writ large! What stopped them in the end in Southwest Africa was ecology: the agricultural toolkit of the Bantu was not suitable for the arid or Mediterranean climates of this region. Until then the Bantu knife cut through the hunter-gatherers like butter, wiping them from the surface of the earth.

With this as a backdrop the hunter-gatherers, and their genomes, are then like mountaintops flooded by farmers. Like the idea that the Azores are the last fragments of Atlantis, so the Pygmies, Hadza, and Khoisan, are the least of the vast domains of pre-Bantu peoples who once inhabited the southern half of the continent. They show deep genealogical relationships, but they are simply a shadow of what once was, and what has disappeared from the face of the earth. The eternal “Dark Continent” was in fact roiled by a demographic revolution of gargantuan proportions within the last 3,000 years. We may debate whether the farmers or hunter-gatherers are the ancestors of modern Europeans, but we know the answer for Africans. The hunter-gatherers lost. In the time of ancient Egypt Africa to the far South was a very different place, the Bantus did not start moving in large numbers until the collapse of the New Kingdom. When the pyramids were young there was a different world across the vast savannas and plains of East and South Africa, and world we’ll never show, and only see through the pinholes provided by the last hunters of the continent.

MORE ABOUT: Africa, Genetics, Genomics
  • Solis

    The Bantu expansion was very thorough. I think Shaka Zulu and his cruel military formations tell us a lot more than small villages of female farmers. No, I’m not suggesting a continent wide confederacy. Rather, the genetic signal from the region of the East-Central Africa lakes is so strong as far south as the Natal that I have difficulty understanding how there could not have been aggressive actions taken against the smaller and less organized hunter-gatherers. Imagine, if you will, the Bantu expansion being driven by men racing forward in Zulu-like military units, clearing territory of hunter-gatherers, killing and starving them out (and taking the occasional girl as a concubine). What stopped them in the end in Southwest Africa was ecology: the agricultural toolkit of the Bantu was not suitable for the arid or Mediterranean climates of this region. Until then the Bantu knife cut through the hunter-gatherers like butter, wiping them from the surface of the earth.

    The eternal “Dark Continent” was in fact roiled by a demographic revolution of gargantuan proportions within the last 3,000 years. The hunter-gatherers lost. In the time of ancient Egypt Africa to the far South was a very different place, the Bantus did not start moving in large numbers until the collapse of the New Kingdom. When the pyramids were young there was a different world across the vast savannas and plains of East and South Africa, and world we’ll never show, and only see through the pinholes provided by the last hunters of the continent.

    I can’t emphasize how much I agree with this. The Bantu expansions acted as a genetic unifier for sub-Saharan Africa.

  • gcochran

    We need another word to describe this process, one that makes clear how it differs from the ‘range expansion’ model.

    I suggest ‘ass-kicking’.

  • Lank

    Regarding the East African influence in the pastoralist Khoikhoi, there is some evidence that points to the demographic significance of the East African pastoralists. Y-DNA E-M293, highly present in the Khoe, peaks in both frequency and diversity in Tanzanian pastoralist groups; Y-chromosomal evidence of a pastoralist migration through Tanzania to southern Africa. M293 peaks in the Datog Nilotes from Tanzania, whose language is strongly influenced by their South Cushitic neighbors, and they apparently derive a majority of their ancestry from those South Cushites (Tishkoff 2009).

    What’s intereresting is that the mutation associated with lactase persistence in the Tanzanian Cushites and Datog also spills over into southern Africa (Itan 2010). So I’d say there’s definitely sufficient evidence for E-M293 carriers migrating from East Africa being responsible for the introduction of pastoralism in southern Africa.

    These South Cushites are closely related to Cushites from Ethiopia, who we know have some Eurasian ancestry. That would explain the results of the Sandawe, who appear to have significant Cushitic ancestry. It seems that the Hadza were not significantly affected by this wave; they have more of an ancient East African affinity, while lacking much of the Eurasian ancestry found in Cushites.

