Why the Neanderthal Theory is a Game-Changer

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 loss 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).

Also See: The Neandertal Genome

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.

Interpreting Ancient DNA

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.

By Razib Khan
Reprinted with permission from Discover