News & Views | Published:

Evolutionary biology

Ancient genomics is born


The reality of a complete Neanderthal genome draws near, as two papers report the sequencing of large amounts of Neanderthal DNA. The results will help to answer some central questions on human evolution.

The study of ancient DNA fascinates everybody. But it has a chequered history, with several high-profile failures — such as false reports of the amplification of DNA sequences from dinosaurs1 — taking the sheen off undoubted success stories. Among these successes are the identification of human-induced changes in the genes of maize2, and the advent of methods for amplifying a single copy of nuclear DNA3. Vocal researchers from the field have argued that it is impossible to sequence the entire genome of an extinct organism, and that the very notion is pure science-fiction because such genomes are typically highly fragmented. How could so many tiny pieces of DNA be sequenced and aligned?

But two papers have now been published that may silence these sceptics. In this issue, Pääbo and colleagues4 (Green et al., page 330) describe the sequencing of more than a million base pairs of Neanderthal DNA; for comparison, the human genome is approximately 3.2 billion base pairs long. In collaboration with Pääbo's group, Rubin and colleagues5 (Noonan et al.) report in Science their use of a different method to recover 65,000 base pairs of Neanderthal DNA. These papers are perhaps the most significant contributions published in this field since the discovery of Neanderthals 150 years ago.

The Neanderthals (Homo neanderthalensis) were our closest hominid relative. Once widespread throughout Europe and Asia, they became extinct approximately 30,000 years ago. The species began to disappear from the Balkans and central Europe about 40,000 years ago6. Modern humans (Homo sapiens) are usually blamed, directly or indirectly, for their demise7. Nevertheless, it seems likely that humans and Neanderthals learnt about tools and ornaments from each other7, and there has been much speculation about possible human–Neanderthal interbreeding, although there is currently no evidence for any such genetic exchange. The results of the new studies4,5 further support this conclusion. But sequencing the entire Neanderthal genome would enable us to examine the possibility of Neanderthal interbreeding with humans in much more detail and to explore the genetic basis of functional differences between Neanderthals and humans. For example, the Neanderthal genes responsible for cognition and mate recognition would be of great interest.

Pääbo and colleagues4 have been able to recover 1 million base pairs of Neanderthal DNA through an advance in DNA sequencing known as pyrosequencing8 (Fig. 1a). This technology achieves more than a 100-fold increase in throughput compared with traditional methods. The latest pyrosequencing machines analyse 25 million DNA bases at a high level of accuracy and in a single four-hour run. Such increased throughput and the ability to directly sequence short fragments of DNA are essential if the entire Neanderthal genome is to be sequenced, as proposed by these authors4. In the past, naysayers have insisted that the reconstruction of ancient genomes is virtually impossible because ancient DNA is typically sheared into short fragments of no more than 100–200 base pairs. But the new pyrosequencing method begins by deliberately breaking genomic DNA into short sequences — so fractured ancient DNA presents no problem.

Figure 1: Sequencing Neanderthal DNA.

DNA can be isolated as short fragments from fossilized Neanderthal bones. a, Pääbo and colleagues4 sequenced more than 1 million base pairs of Neanderthal DNA using an approach known as pyrosequencing. In this method, the isolated DNA fragments are attached to beads that are placed in wells, where the DNA is sequenced directly. About 6% of the sequences obtained were identified by the authors as putative Neanderthal DNA; the remaining sequences were unidentified or represented contamination. The authors localized the Neanderthal sequences (red rectangles) to particular chromosomes by matching them to similar sequences in the human genome. b, Rubin and colleagues5 used metagenomics to sequence 65,000 base pairs of Neanderthal DNA. In this method, the DNA fragments (red for Neanderthal DNA, blue for contamination) are incorporated into loops of DNA known as plasmids, which are amplified by replicating them in bacteria. The resulting DNA library is then further amplified enzymatically. Because Neanderthal DNA is very similar to DNA from modern humans, the authors used known gene sequences of human DNA to 'trap' complementary strands of the amplified Neanderthal DNA, so isolating them from contaminating sequences. These isolated Neanderthal DNA fragments were then sequenced.

