Max-Planck Institute aims to build Neandertal genome

5 minute read

It's a short story from the AP, but here's the lede:

FRANKFURT, Germany - German and U.S. scientists have launched a project to reconstruct the Neanderthal genome, the Max-Planck Institute for Evolutionary Anthropology said Wednesday.
The project, which involves isolating genetic fragments from fossils of the prehistoric beings who originally inhabited Europe, is being carried out at the Leipzig-based institute.

There are no details.

Now, what is the real story here? I suggest some points to consider:

  1. The head of Max-Planck Evolutionary Anthropology is Svante Pääbo, who is the world's leading expert on DNA recovery from ancient remains. In my assessment, he is very unlikely to announce a project like this one without some advance knowledge that there will be results.
  2. The recovery of a complete mitochondrial genome from Neandertals is certainly achievable. But nobody from Max-Planck or anywhere else has yet published such a sequence. It would also be actually informative, in that it would give us some information about the amino acid sequences coded by human mtDNA prior to the common ancestor of all living human mtDNA. This would be the opportunity to see which evolutionary changes in mtDNA might have been targets of recent selection.
  3. But, there would be no reason for a press release about a complete mtDNA genome in the absence of details. Anybody can see that a complete mtDNA sequence is achievable, and while it would be important news, it would not justify this kind of attention.
  4. Therefore, they really are talking about recovering nuclear DNA information from Neandertals.
  5. Referring to point (1) above, this means that someone in Pääbo's lab has very likely already recovered some nuclear sequence from a Neandertal.
  6. That lab has, in recent publications, focused at some length on the problem with contamination inherent in Neandertal remains, and has been skeptical of results from other labs reporting modern human sequences from post-Neandertal modern Europeans. The clear trend over the past two years is to reject any modern sequence as possible contamination.
  7. Using this standard, we may conclude that at least some nuclear DNA thus far recovered from some Neandertal specimen exhibits some interesting difference from living humans.

Much of this train of logic follows directly from point (1), that Pääbo is unlikely to make an announcement like this without already knowing some results. This assumption is pretty justifiable for anybody familiar with the literature. After all, Pääbo has himself been highly critical of botched ancient DNA recoveries by other labs. Also, the very existence of news stories suggests strongly that a negative result is not the end point.

If you have any lingering doubts, then read this 2003 story from US News and World Report, where Richard Klein says "Svante is going to be the first anthropologist to win a Nobel Prize." I think my assessment is very likely; this is not someone who is going to be associated with a highly publicized failed effort.

Assuming that some Neandertal nuclear DNA is coming, what would be useful? First of all, the stated goal of "reconstructing the Neandertal genome" is completely impossible, unless by this they mean the mtDNA genome. The focus will be on highly polymorphic sites that contain lots of single nucleotide (or other short) polymorphisms. This is a good reason to expect there will not be Y chromosome evidence, despite its probable interest, because the Y is just not very polymorphic. Longer length polymorphisms, like microsatellites, likewise are not very realistic because they would be very difficult to clone without error from a damaged sequence, and they extend over dozens to hundreds of base pairs, making it difficult to find them (along with enough flanking sequence for primers) intact. It is possible that they will focus on the X chromosome, which is considerably easier to deal with in males because they do not carry two potentially different copies.

The twist is that you are not going to confirm the DNA is real (and not contaminated) unless you can show a difference from living humans. My gut tells me that for most nuclear genes those kinds of differences are unlikely to be found. A real advance in studying these ancient remains would be some theoretical movement on the likelihood of observing certain kinds of presently rare alleles in the same individual. If I were doing this work, I would work out the odds of observing certain combinations of biallelic sites in present-day Europeans, and then assess the same sites in a Neandertal fossil. This would allow a more-or-less standard admixture study, albeit with a very small ancestral sample. Or, I would try to find specific short functional alleles that are polymorphic in humans. For example, MC1R would be very interesting, because you could try to find haplotypes selected for skin or hair color. This story would be about the antiquity of selection for skin color, although it might not necessarily demonstrate a Neandertal origin for light skin in Europeans.

But I don't think they'll do what I would do.

My best guess is that we will see first the sequence for FoxP2 from one or a few Neandertal specimens.

Pääbo's lab has done much of the molecular work on FoxP2 (OMIM), including assessment of the variation in other primates. This means they have primers for the gene that work both within humans and among different primate species. Furthermore, the gene underwent a recent selective sweep (dating to the last 200,000 years) in humans, meaning that Neandertals are fairly likely to lack the present human allele. Humans differ from chimpanzees by two amino acid sites, so it is likely that one of these differs between recent humans and their ancestors. Although it is possible that Neandertals are identical to humans, there is still good reason to suspect that Neandertal FoxP2 sequences will be different. And such a difference would help pass the contamination test.

It's like the perfect genetic storm.

You can expect that if this is the finding, the conclusion will be a speculation about Neandertal speech abilities. I have recently written with some colleagues (Wolpoff et al. 2004) about this issue, concluding that FoxP2 just doesn't tell us much about speech and language in Neandertals. So it will be interesting to see anthropology and genomics interacting if we get more information about Neandertal genetic variation.

To you graduate students in Leipzig, please watch out! I know that many archaeologists think Neandertals were dumb. But you can bet that if the velociraptors could learn to open doorknobs, then cloned Neandertals can too.

UPDATE (7/7/05): A reader has sent me a reference to a paper in press describing nuclear DNA recovery from cave bear remains. I discuss it in a later post. That success may explain the new confidence about the prospect for Neandertal DNA, but I'm still thinking there probably is already some positive result from them also. The scope of the bear research is very suggestive about the method, however, and it is much more akin to what I think would be informative than I feared.


Wolpoff MH et al. 2004. Why not the Neandertals? World Archaeol 36:527-546. PDF available online