Archaic humans

Four makes a party

Adding the first high-quality Neanderthal sequence to genomic comparisons of archaic and modern humans sheds light on gene flow, population structure and adaptation, and suggests the existence of an unknown group. See Article p.43

Archaic humans have captured the popular imagination since the nineteenth century, when the remains of Neanderthals were discovered in the Neander valley of northern Germany and elsewhere in Europe. Until recently, Neanderthals and other archaic humans were known only from bones and various artefacts, but DNA-sequencing technology is now providing us with new perspectives on these early groups and their relationships to modern humans. In this issue, Prüfer et al.1 (page 43) report the first high-quality genome sequence of a Neanderthal individual. Their work adds to an emerging story about a tangled web of gene flow among modern humans and different early hominins (humans and archaic groups that are more closely related to humans than to chimpanzees), and hints at the existence of an unknown, highly diverged hominin group that contributed to this archaic gene pool.

Neanderthals are thought to have persisted in southern Europe until around 30,000 years ago2, thus potentially overlapping with modern humans. As a result, there has long been interest in whether Neanderthals might have interbred with early Europeans. In the 1990s, the first comparisons of DNA sequences from modern humans and Neanderthals3,4 suggested a rather simple story: that modern humans emerged from Africa during the past 100,000 years, and spread around the globe without receiving genetic contributions from hominins that had left Africa much earlier.

These early studies were based on sequences from mitochondrial DNA, which is easier than nuclear DNA to capture in ancient samples but represents only a tiny fraction of the human genome. However, the past few years have seen a revolution in our ability to obtain nuclear-genome sequences from ancient samples5,6,7,8,9, and these data are providing startling insights. One surprise was the first clear evidence for interbreeding between Neanderthals and modern humans5; another was the discovery of a second type of archaic hominin in Eurasia in addition to Neanderthals. This group, dubbed the Denisovans, is known mainly from the genome sequence of a single finger bone found in a cave in the Altai Mountains in Siberia6,7.

Although the Neanderthal bone from which Prüfer et al. derived their genomic sequence was found in the same Siberian cave, its owner is estimated to have lived several thousand years earlier than the Denisovan individual, and the two populations that the individuals represent are not closely related. The ancestors of Neanderthals and Denisovans diverged from the main human lineage about 600,000 years ago, and then split from each other around 400,000 years ago (Prüfer et al. discuss these estimates and associated caveats in detail). Thus, Neanderthals and Denisovans were quite distinct populations, having been separated for roughly three times longer than any modern human populations.

Prüfer and colleagues' sequence comparisons provide further detail about the extent of interbreeding between the different hominin groups living during the Pleistocene period (see Fig. 8 of the paper1). The authors offer a more confident estimate of the Neanderthal contribution to the genomes of modern humans: about 2% for non-Africans (Africans have no detectable Neanderthal ancestry). They also report gene flow from Neanderthals into Denisovans that includes input at functionally important genomic regions involved in immunity and sperm function. Earlier work had shown that the main Denisovan contribution to modern humans is found in some populations in Oceania and, to a lesser extent, in east Asians6,7.

Most provocatively, Prüfer et al. find evidence for modest levels of gene flow into Denisovans of sequence that is different from that of any known group, implying that there is at least one more, so far undiscovered, archaic-hominin group (Fig. 1). Low levels of gene flow have been observed in other radiations of species, so evidence for inter-hominin breeding should not be a tremendous surprise10; however, it does seem that Eurasia during the Late Pleistocene was an interesting place to be a hominin, with individuals of at least four quite diverged groups living, meeting and occasionally having sex.

Figure 1: Gene flow from an unknown ancient population.

Prüfer et al.1 calculate that modern Africans show greater genomic similarity to Neanderthals than to Denisovans. Average sequence divergence along the lineage leading to modern Africans is 7.47% since the last common ancestor with Neanderthals, and 7.71% since the last common ancestor with Denisovans (both numbers represent the percentage of divergence since the human–chimpanzee split). This difference is highly significant, and is inconsistent with a simple model in which the entire Neanderthal and Denisovan genomes come from the same source population. The best alternative model identified by the authors is that there was flow of a small contribution of genomic material (0.5%–8%) into Denisovans from a highly diverged, unknown population.