    This is interesting as Sandawe has proposed connections to the Khoe languages of southern Africa, lacking in the Hadza, whose only connection with southern African Khoisan languages appears to be the presence of clicks in their language. The less ancient history of the Sandawe may explain why they fit a relatively simple admixture of Khoisan, Cushitic, and Bantu, whereas the story for the Hadza seems to be a bit more complicated (at least on the East African side).

  • Joshua Gatera

    The Eurasian migration twig detected between the Sandawe and the Eurasian samples included in the study is not particularly accurate. The absence of NE African samples from the Horn of Africa skews the results of the analysis, the inclusion of these aforementioned groups would have resulted in a migration stream from groups in Ethiopia, like the Oromo for example, to the Sandawe. The Sandawe are the probable descendants of an indigenous SE African hunter-gatherer population, related to the Khoe, that ultimately absorbed gene-flow from intruding South Cushites, Bantus, and Nilotes.

    @ Lanks

    The Hadze likely descended from the same indigenous SE African hunter-gatherer population ancestral to the Sandawe, but simply absorbed more Nilotic ancestry than the latter who largely integrated southbound South Cushites. Their Eurasian affinity is more than just “affinity”, but via gene-flow from neighboring Nilotes in South Sudan and the vicinity.

  • Nick Patterson

    I’m a co-author of the Pickrell paper.
    The genetic data in Southern Africa suggests a more complicated pattern
    of Bantu expansion than a simple military takeover.

    Two data items:
    1) There is little evidence of “gender-biased gene flow” with male Bantu mating with
    female click-speakers but not the reverse. This is in contrast to (for instance) the
    “Cape Coloured” where I believe that almost all European ancestry was from males.

    2) The Damara are a genetically Bantu population in Namibia who often acted as servants
    to Nama pastoralists, who have primarily Khoisan genetics. History is complicated!

  • jb

    I’m not disputing your basic thesis, that the Bantus aggressively displaced the hunter-gatherers. This seems very plausible to me. But I was under the impression that Shaka’s military formations were highly innovative, and gave his Zulus a great advantage even over the other Bantu nations. So invoking them as a model for the Bantu expansion as a whole may give the wrong impression. A better model might be the small scale cattle raiding that you still see today in many parts of Africa. Given the higher population densities of the Bantus, and their superior iron weaponry, it seems to me that this is all that would be required to allow the Bantu simply walk in and take over on their own schedule.

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    #5, the issue for me is that i’m not seeing any evidence of admixture cline from kenya to mozambique. who lived here before the bantus arrived? the archaeology indicates they arrived ~2000 years BP. there is obviously a cline in south africa, botswana, etc. many more xhosa resemble khoisan than zulus physically, and would surely be validated by genetic analysis.

    #6, i was simply suggesting that the expansion was more than simply the introducing of an agricultural toolkit.

  • Lank

    The Hadze likely descended from the same indigenous SE African hunter-gatherer population ancestral to the Sandawe, but simply absorbed more Nilotic ancestry than the latter who largely integrated southbound South Cushites. Their Eurasian affinity is more than just “affinity”, but via gene-flow from neighboring Nilotes in South Sudan and the vicinity.

    I see no evidence for Nilotic ancestry in the Hadza. Look at K=4 to K=9 in Pagani’s ADMIXTURE runs (Fig. S1C). At K=4, the Hadza are a rather unique blend of an Ancestral East African with a hunter-gatherer component. They rival the East African of Nilotes, despite lacking the West African component (save for some individuals, definitely with Bantu ancestry) found in significant levels among Sudanese Nilotes. Tishkoff 2009, linked in my previous comment, also found evidence for this distinct East African component in the Hadza. The Sandawe, on the other hand, look very much like a recent composite in Pagani’s ADMIXTURE runs.

