How much of the Neandertal genome came from contemporaries in Africa?

Back in June of last year, Melissa Hubisz, Amy Williams, and Adam Siepel coworkers put out a preprint with a new method for looking at introgression using the ancestral recombination graph: “Mapping gene flow between ancient hominins through demography-aware inference of the ancestral recombination graph”.

The paper has not yet appeared in a journal, but at the time I noted one interesting conclusion:

Applying this method to modern and archaic hominins, we confirm that a significant proportion of the Neanderthal genome consists of regions introgressed from ancient humans. While we identified 3% of the Neanderthal genome as introgressed, a rough extrapolation based on our estimated rates of true and false positives suggests that the true amount is around 6%. Thus, the Neanderthal genome was likely more influenced by introgression from ancient humans, than non-African human genomes are by Neanderthal introgression. Our follow-up analysis suggests that the Hum→Nea gene flow occurred between 200-300kya. This time estimate is largely based on the frequency of introgressed elements among the two diploid Neanderthal genomes, and thus will be sensitive to the accuracy of the demographic model we used for simulation, as well as other factors such as mutation rate and generation time.

With the first sequencing of the first Neandertal genome, and for several years afterward, many geneticists promoted a scenario in which gene flow between Neandertals and modern humans had been a one-way arrangement. The idea was that moderns got some DNA from Neandertals, but the Neandertals never got any from modern humans. This conclusion was based on some data, but it was premature. The early methods applied to the Vindija low-coverage genomes could not detect “modern” genetic input into the Neanderthal population unless that modern input came from the ancestors of some modern populations and not others. The data did rule out that the Vindija Neandertals had genetic input from the immediate ancestors of living Europeans. But the analyses could not test for older introgression from African-derived populations not closely related to one living population or another.

In the last year or two, a number of analyses have started to identify introgression of particular segments of the genome. Geneticists have also added larger samples of DNA from today’s African populations. African peoples are still badly underrepresented in genetic datasets, and the recent additions are a drop in the bucket of what is needed. But it’s impressive how a better representation of African variation and better methods have started to illuminate deeper phases of population mixture in human evolution.

One of those is the deep introgression into Neandertals from their African contemporaries. The African origin of Neandertal mitochondrial DNA, revealed a few years ago, was the first major element of our emerging understanding. Mitochondrial DNA was not alone. Neandertals were repeatedly connected to African populations in the time after 350,000 years ago. They derive a substantial fraction of their genetic variation from such contacts with African populations.

The paper from Lu Chen and coworkers earlier this year provided strong evidence of the importance of gene flow from Africans into Neandertals in the period after 150,000 years ago (“Neandertal ancestors of African populations”). This new paper from Hubisz and coworkers is pointing to gene flow in an earlier period of time, more similar to that time when the Neandertal mtDNA introgressed.