denisovan
© MAAYAN HARELThis artist's reconstruction, based on anatomical estimates from a new method, shows the face of a Denisovan girl from Siberia in Russia
A team of European researchers led by DeCode Genetics' Kari Stefansson examined the effects of the admixture of modern humans and Neanderthals, and found that Icelandic genomes had more Denisovan-like DNA fragments than expected.

As the researchers described in a study published on Wednesday in Nature, they analyzed 14.4 million putative archaic chromosome fragments that were detected in fully phased whole-genome sequences from 27,566 Icelanders. These fragments corresponded to a range of 56,388 to 112,709 unique archaic fragments that covered 38 percent to 48.2 percent of the callable genome. On the basis of similarity with known archaic genomes, the researchers assigned 84.5 percent of the fragments to an Altai or Vindija Neanderthal origin and 3.3 percent to Denisovan origin, while 12.2 percent of the fragments were of unknown origin.

The unexpectedly large proportion of Denisovan DNA in the Icelandic genomes is likely explained by gene flow either into ancestors of the introgressing Neanderthals or directly into humans, the researchers said.

"One possible explanation is introgression from groups with genomes that are related to the Denisovan and Altai genomes into introgressing Neanderthals before this group contributed to modern human genomes," the authors wrote. "Alternatively, the introgressing Neanderthals could have been a Vindija-like group that carried some anciently diverged haplotypes due to incomplete lineage sorting that now seem, by chance, to be more similar to the Altai and Denisovan genomes than the Vindija genome."

The researchers performed extensive simulations under different demographic and admixture models and found that the observed characteristics of Denisovan-like fragments in Icelanders were not compatible with a simple introgression from a Vindija-like group without that population having had prior admixture with a Denisovan-like group. They also theorized that there could have been direct admixture from a Denisovan-like group into the common ancestors of non-Africans before the main Neanderthal admixture event.

When they compared within-individual, paired archaic fragments to syntenic non-archaic fragments, they found that the overall rate of mutation was similar in humans and Neanderthals during the 500,000 years that their lineages were separate, but also that there were differences in the relative frequencies of mutation types, perhaps due to different generation intervals for males and females.

The team also studied the influence of archaic variants on the phenotypic diversity of contemporary humans by assessing 271 phenotypes. After several filtering and analysis steps, they found only five independent archaic variants that were likely to be a true source of phenotypic association.

One was an example of a strong association with a reduced level of prostate-specific antigen driven by rs17632542, which has been reported to reduce the risk of prostate cancer; however, the introgressed status of the variants was not reported. Another was rs28387074, which decreases the concentration of hemoglobin and survives adjustment for non-archaic variants that are responsible for another strong signal in the same region. The remaining three association signals due to archaic variants were for reduced height, decrease in mean corpuscular hemoglobin, and an increase in plasma prothrombin time.

"We show that a large part of an archaic genome can be mined from contemporary descendants of populations that were recipients of introgression around 50 to 60 thousand years ago," the authors concluded. "The recovered archaic fragments are consistent with being descended from multiple archaic individuals, who belonged to an archaic population similar to the Vindija Neanderthal.

However, the considerable proportion of archaic fragments that are closer to the Denisovan genome cannot be explained by incomplete lineage sorting. Rather, they require Denisovan introgression, either directly into humans or into Neanderthals who later mixed with humans, which must have occurred soon after they migrated out of Africa, because its signal is found in all contemporary non-African populations from the Simons Genome Diversity Project. This raises the possibility that there were Denisovan-like groups west of the Altai mountains, where such gene flow into humans must have occurred."