Razib Khan One-stop-shopping for all of my content

May 7, 2010

The three layers of the Neandertal cake

I assume by now that everyone has read A Draft Sequence of the Neandertal Genome. It’s free to all, so you should. At least look at the figures. Also, if you haven’t at least skimmed the supplement, you should do that as well. It’s nearly 200 pages, and basically feels more like a collection of minimally edited papers than anything else. There’s no point in me reviewing the paper, since you can read it, and plenty of others have hit the relevant ground already.

Since there seem to be three main segments of the paper, here are a few minimal thoughts on each.

First, the draft genome. What would you have said if someone came up to you ten years ago and told you that you’d live to to see this? Svante Paabo himself admitted he didn’t think he’d see something like this in his lifetime. There was a lot of hard work that went into figuring out how to get at the genetic material, purify it, and confirm that it was actually from the samples in question and not handler contamination and such (remember that there was a problem with contamination a few years back). To a great extent the focus on the results, instead of the methods, is like critiquing a set of landscape photographs taken from a very high peak. We can’t forget the effort and energy that went into scaling the peak itself. A lot of labor input obviously went into this, but additionally we can thank the fact that we live in a technological society where progress is not only expected, but often can’t be accounted for in our projections of future possibilities. I think that’s a very hopeful thing which makes me a little less pessimistic about the possibility of the magic carpet economy.

Second, the are the comparisons between Neandertals, modern humans, and chimpanzees. As Carl Zimmer noted there are an alphabet soup of genes thrown at you in the results. It is hard to make sense of it all, though I did note that genes involved in skin function and phenotype seem to have been the subject of differential evolution between Neandertals and modern humans (i.e., SNP differences in regards to substitutions in the lineages). We already know that there are suggestive signs that Neandertals lost function on pigmentation independently from modern humans. That shouldn’t be too surprising, given that it seems that West and East Eurasians evolved light skin independently. There are some uncertainties about the timing of this, but the different genetic architecture implies that it was unlikely to have occurred immediately after the Out-of-Africa event, and in fact some of the loci imply that depigmentation may have occurred in the Holocene. Skin is famously our biggest organ, so it shouldn’t be that shocking that it is possibly a target of selection, but curious nonetheless (recall that it seems that humans evolved darker skin from a paler ancestor as we lost our fur in the tropics).

Additionally, I think the finding that Neandertals and modern humans seem to share most of the same HARs, regions of the genome where our human lineage seems to differ from other mammals in exhibiting a lot of evolutionary change, is of great interest, though not necessarily surprising. When pointing to Luke Jostins’ post on rates of encephalization, I observed that in some ways it seems like there was a very powerful and consistent lineage specific trend toward greater cranial capacity which had incredible time depth. In The Dawn of Human Culture Richard Klein puts the emphasis on the sharp break between those populations before ~50 thousand years ago, and after. This period is marked by the shift toward behavioral, as opposed to just anatomical, modernity (there were anatomically modern humans in Africa ~200 thousand years ago). Klein’s thesis is that some mutation triggered a radical biocultural change, and was responsible for the Great Leap Forward, the efflorescence of creative symbolic culture which we truly consider the sin qua non of culture. The sharing of HARs between Neandertals and pure humans, and the consistent trend toward encephalization (aside from the post-Ice Age reversal), makes me shift the priors a touch more toward inevitable continuity and away from contingency. I find much of the politics of Robert J. Sawyer’s Neanderthal Parallax series a bit heavy-handed, but his depiction of Neandertals as fundamentally intelligent creatures who differ only on the margins seems a lot more plausible to me now than it was when I first read it in the early to mid-aughts.

Third, and finally, there’s the story of admixture and sex. This is getting all the press, but of course this is the most uncertain, inferential, and speculative aspect of the paper. It’s impressive, but it should open to skepticism, especially after the Out-of-Africa totalism which was ascendant until recently. John Hawks accepts the thrust of the findings, but obviously has his own ideas as to modifications, extensions and qualifications. Dienekes Pontikos favors an alternative interpretation of the data, which the authors point to in the text but dismiss as less parsimonious. My own inclination is to favor the authors in their interpretation of parsimony, but I will admit that this assertion is disputable. Dienekes and others would suggest that it is just as, or more, plausible that the shared variants between non-Africans and Neandertals arise from their common northeast African ancestral population (or some ancestral population of non-Africans and Neandertals). He rightly points out that there may be ancient population substructure within Africa, and using a particular African group as a “reference” for the whole continent may lead to false inferences. The main issue is that the probability of retrieving ancient DNA from northeast African samples in the near future seems low because the conditions for preservation are not optimal  (tropical climates famously degrade and recycle biological material more efficiently than temperate or boreal climates). Additionally, using modern northeast African populations is somewhat problematic because there has clearly been some back-migration from the nearby Arabian populations into this area in the medium-term past (the languages of the Ethiopian highlands are Semitic). One supposes that one could differentiate between the African and Arabian components of the genome of Ethiopians and Somalis, but if the admixture event was two to three thousand years ago I presume it would be technically more challenging than an African American, where very few generations have passed since admixture for recombination to fragment long genomic regions attributable to one ancestral population. In other words, how do you distinguish Neandertal variants which arrived back from Eurasia from ancient African ones? (I suppose that the haplotypes would differ so that the genuinely African ones would be more diverse)