To reassemble a randomly fractured genome, it is essential to have the complete genome of a closely related organism for comparison. This acts as a kind of scaffold on which isolated DNA sequences can be 'hung'. Pääbo and colleagues4 used the human genome, mapping Neanderthal DNA sequences to particular regions on the human chromosomes. Studies of the extinct mammoth9 used the elephant genome in a similar way. The sequenced genomes from mice, fruitflies and chickens are likely to act as DNA scaffolds for a range of extinct organisms in the future.

With more than 1 million base pairs analysed, the authors4 compared their Neanderthal DNA sequence with the chimpanzee and human genomes. Assuming that humans and chimpanzees diverged 6.5 million years ago, and using a combination of likely population sizes and divergence times, Pääbo and colleagues4 estimate that human and Neanderthal DNA diverged between 465,000 and 569,000 years ago, with a best estimate of about 516,000 years ago. Interestingly, the authors also show that Neanderthals were derived from a very small ancestral population of about 3,000 individuals. This is similar to the estimated ancestral population size of humans — that is, the number of individuals needed to produce the amount of observed genetic diversity within the current population.

Rubin and colleagues5 used a 'metagenomic' approach to sequence the Neanderthal genome (Fig. 1b). Metagenomic techniques were originally developed to recover genomes from environmental samples, enabling the study of organisms that are not easily grown in the laboratory. The authors have adapted this method to recover large tracts of Neanderthal DNA. Their results agree broadly with those of Pääbo and colleagues4 — for example, Rubin and colleagues estimate the divergence time for humans and Neanderthals to be between 120,000 and 670,000 years ago, with their best estimate being 370,000 years ago.

The metagenomic approach5 recovered only small amounts of Neanderthal DNA compared with the pyrosequencing4 method. But metagenomics does offer some advantages, as it enables Neanderthal sequences that are similar to human genes to be targeted. Crucially, Rubin and colleagues5 recovered 29 out of 35 genes that were targeted in this way. So although pyrosequencing recovers larger amounts of ancient sequence, metagenomics allows the focused recovery of specific regions of interest. Which of the two methods will be the more appropriate for future applications will depend on the question being addressed.

Although these studies4,5 will not immediately answer many of the questions about the biological differences between Neanderthals and humans, they foreshadow an exciting development — the recovery of the complete Neanderthal genome. Pääbo and colleagues4 promise this within two years, and propose to sequence the Neanderthal genome six times over. This will allow any rare mistakes in sequencing to be identified using the consensus from independent reads. Multiple sequencing of genomes is standard practice — for example, the human genome was sequenced, on average, ten times over. But a big difficulty for Pääbo's team is the cost. Each pyrosequencing run is very expensive, and the authors estimate that 6,000 runs are required. However, the authors allude4 to improvements in the technique that will reduce the number of runs tenfold.

The full Neanderthal genome will provide a precise assessment of differences between us and our closest relative. It may also help to resolve a debate that started more than 20 years ago10, when studies of ancient DNA began: in evolution, how important are mutations in genes that result in structural and physiological changes, compared with mutations that affect the regulation of those genes? More fundamentally, these combined studies show that, when predicting the limits of science, one should never say never.


  1. 1

    Woodward, S. R., Weyand, N. J. & Bunnell, M. Science 266, 1229–1232 (1994).

  2. 2

    Jaenicke-Despres, V. et al. Science 302, 1206–1208 (2003).

  3. 3

    Huynen, L., Millar, C. D., Scofield, P. R. & Lambert, D. M. Nature 425, 175–178 (2003).

  4. 4

    Green, R. E. et al. Nature 444, 330–336 (2006).

  5. 5

    Noonan, J. P. et al. Science 314, 1113–1118 (2006).

  6. 6

    Pettitt, P. B. Oxf. J. Archaeol. 18, 217–240 (1999).

  7. 7

    Jankovic, I. Coll. Anthropol. 28 Suppl. 2, 379–401 (2004).

  8. 8

    Margulies, M. et al. Nature 437, 376–380 (2006).

  9. 9

    Poinar, H. N. et al. Science 311, 392–394 (2006).

  10. 10

    Higuchi, R., Bowman, B., Freiberger, M., Ryder, O. A. & Wilson, A. C. Nature 312, 282–284 (1984).

Download references

Author information

Rights and permissions

To obtain permission to re-use content from this article visit RightsLink.

About this article

Publication history

  • Published

  • Issue Date


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.