By amplifying the melanocortin type 1 receptor from the woolly mammoth, we can report the complete nucleotide sequence of a nuclear-encoded gene from an extinct species. We found two alleles and show that one allele produces a functional protein whereas the other one encodes a protein with strongly reduced activity. This finding suggests that mammoths may have been polymorphic in coat color, with both dark- and light-haired individuals co-occurring.
Missing from the story is this: it's the first insight about phenotypic variation that anyone has derived from ancient DNA.
It's a very short paper; the essential elements are that they found the polymorphism by trying to reconstruct the Mc1r sequence in an individual that turned out to be a heterozygote for three amino acid-coding sites. They then sequenced the polymorphisms in "additional specimens" including some homozygotes, which allowed them to work out the linkage among the polymorphisms. Sequencing of additional individuals helped to rule out the possibility that the candidate polymorphisms resulted from idiosyncratic DNA damage and not genuine biological variation. One of the three polymorphisms is frequently variable among other mammal species; two of them are highly conserved -- pointing to a possible important functional difference. They then constructed a model allele including these polymorphisms to see how it would be expressed in a cell culture; they found that one of the mammoth alleles showed only partial activity. From this they inferred that the polymorphism probably had phenotypic effects on pigmentation.
Of course it's a bit of a stretch from this conclusion to blonde hair; but they are helped in this by a little-known fact: frozen mammoths show variation in their fur coloration! So the genetic observations help to explain an already-known phenotypic variation.
Would they have had the confidence to make this argument if such pigment variations were not already known? Hard to say, but one consequence of this paper has to be an increase in the confidence in such conclusions for future studies, where phenotypes aren't already known.
I would say that the most important aspect of the paper is that it shows the importance of having sequences from multiple individuals. In this study, the sequencing of other individuals helped substantiate that the polymorphisms were not DNA damage. Since these sequences are broken into short pieces, reconstructing the alleles from a heterozygote is difficult if not impossible. So finding homozygotes among these individuals was necessary to reconstruct the two alleles, which gave the opportunity to assess their functional differences. And although this study didn't include a discussion of the origin of this polymorphism, reconstructing haplotypes is essential to that question also, another enterprise that is not possible without samples form multiple individuals.
So there is only a limited amount of information that can be obtained from a single individual, especially under circumstances where the DNA is fragmented or possibly damaged. Sampling multiple individuals helps these questions a lot -- even if it may introduce problems that haven't yet been considered.
And of course, from this work (undertaken in part at the Max Planck Institute for Evolutionary Anthropology) you can see some of the strategies that will ultimately be applied to the Neandertal genome. Mc1r is sort of an obvious candidate to look at, and they are certainly going to be trying to find FoxP2 polymorphisms also. But I anticipate that the story will be much deeper -- we may spend much more time figuring out the Neandertal genome than the 150 years since the Feldhofer cave discovery.
Römpler H, and 8 others. 2006. Nuclear gene indicates coat-color polymorphism in mammoths. Science 313:62. DOI link