But even if you reject the top-line finding, that most of us are not pure human, I think the paper is a game-changer in terms of shifting your priors in relation to evaluating the plausibility of a result which suggests admixture from an ancient non-African population. I found out about the high likelihood of this paper just before the UNM results were presented at the American Anthropological Society meeting, and it is clear in hindsight with the large author list that many people knew what was coming down the pipepline and had recalibrated their assessment of results which indicated admixture. It is perhaps time to go back and take a second look at papers which you skipped over before because it seemed that they may have been spurious or reporting a statistical quirk because they lay outside of the orthodox paradigm. This is clearly a case where it is good to live in interesting times.

Citation: Green, R., Krause, J., Briggs, A., Maricic, T., Stenzel, U., Kircher, M., Patterson, N., Li, H., Zhai, W., Fritz, M., Hansen, N., Durand, E., Malaspinas, A., Jensen, J., Marques-Bonet, T., Alkan, C., Prufer, K., Meyer, M., Burbano, H., Good, J., Schultz, R., Aximu-Petri, A., Butthof, A., Hober, B., Hoffner, B., Siegemund, M., Weihmann, A., Nusbaum, C., Lander, E., Russ, C., Novod, N., Affourtit, J., Egholm, M., Verna, C., Rudan, P., Brajkovic, D., Kucan, Z., Gusic, I., Doronichev, V., Golovanova, L., Lalueza-Fox, C., de la Rasilla, M., Fortea, J., Rosas, A., Schmitz, R., Johnson, P., Eichler, E., Falush, D., Birney, E., Mullikin, J., Slatkin, M., Nielsen, R., Kelso, J., Lachmann, M., Reich, D., & Paabo, S. (2010). A Draft Sequence of the Neandertal Genome Science, 328 (5979), 710-722 DOI: 10.1126/science.1188021

No scientists had to die for this paradigm shift!

Filed under: Genetics,Genomics,Human Evolution,Neandertal,Neandertal Genome — Razib Khan @ 12:17 am

In Science Ann Gibbons has a very long reported piece, Close Encounters of the Prehistoric Kind. It’s well worth reading, but behind a pay wall. If you don’t have access though, I want to spotlight one particular section:

The discovery of interbreeding in the nuclear genome surprised the team members. Neandertals did coexist with modern humans in Europe from 30,000 to 45,000 years ago, and perhaps in the Middle East as early as 80,000 years ago (see map, p. 681). But there was no sign of admixture in the complete Neandertal mitochondrial (mtDNA) genome or in earlier studies of other gene lineages…And many researchers had decided that there was no interbreeding that led to viable offspring. “We started with a very strong bias against mixture,” says co-author David Reich of Harvard Medical School in Boston. Indeed, when Pääbo first learned that the Neandertal DNA tended to be more similar to European DNA than to African DNA, he thought, “Ah, it’s probably just a statistical fluke.” When the link persisted, he thought it was a bias in the data. So the researchers used different methods in different labs to confirm the result. “I feel confident now because three different ways of analyzing the data all come to this conclusion of admixture,” says Pääbo.

The finding of interbreeding refutes the narrowest form of a long-standing model that predicts that all living humans can trace their ancestry back to a small African population that expanded and completely replaced archaic human species without any interbreeding. “It’s not a pure Out-of-Africa replacement model—2% interbreeding is not trivial,” says paleoanthropologist Chris Stringer of the Natural History Museum in London, one of the chief architects of a similar model. But it’s not wholesale mixing, either: “This isn’t like trading wives from cave to cave; the amount of admixture is tiny,” says molecular anthropologist Todd Disotell of New York University in New York City. “It’s replacement with leakage.”