The Neanderthal and Denisovan genomes also share another intriguing feature: they both have extremely low genetic diversity, with only about two heterozygous sites (sequence differences between the paired homologous chromosomes) per 10,000 nucleotides. This equates to only around one-quarter of the genetic diversity of modern humans. The Neanderthal individual sequenced by Prüfer et al. had reduced heterozygosity in part because she was inbred (her parents were as related as half-siblings). However, the authors' analysis suggests that the primary cause of the low variability is that both groups had extremely small effective population sizes for the preceding 100,000 years or more.

Not only are these diversity estimates low compared with the genetic diversity of modern humans, they are also among the lowest levels of genetic diversity reported for any organism11. These small population sizes seem paradoxical given the large geographical range of Neanderthals (and perhaps also of Denisovans), but they suggest that the population densities of these hominins were extremely low. Might these archaic hominins have been on their way to extinction even in the absence of any competition they may have experienced from modern humans?

The new Neanderthal genome will also provide insight into the evolution of modern humans. Prüfer et al. report that there are just 96 protein-coding positions at which the Neanderthal sequence differs from that of all modern humans, with around a further 35,000 such differences at non-coding positions, some of which may affect gene regulation. This catalogue is an intriguing starting point for studying the functions of genetic differences between these groups; for example, this list is short enough to imagine creating cell lines or mouse models that contain each specific change. However, one must be mindful that many human attributes, such as bipedal gait and complex culture, probably evolved before this period of hominin diversification, and that additional important variants may lie in parts of the genome that are difficult to sequence using current methods.

After years of challenges, ancient-DNA studies are coming into their own, but they are raising as many questions as they answer. How many distinct archaic hominin groups were around in the Late Pleistocene? What were their geographical distributions? How did they help to shape the genetic make-up of modern humans? The recent sequencing of a 24,000-year-old Siberian specimen9 and the recovery of mitochondrial DNA from a 400,000-year-old hominin12 are examples of how each new ancient genome adds significantly to our understanding of both recent and more distant human history. We can expect many more exciting stories in the coming years.


  1. 1

    Prüfer, K. et al. Nature 505, 43–49 (2013).10.1038/nature12886

    CAS  ADS  Article  PubMed  PubMed Central  Google Scholar 

  2. 2

    Finlayson, C. et al. Nature 443, 850–853 (2006).

    CAS  ADS  Article  Google Scholar 

  3. 3

    Vigilant, L., Stoneking, M., Harpending, H., Hawkes, K. & Wilson, A. C. Science 253, 1503–1507 (1991).

    CAS  ADS  Article  Google Scholar 

  4. 4

    Krings, M. et al. Cell 90, 19–30 (1997).

    CAS  Article  Google Scholar 

  5. 5

    Green, R. E. et al. Science 328, 710–722 (2010).

    CAS  ADS  Article  Google Scholar 

  6. 6

    Reich, D. et al. Nature 468, 1053–1060 (2010).

    CAS  ADS  Article  Google Scholar 

  7. 7

    Meyer, M. et al. Science 338, 222–226 (2012).

    CAS  ADS  Article  Google Scholar 

  8. 8

    Keller, A. et al. Nature Commun. 3, 698 (2012).

    ADS  Article  Google Scholar 

  9. 9

    Raghavan, M. et al. Nature (2013).

  10. 10

    Bachtrog, D., Thornton, K., Clark, A. & Andolfatto, P. Evolution 60, 292–302 (2006).

    CAS  Article  Google Scholar 

  11. 11

    Leffler, E. M. et al. PLoS Biol. 10, e1001388 (2012).

    CAS  Article  Google Scholar 

  12. 12

    Meyer, M. et al. Nature (2013).

Download references

Author information



Corresponding authors

Correspondence to Ewan Birney or Jonathan K. Pritchard.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Birney, E., Pritchard, J. Four makes a party. Nature 505, 32–33 (2014).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.