Figure 1: Tourists in front of the Denisova Cave, where the Denosivan bone fragment was found. Image courtesy of Wikimedia.

Recent advances in DNA preparation and sequencing have enabled high-quality analysis of archaic human genomes. Previously sequenced genomes of archaic humans, including the Neanderthal genome in 2010, were relatively low in quality due to the degraded state of the DNA samples. Information that could be inferred from these low-quality genomes was limited. A new method, invented in 2012, enabled a higher-quality rendition of the previously sequenced Denisovan genome. Analysis of the high-quality Denisovan genomic sequence suggests inbreeding between ancestors of Denisovans and ancestors of modern humans. The high-quality sequencing also gave insights to Denisovan population characteristics.

Initial Sequencing

In March 2010, scientists found a humanoid finger bone belonging to a juvenile female who lived 50,000 years ago in Denisova cave in the Altai Mountains, Siberia (1). An international group led by Svante Pääbo, a paleogeneticist at the Max Planck Institute, set out to sequence the DNA fragments left in the finger bone. From the initial analysis published in May 2010, the group found that the finger bone was neither Neanderthal nor modern human but rather an undiscovered archaic human. The archaic human was named Denisovan after the location where the bone fragment was found. Due to the age and small amount of DNA remaining in the finger bone, the group was only able to achieve 1.9x coverage (every nucleotide in the genome was sequenced 1.9 times on average) (1). This means the genome was too low in quality to be used to accurately document the finer genetic differences between Denisovans and modern humans.

New Method

In August 2012, the group published a new, high-quality version of the Denisovan genome. A new preparation method for DNA libraries, a collection of DNA fragments stored for sequencing, developed by Matthias Meyer allowed the group to increase the Denisovan genome coverage (2). DNA for sequencing is usually prepared in the double-stranded state. However, sequencing single-stranded DNA might produce more data from archaic DNA sources. The denaturing of double-stranded DNA into single-stranded DNA effectively doubles representation in the library (2). The protocol did not use DNA purification methods to prevent a possible loss in yield.

The libraries were sequenced with the Illumina Genome Analyzer platform. This produced reads, or short sequences of DNA, that were about 35 base pairs. The short reads were aligned to the modern human genome. This resulted in around 31x coverage of the approximately 1.86 billion base pairs of the human autosomal genome to which the reads were mapped. 99 percent of the nucleotides have been sequenced 10 or more times (2).

The group calculated the average error rate to be 0.13 percent of nucleotides sequences. This was done by comparing the constructed sequence with highly conserved regions within primates, with the assumption that deviances from highly conserved base pairs were errors. The group inferred 0.07 percent male modern human DNA contamination (2). Since the Denisovan DNA sample came from a female individual, any alignments to the human Y chromosome were considered contaminants.

Figure 2: The Denosivan fossil finger (a replica shown here with a U.S. penny). Image courtesy of Max Planck Institute for Evolution and Anthropology.


For comparison, the group also sequenced the genomes of 11 modern day humans from various ethnic groups: a San, Mbuti, Mandenka, Yoruba, and Dinka from Africa; a French and Sardinian from Europe; a Han, Dai, and Papuan from Asia; and a Karitiana from South America (2). The group identified sites that are variable in a modern West African individual, who is not thought to be influenced by Neanderthal or Denisovan genetic flow (3), and counted the occurrence of derived alleles in Neanderthal or Denisovan but not in Chimpanzee. The group estimates that the Denisovan and modern human populations diverged around 170,000 to 700,000 years ago (large span due to uncertainty of human mutation rates) (2).

From the initial analysis in 2010, the group found Denisovan DNA in Southeast Asian populations and concluded that ancestors of modern humans probably interbred with the Denisovans (1). Out of the 11 modern human populations sequenced in the new study, Papuans derive the most genetic material (6 percent) from Denisovans  (2). Papuans also share more alleles with Denisvan genomes on the autosome chromosomes than on the X chromosome. A possible explanation is that genetic flow introduced to Papuan populations originated from Denisovan males.

The high quality of the new Denisovan genome allowed accurate measurement of its heterozygosity, or the percent of nucleotides that are different between the maternal and paternal genomes. Heterozygosity can also be used to measure genetic variation in the population. The group estimated that Denisovan heterozygosity to be 0.022 percent (2). This is a low percentage compared to the heterogyzosity of modern humans. Unusually long stretches of homozygosity were not detected, excluding immediate inbreeding as a possible explanation. The conclusion was that Denisovans have very low genetic diversity compared to that of modern humans. This implied that the Denisovans had a dwindling population at the same time when modern humans had an expanding population.

Although no phenotype information exists about Denisovans, the group could infer that the Denisovan girl had brown hair, skin, and eyes. The Denisovan female sequence carried alleles that are associated with dark pigmentation in modern humans.

The group also concluded that Denisovans contained 23 chromosomes, not 24 as in chimpanzees. A fusion of two chromosomes created human chromosome 2. The fusion event is characterized by a human-specific DNA sequence repeat. There are 12 occurrences in the Denisovan genome where sequences contained the unique repeat. The same test performed with the chimpanzee genome failed to produce the unique repeat. The group concludes that Denisovans shared the same two chromosome fusion event as shown in modern humans (2).

The sequencing of an archaic human allows comparison with modern humans for the identification of traits unique to modern humans. The group identified 111,812 single nucleotide changes (SNCs), including 269 changes that cause amino acid changes, and 9,499 nucleotide insertions and deletions distinguishing modern human genomes from that of Denisovans and chimpanzees. These changes were found in strongly conserved genes, including genes associated with brain function. One of the genes, CNTNAP3, is associated with speech formation. There is speculation that critical components of speech and synaptic transmission may have only arisen in modern humans. There were also 34 differences in genes associated with skin and eye diseases in modern humans. It seems possible that physiology of the skin and eye have diverged since the last common ancestor between Denisovans and modern humans (2).


There are many unresolved questions about the Denisovan humans. The observation that the Denisovans have low genetic variability conflicts with the diversity of the Denisovan’s habitats, which ranged from Siberia to south eastern Asia and Austrialia. While it is possible that the Denisovans had a quick population growth phase, more research is needed to determine their true geographic range and to see whether populations in Siberia and populations further south were genetically related.

The new method employed in the 2012 study will provide a path for other ancient humans to be sequenced. Two Denisovan teeth were also discovered in the Denisova cave. More Denisovans could be sequenced to determine the population characteristics of the ancient humans. The new method for sequencing archaic DNA could be used to improve the Neanderthal genome. A new high-quality sequence of Neanderthal genome could be used to determine the timeline of Neanderthal population expansion. If the Neanderthal expansion timeline is consistent with the Denisovan timeline, the population model could be used to inform the diversity of the population in the Out-of-Africa theory.


Although Denisovans are thus far only represented by one finger bone and two teeth, they are currently the most well-known archaic human genetically – including Neanderthals of which there are hundreds of fossil records (4).

The differences between modern and archaic humans are important to catalog in order to further understanding of why archaic humans eventually became extinct and how modern humans expanded their population.


Contact Xiongfei Zhang at



1. D. Reich et al., Nature. 468, 1053-1060 (23 December 2010)

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

3. Green et al., Science 328(5979), 710-722 (2010).

4, Ann Gibbons, Science. 337(6098), 1028-1029 (2012).