The power of data to overwhelm human prejudice is sometimes very awesome. And the bias which Reich and Pääbo exhibited was not unfounded; Pääbo was involved in the sequencing of the Neandertal mtDNA, and found no evidence of admixture there. These data were strong, and I believe they should now shift our assessment of probabilities in relation to earlier papers which claimed some admixture between the population which derives from the Out-of-Africa expansion, and the Others.

In the second section it is notable that Chris Stringer has discarded replacement as not viable. He uses the term “not trivial,” which means that it’s a significant finding of note which one can’t simply ignore when generating inferences from a set of facts which one takes as axiomatic. Disotell’s attempt to minimize the finding is more a matter of rhetoric. He does not dismiss the admixture, he simply consigns it to the undefined category “tiny.” To some extent this reminds me of the neutralist vs. selectionist arguments of the 1970s, and more recently of the Out-of-Africa vs. Multi-regionalism disputes in human evolutionary origins. An argument over the meaning of words is a matter of law, an argument grounded in empirical data and quantitative estimates is an argument about science. No one holds to the extreme caricatures of any four of the models at this point; we’ve established that all these paradigms are unchaste, now we’re just haggling over price. We know that humans and the Others did the deed, we’re now mapping out the what, where, and how often.

But this is not the closing of the gate of itjihad. Dienekes presents an alternative model which may explain the data:

There is an alternative explanation. It involves the emergence of Homo sapiens and Homo neanderthalensis from a common ancestor and the subsequent admixture of Homo sapiens with populations that have branched out before this divergence. This would account for increased similarity between Eurasians and Neandertals, but without the problem of explaining how “Neandertal” ancestry is so similar in Europeans and East Asians.

What about Africans? Why do they stand further away from Neandertals? The answer is simple: low-level of admixture with archaic humans in Africa itself. It is fairly clear to me that the sapiens line whose earliest examples are in East/South Africa must have been an offshoot of an older African set of populations. We are lucky that Neandertals lived in a climate conducive to bone (or even DNA) preservation, while the African populations inhabiting the tropics left no traces of their existence.

In conclusion: I am not at all convinced that the authors have uncovered evidence of Neanderthal admixture in Eurasians; the alternative explanation is that modern humans and Neandertals were related, modern humans spread from East Africa/West Asia and as they entered deeper into Africa, they interacted with archaic human populations there.

In his magisterial post John Hawks has hinted at more wheels within wheels. From a Kate Wong story in Scientific American:

Intermixing does not surprise paleoanthropologists who have long argued on the basis of fossils that archaic humans, such as the Neandertals in Eurasia and Homo erectus in East Asia, mated with early moderns and can be counted among our ancestors—the so-called multiregional evolution theory of modern human origins. The detection of Neandertal DNA in present-day people thus comes as welcome news to these scientists. “It is important evidence for multiregional evolution,” comments Milford H. Wolpoff of the University of Michigan, the leading proponent of the theory.

The new finding shows that “gene flow across taxonomic boundaries happens,” observes geneticist Michael F. Hammer of the University of Arizona. Hammer is among the minority of geneticists who have espoused the idea of gene flow between archaic and modern populations. His own studies of the DNA of people living today have uncovered, for example, a stretch of DNA that seems to have come from encounters between moderns and H. erectus.

I assume Wolpoff is exultant. I do not personally think that this finding necessarily is going to result in a renaissance in Multi-regionalism, but Wolpoff has been the subject of a rising tide of skepticism and dismissal these past few decades. But rather than a more robust discussion between a revived Multi-regionaism and Out-of-Africa, I think these findings, and those that are likely to follow, will force us to move past simplistic typologies and accept that human evolutionary history works itself out through the principles of population genetics, and so can only roughly be modeled in words. The devil is in the parameters.

May 6, 2010

Neandertal genome open thread

Filed under: Anthroplogy,Neandertal Genome — Razib Khan @ 2:21 pm

My thoughts on the topic are pretty disjointed, and I can’t come up with a post that adds anything to what others have already said. And there’s still the primary documents to digest in full. So I’m going to open this post up to stray thoughts/comments (though try to keep it at least at the level of a Neandertal cognitively). Question: what other human evolution story is of the same order of magnitude in terms of significance? I think the last one of this magnitude was the Cann, et. al. “mitochondrial Eve” narrative of the mid-80s. Before that Lucy?

Neandertals, admixture, etc.

Filed under: Evolution,Genetics,Genomics,Neandertal Genome,Neandertals — Razib Khan @ 10:12 am

Read Carl Zimmer’s post, Skull Caps and Genomes. The papers aren’t on the Science website yet. And of course, Google News. I think I’m just going to wait on the papers before I say much more….

Good time to be alive, no?

Update: John Hawks.

Update II: Science got the page up.

Powered by WordPress