Mailbag: Denisova and the hobbit

Re: Denisova at high coverageā€

Dear John I can hardly keep up with all the archaic human stuff coming out - partly as my main occupation is health genetics - but I wanted to run something past you. Given that there is reasonable agreement that Papuans and to a lesser degree Han Chinese (and peoples in between) have Denisovan ancestry, is there still any evidence that they have Neandertal ancestry? Or is the signal of presumed Neandertal ancestry just a signal of archaic ancestry which is in fact from Denisovans? Of course, if there is no strong evidence of Neandertal ancestry in Papuans and Chinese then it removes the need for the interbreeding to have occurred immediately out of Africa. It already seems that modern humans interbred with Neandertals and Denisovans (and perhaps with some other archaic hominin in Africa), but it seems to me that the place and time of Neandertal interbreeding could be significantly different from where first proposed? Then the timing of the Denisovan admixture would be of interest - I haven't read today's paper yet. Did you or anyone look in Japan and Tibet (and indeed the Andamanese) where we see relict Y chromosome lineages (and to a lesser extent relict mtDNA lineages eg M12). Perhaps they could have higher proportions of Denisovan as they have been less affected by the expansions accounting for the majority of Han ancestry? Perhaps not, but there is GWAS data available for all, so methods using this could be applied. Finally - getting out on a limb - I wonder if the Flores hominins were Denisovans? Can only be speculation, but has anyone commented on this. Thanks for your thoughts, I hope you'll spare a second to reply.

Many thanks for your kind words. It is indeed hard to keep up, even for us!

We are very confident that these populations do have Neandertal ancestry and that it does not all come via Denisovans. The Neandertal and Denisovan genomes each have a large number of unique alleles that allow us to differentiate these signatures. In Australians and PNG, the fraction of around 6% Denisovan is in addition to around 3% Neandertal.

I agree, Tibet is a promising area. The paper by David Reich et al. last year did include a broad range of Chinese regional and ethnic minority populations, including Tibet, and did not find any evidence of Denisovan admixture there. This is very strange to me, and it remains to be satisfactorily explained.

I doubt that Flores was Denisovan although it is possible. The initial habitation of Flores more than 1 million years ago was too early to have been accomplished by the same people that gave rise to Denisovans. The Denisovans diverged from Africans and Neandertals less than 600,000 and maybe as recently as 300,000 years ago. However, the initial habitation of Flores is known only from archaeology, and the hobbit itself is much more recent, only 18,000 years old. So possibly these are not a single population, in which case it is conceivable that the hobbit is Denisovan-derived.