Doing population genomics with archaic hominins is tricky. We’re severely constrained not only by the limitations of the fossil record, but also the vagaries of DNA preservation. Preservation of DNA within fossils depends on incredibly finicky burial conditions, and there’s a current upper limit of DNA preservation of 700,000 years (although that date theoretically could be pushed back a bit further). Adding to the preservation difficulties is the unfortunate fact that anything and everything poses a contamination danger to the fossil genomes. You’ve got to be pretty comfortable with constant failure to do this kind of work.
But every once in a while, conditions are just right for something truly spectacular to be discovered. This was the case back in 2010, when a fossil hominin finger phalanx from Late Pleistocene deposits in the Denisova cave in the Altai mountains of Siberia yielded a mitochondrial genome that was unlike any previously sequenced. Phylogenetic analyses of the mitochondrial genome showed that it last shared a common ancestor with humans and Neanderthals 1 million years ago. Subsequent analysis of the nuclear DNA further refined the phologenetic relationship of the Denisovan individual and other hominins, showing that it was a sister group to Neanderthals. Denisova 3 (as the phalanx was named) shared some alleles with contemporary Melanesians (about 4-6% of their genome), suggesting that there had been ancient admixture between the archaic hominins and the ancestors of New Guineans, Australians, and Mamanwa (Reich et al. 2011). Interestingly, the population histories of mitochondrial and nuclear genomes from the Denisovan 3 individual were different, suggesting that at least part of the ancestry of the Denisova 3 was derived from admixture with an unknown hominin (i.e. not Neanderthal or H. sapiens).
A single genome can provide information about many ancestors in a population, but confident inferences of population history from just one individual can be problematic, particularly if he or she comes from a group with population structure or inbreeding. Conclusions derived from these data have to be subjected to testing (and possible revision) with the addition of other genomes. Unfortunately, “other genomes” is a lot to demand from the fossil record for reasons I discussed above. But the Denisova cave has been remarkably kind to us, and yielded other archaic hominin remains. A Neanderthal genome was recovered from a toe phalanx, showing evidence of admixture between Neanderthals and Denisovans (Prüfer et al. 2014). Another mitochondrial genome, very closely related to Denisova 3, was sequenced from a molar found in the cave (Krause et al. 2010). And a third Denisovan individual (Denisova 8), represented by another molar, was discovered in the cave a decade later. The nuclear genomes from Denisova 4 and 8, and the mitochondrial genome of Denisova 8, were sequenced and published last week in PNAS (Sawyer et al. 2015).
The genetic data from these two individuals gives us additional clarity in understanding the relationships between humans, Denisovans, and Neanderthals. The mitochondrial genomes of Denisova 4, 3 and 8 form a clade, with 3 and 4 very closely related to each other. From the autosomal genomes, we have confirmation that the Denisovans are more closely related to Neanderthals than to us. The three Denisovan individuals have low nuclear genomic diversity compared to modern humans, although that of Neanderthals was even lower. The autosomal genome of Denisova 8, like Denisova 3, shares a small number of alleles with contemporary peoples of Oceania, but the Denisova 4 genome did not yield enough high quality data to assess this question.
Interestingly, Denisova 8 has considerable mitochondrial and autosomal divergence from individuals 3 and 4 and is also much older (estimated by various methods to be as much as 60,000 years older). The authors speculatively attribute the genetic differences to possible gene flow from other, as yet unknown archaic hominins (such as was inferred from the Denisova 3 genome), but this needs to be addressed with genomes from additional individuals.
So now we have three. Considering that we didn’t even suspect the existence of Densiovans until the first genome was sequenced in 2010, it’s a remarkable step forward. But there are too many unresolved questions for me to be satisfied with just these three. Most importantly: what was the mysterious ancient hominin who left traces of ancestry in the Denisovan genomes? Was it Homo erectus? Some other, unknown hominin?
It seems presumptuous to ask Denisova Cave to give us more genomes, but we do need them to solve this mystery. Even better would be sequences from Denisovan individuals from sites elsewhere in Asia. After all, the finding of admixture between Denisovans and the ancestors of New Guineans, Australians, and Mamanwa indicates that they must have been more widely distributed than just this single cave. And while I’m writing my Denisovan Christmas wish list to the fossil record, I’d also like to ask for additional skeletal elements so we could have a formal description and taxonomic classification.
I have no good reason to believe that any of these requests will be granted, but if a decade of working with ancient DNA has taught me anything, it’s to be optimistic in the face of overwhelming odds.
References and further reading
Dabney et al. 2013. Ancient DNA damage. Cold Spring Harb Perspect Biol 2013;5:a012567 doi: 10.1101/cshperspect.a012567
Krause et al. 2010 The complete mitochondrial DNA genome of an unknown hominin from southern Siberia. Nature 464: 894-897. doi:10.1038/nature08976
Meyer M, et al. 2012. A high-coverage genome sequence from an archaic Denisovan individual. Science 338(6104):222–226.
Prüfer K, et al. 2014. The complete genome sequence of a Neanderthal from the Altai Mountains. Nature 505(7481):43–49.
Reich et al. 2011. Denisova Admixture and the First Modern Human Dispersals into Southeast Asia and Oceania. The American Journal of Human Genetics 89, 516–528
Sawyer et al. 2015. Nuclear and mitochondrial DNA sequences from two Denisovan individuals. PNAS
