NEW YORK (GenomeWeb) – A new analysis of archaic hominin sequences in human genomes is providing a refined look at past human relationships with Neanderthals and Denisovans, as well as the types of archaic sequences remaining in present-day human populations.
As they reported in Science today, researchers from the University of Washington, the Max Planck Institute for Evolutionary Anthropology, and elsewhere developed a method for picking potential archaic hominin sequences out of more than 1,500 human genomes — a global collection made up of sequences from the 1000 Genomes Project and 35 new genomes they sequenced to an average of 40-fold coverage, representing individuals from 11 sites in Papua New Guinea's Bismark Archipelago in Northern Island Melanesia.
The team's analysis of archaic hominin sequences reflected widespread Neanderthal mixing with the ancestors of human populations outside of Africa, likely over multiple admixture events. Denisovan DNA was less common in human genomes, mainly turning up in individuals from Melanesia and other parts of Papua New Guinea. And by digging into parts of the human genome where Neanderthal or Denisovan DNA has persisted or been purged, the group was able to refine its view of the impact of such mixing events.
"We've known for a long time — maybe five years or something like that — that non-Africans have Neanderthal ancestry of around 2 percent. Around that same time, we figured out that Melanesians have also some amount of Denisovan ancestry," first author Benjamin Vernot told GenomeWeb. Vernot was based at the University of Washington when the study was done, and is currently a post-doctoral researcher in Svante Pääbo's lab at the Max Planck Institute for Evolutionary Anthropology.
"The big difference here is going from saying, 'This individual has 3 percent Denisovan ancestry' to saying, 'This individual has genes X, Y, and Z that come from Denisovans' — actually mapping them out on the genome," Vernot explained. "That sort of thing had not been studied before in these populations and you can learn a lot more by identifying the particular [archaic] fragments in the genome."
A series of past studies have started to untangle the types of Neanderthal sequences that remain in non-African populations and their potential roles in human health and disease. But adaptive introgression by sequences from the Denisovan has yet to be explored in the same detail in modern humans, the authors noted in their paper.
To get a clearer look at archaic admixture patterns across populations, Vernot and his colleagues first came up with a strategy for finding archaic hominin sequences in present-day genome sequences. Scanning each human genome, they highlighted sequences suspected of having archaic hominin origins due to their enhanced diversity in non-Africans over Africans — bucking the typical trend of elevated diversity in sequences from African individuals.
Since ancestral populations in Africa are thought to have experienced little or no admixture with Neanderthals or Denisovans, Vernot explained, present-day African genomes should be largely free of sequences from either archaic hominin.
While the first stage of the archaic sequence search might pick up sequences from still-to-be-discovered archaic hominins, he noted that such sequences would have been weeded out when compared side-by-side to the Neanderthal and Denisovan genomes in the second stage of the analysis.
Finally, because the Neanderthals and Denisovans are expected to share some haplotypes due to their closer relationship to one another than to humans, the team performed an additional computational step to classify the potential archaic hominin sequences as Neanderthal, Denisovan, or neutral/ambiguous.
In total, they searched 1,523 worldwide human genomes with this archaic hominin sequence-finding strategy, and detected more than a billion bases of predicted Neanderthal sequence and 304 million potential Denisovan bases.
Large chunks of the human genome had lower-than-usual Neanderthal sequence content, while Neanderthal sequences were enriched in other genome sites. Denisovan sequences were largely restricted to Oceanic individuals, showing up to a lesser extent in the genomes of individuals from East Asia or South Asia. The team noted that the newly sequenced Melanesian genomes, on average, contained tens of thousands of bases from both Neanderthals and Denisovans.
As it turned out, though, Denisovan sequences were missing from many of the same genome sites designated as Neanderthal-depleted in other populations, consistent with the idea that there may have been selection against archaic sequence introgression in certain parts of the human genome.
"The Melanesian population does not have Neanderthal ancestry where the other non-African populations don't have Neanderthal ancestry, so that's encouraging. And they also don't seem to have Denisovan ancestry in those regions," Vernot said. "That strengthens the argument that there's really something different between humans and archaics in those regions that might be driving that signal."
Intriguingly, at least some of the sites where archaic hominin sequences are missing are also home to genes linked to language or brain development, genes that are highly expressed in the brain, and/or genes mutated in individuals with autism spectrum disorder. On the other hand, the retained Neanderthal and Denisovan sequences were enriched for genes contributing to metabolic and immune pathways.
Vernot cautioned that while such potential ties are tantalizing, it is too soon to say how significant they are, particularly for the Neanderthal and Denisovan sequence-depleted regions, which span large regions containing hundreds of genes.
Members of the teams plan to take a closer look at positive and negative selection on archaic haplotypes. They also expect to keep tweaking the models of mixing between modern humans and their archaic cousins as more and more human populations are profiled and new archaic hominin samples are discovered and sequenced.
From the latest data, for example, the team estimated that there were at least three distinct admixture events involving Neanderthals and humans, including one round of mixing that modern Melanesian's ancestors seem to have missed out on.
"It was not just this simple thing where we ran into Neanderthals as we left Africa in the Middle East," Vernot said. "Maybe we ran into them several times and it was more complicated than it could have been."