    The affinities of the Hadza are not strange considering their mtDNA. Hadza are mainly L4, which descends from the same L3’4 clade as L3. Their divergent hunter-gatherer ancestry seems to be found on the male side, where they carry the B2b lineage that is also characteristic of Biaka Pygmies, and found occasionally in Sandawe and southern African Khoisan.

    I agree on the Hadza and Sandawe descending from a common East African hunter-gatherer population. That is evident in their Y-DNA, mtDNA and autosomal affinities. The other components seem to differ, though. The Sandawe have a less ancient history as a distinct genetic group.

  • Joshua Gatera

    My assertion was based on the amateur results of this particular run https://docs.google.com/spreadsheet/ccc?key=0AqWOT2m6btAGdEpjTExTRG1NU1NKRnJfd3hOeWkwdkE#gid=6 by an interested party at another internet forum; in which the admixture run detected notable “Nilo-Saharan” ancestry in all of the five Hadze sampled. This trend is noted from K=7 onward with the crystallization of the Nilo-Saharan cluster which peaks in the Anuak, Gumuz, and S. Sudanese samples. The aforementioned results could however be a shortcoming of STRUCTURE? What do you think?

    I agree, the Sandawe are a much more cosmopolitan group than their respective hunter-gatherer counterparts the Hadze; they seemed to have easily assimilated (and continue doing so) various waves of South Cushitic, Nilotic, and Bantu gene-flow into an already pre-existing hunter-gatherer base.

    Regarding the Southern Sudanese samples, some of these samples definitely consist of individuals with more recent West African gene-flow, in particular individuals who speak a Central Sudanic language; Central Sudanic speakers have a long and intertwined history with speakers of the Ubangian languages. But there seems to have already been some level of gene-flow with a population genetically similar to contemporary West Africans at a relatively early stage of the Nilo-Saharan phylum. In the order of ~15% since the Anuak of Ethiopia and I’m assuming Dinka/Nuer (Pagani et al. 2012) possess minor affinities to West Africa. A Kordofanian origin for this admixture is a possibly, i.e. Niger-Kordofanian languages located in Central Sudan? The fact that the Gumuz lack significant West African admixture implies that they were already differentiated from other Nilo-Saharan speakers at the time of admixture, something that linguistics supports.

  • Lank

    Regardless of how Nilo-Saharan speakers attained their ‘West African’ affinity, it would definitely have been present in Nilotes expanding south from Sudan.

    Remember, the East African ancestry of the Hadza has to be assigned somewhere. I wouldn’t pay much attention to their ‘Nilo-Saharan’, when lower K values make it clear that Hadza cannot have significant ancestry from the Nilotic groups that migrated to East Africa relatively recently.

  • Joshua Gatera

    I agree, my statement regarding West African admixture among East African Nilotes was independent from our previous discussion in respect to the genetic affinities of the Hadze. I wanted to know your opinion?

  • Lank

    I don’t know how to explain the West African. A Kordofanian source is possible. That’s about the only explanation I can think of for the widespread West African in Nilo-Saharans, if recent admixture is the reason. Alternatively, Sudanese Nilo-Saharans may just not have been as differentiated from West Africans as the people living farther east.

    BTW, the Gumuz may have a reduced West African affinity, but that is partially due to their Ethiopian highland ancestry.

  • ryan

    >the genetic signal from the region of the East-Central Africa lakes is so strong as far south as the Natal

    My understanding was that the original homeland was in the Cameroun/Nigeria area. Is the genetic signal not as strong going that far back? Or were you just speaking of the general sweep of the genetic signal, rather than intending to show the origin and direction of the change?

  • Lank

    #13, some southern African Bantus are known to descend from Bantu groups that migrated from the African Great Lakes region. Ultimately, all Bantu groups originally came from the Cameroonian/Nigerian homeland, that’s right.

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com

ADVERTISEMENT

See More

ADVERTISEMENT

RSS Razib’s Pinboard

Edifying books

Collapse bottom bar
+

Login to your Account

X
E-mail address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »