October 23, 2012

The great human expansion (Henn et al. 2012)

I have been a rather outspoken critic of the "standard recent Out-of-Africa model" of human origins. A new paper by Henn, Cavalli-Sforza, and Feldman presents an up-to-date version of that model, and is quite useful as an overview of what I believe to be (and I'm sure the authors do not!) the passing paradigm.

From the paper:
Genetic data indicate that, approximately 45 to 60 kya, a very rapid population expansion occurred outside of Africa, and spread in all directions across the Eurasian continents, eventually populating the entire world.
This is true. The question is whether this expansion originated in Africa itself, or in Eurasia, from people who had left Africa at a much earlier time. One aspect of this expansion that is often brought in defense of this hypothesis is the orderly diminution of genetic diversity outside Africa from the Near East. But, we ought to remember that "clines don't carry dates", and that particular one is consistent with an Out-of-Arabia dispersal of modern humans during the time in question.

From the paper:
However, current evidence indicates that this near-modern population did not persist in the Near East and was subsequently replaced by Neanderthals during the following glacial period, with little evidence of temporal overlap (5, 6). It is not until at least 50,000 y ago that evidence of behaviorally modern humans occurs in the archaeological record in the Near East.
The evidence for behavioral modernity (the transition to the Upper Paleolithic and/or Lower Stone Age) appears near simultaneously around the planet and is thus no evidence for an Out-of-Africa event accompanying it.

If out species became behaviorally modern due to a population expansion circa 50ka, then we would expect different human populations to have split times of ~50ka. This is not, however, what we observe, but, rather, in all genetic systems (mtDNA, Y-chromosomes, and autosomal DNA), there is evidence for population splits within Homo sapiens of order 200ka, which were not, however, complete, but were followed by later episodes of admixture.

A recent paper estimated that the Khoe-San split from the rest of us ~100ka. Even if we disregard the use of a now-outdated mutation rate, this is still twice as old as the UP/LSA transition. The implication is clear, that at least in some part of our species, behavioral modernity c. 50ka did not spread through the spread of a new population, but through the spread of an idea. Now, let's flip this around, and go from South Africa to the Levant, where we do have evidence for a ~100ka split between a group of modern humans (the Mt. Carmel ones) and African humankind. If a population that split off ~100ka (and indeed, more likely 200 ka) within Africa is, nonetheless fully behaviorally modern by "cultural osmosis", so could the population of modern humans who lived in Asia pre-100ka: no need to invoke population replacement to explain the appearance of behavioral modernity.

The argument is simple: deep genetic population splits are no obstactle to the flow of culture in the case of Africa, so why postulate an obstacle to cultural flow (in whatever direction) between human groups with equal, or indeed much shallower split times?

I have written before about my distaste for Biblical-level bottlenecks, and here they are presented explicitly:
Resequencing studies have estimated the ancestral effective population size at 12,800 to 14,400, with a 5- to 10-fold bottleneck beginning approximately 65,000 to 50,000 y ago (although see ref. 15 for a bottleneck to only 450 individuals). It is generally assumed that the bottleneck occurred as a small group(s) with an effective population size of only approximately 1,000 to 2,500 individuals moved from the African continent into the Near East.
This is of course, possible. But, the model writes the story, and if one assumes tree-like divergence of human populations sans admixture, then one will doubtlessly infer a story of migration going from the most diverse human populations, to the least diverse ones.

But, admixture matters. In the proximate sense, the diminution of genetic diversity from East Africa has never been established securely: to do so, one would need to isolate what is "diverse by admixture" and "diverse by antiquity". These proximate causes of increased diversity can be addressed because there are relatively unadmixed groups of people still in existence, and admixture LD has not had sufficient time to decay. This is particularly the case for many intermediate populations in the road Out-of-Africa, including East Africans, Near Eastern populations, South Asians, etc., all of which have evidence for recent admixture. Indeed, recent work has also established admixture within Africa itself, of both the recent and the archaic kind.

But, if the principle of admixture is accepted, then the possibility that it may have occurred in the distant past must also be entertained. In the absence of both LD-based evidence (which decays exponentially), and extant unadmixed populations (which tend to be absorbed or die out), older episodes of admixture will manifest themselves as little more than an excess of polymorphism, all the greater depending on the size of the introgressing element and its genetic divergence: a little admixture from a much diverged element will contribute a similar number of new alleles as a lot of admixture from a less diverged one.

In fact, we do see such an excess of polymorphism in Africans, and it will serve us well to remember that the Out-of-Africa bottleneck may have joined forces with an In-Africa-Admixture to create the contrast between African and Eurasian effective population sizes.

The story told in The great human expansion is, in my opinion, no longer believable. Three reasons have contributed to make it so:

  1. The publication of the Neandertal and Denisovan genomes have killed off any notion that the human tree blossomed in a vacuum, unperturbed by the other denizens of the Homo forest. 
  2. Recalibration of the human autosomal mutation rate have revealed deep autosomal divergences within our species. These can be consistent with either (i) a recent expansion followed by admixture with divergent lineages, or (ii) old population structure within the species accompanied by cultural flow of behavioral modernity. I tend to support a mix of these ideas. What cannot have happened, however, is the model of a recent, simultaneous expansion responsible for both the spread of modern humans and behavioral modernity. 
  3. While the ~60ka Out-of-Africans remain elusive, archaeologists have made steady progress in uncovering real links between Africa and Eurasia prior to 100ka. The Mousterian-using Mt. Carmel members of our species can no longer be discounted as the Out-of-Africa that failed, because they are now accompanied by Nubian Complex and Jebel Faya Arabians at around the same time.
Point #3 is particularly important: these are real archaeologically demonstrated links between Africa and Asia. It is no longer possible to discount Skhul/Qafzeh as the little Levantine colony of modern humans that failed, because they're no longer the only evidence for pre-100ka Out-of-Africa: a model must now demonstrate either why (i) all pre-100ka modern humans failed, or (ii) the Arabians-with-African-technologies in places like Dhofar would not have been modern humans.

PNAS doi: 10.1073/pnas.1212380109

The great human expansion

Brenna M. Henn et al.

Genetic and paleoanthropological evidence is in accord that today’s human population is the result of a great demic (demographic and geographic) expansion that began approximately 45,000 to 60,000 y ago in Africa and rapidly resulted in human occupation of almost all of the Earth’s habitable regions. Genomic data from contemporary humans suggest that this expansion was accompanied by a continuous loss of genetic diversity, a result of what is called the “serial founder effect.” In addition to genomic data, the serial founder effect model is now supported by the genetics of human parasites, morphology, and linguistics. This particular population history gave rise to the two defining features of genetic variation in humans: genomes from the substructured populations of Africa retain an exceptional number of unique variants, and there is a dramatic reduction in genetic diversity within populations living outside of Africa. These two patterns are relevant for medical genetic studies mapping genotypes to phenotypes and for inferring the power of natural selection in human history. It should be appreciated that the initial expansion and subsequent serial founder effect were determined by demographic and sociocultural factors associated with hunter-gatherer populations. How do we reconcile this major demic expansion with the population stability that followed for thousands years until the inventions of agriculture? We review advances in understanding the genetic diversity within Africa and the great human expansion out of Africa and offer hypotheses that can help to establish a more synthetic view of modern human evolution.

Link

2 comments:

Aaron said...

I agree with your comments.

If you were to take all of the genetic information of people today, without any understanding of our recent history, and use the current genetic models, you would wrongfully assume an out of America hypothesis.

America has the most genetic diversity; the alleles from Europeans, Africans, Amerindians, and Asians would be represented. Each of the other continents would simply be bottleneck migrations from America.

I realize that allele diversity can be created over time, but admixture has a much stronger effect than time. In fact East Africa has the most diversity simply as a result of recent admixture between Eurasians and Sub-Saharan Africans, not necessarily because Eurasians and Sub-Saharan Africans came out of East Africa.

terryt said...

"I have written before about my distaste for Biblical-level bottlenecks, and here they are presented explicitly"

Beliefs established during our childhood are extremely difficult for us to discard.

"One aspect of this expansion that is often brought in defense of this hypothesis is the orderly diminution of genetic diversity outside Africa from the Near East. But, we ought to remember that 'clines don't carry dates', and that particular one is consistent with an Out-of-Arabia dispersal of modern humans during the time in question".

And the cline may even have been established during the migration of H. erectus out of Africa.

"In fact East Africa has the most diversity simply as a result of recent admixture between Eurasians and Sub-Saharan Africans, not necessarily because Eurasians and Sub-Saharan Africans came out of East Africa".

And to me the diversity in southern China is also most easily explained as the result of mixture between the Chinese Neolithic and the pre-existing Papuan-type people in South China and SE Asia. This seems to be an unpopular view at present.