This web page was produced as an assignment for an undergraduate course at Davidson College. For information please contact Sarah Pyfrom ( or Malcolm Campbell (



Review of
Targeted Retrieval and Analysis of
Five Neandertal mtDNA Genomes

In this study, the relative genetic diversity among several Neandertal DNA samples is examined and new method for extracting target DNA sequences from unwanted background DNA is proposed and explained. Using six samples from Neandertals across their geographical range, researchers have determined that Neandertals have low mtDNA diversity relative to modern humans. This suggests that there was a relatively small population of Neandertals spread across a large geographical area.
Five samples of DNA were taken from Neandertals found in Croatia (undated), Germany (Feldhofer 1 and 2 ~40,000 years before present (BP)), Spain (Sidron 1253 ~39,000 yr BP) and Russia (Mezmaiskaya 60,000-70,000 yr BP). For each individual, the entire mitochondrial genome was sequenced using small, overlapping DNA samples and a new procedure called primer-extension capture.  Specifically, they looked at the hypervariable regions (HVRs) within the mtDNA, which was most accessible and best preserved. These sequences were compared to a previously discovered Neandertal mtDNA genome that was sequenced using the much more time-expensive and costly high-throughput shotgun method.
 As the paper states, complications often arise when sequencing Neandertal DNA due to an overwhelming number of environmental contaminants and the age and fragility of the ancient DNA. They managed to sequence the entire mtDNA genome from one Croatian Neandertal and used it is a template for further sequencing procedures. However, this first genome required thousands of runs using the high throughput whole shotgun sequencing method and was too costly to implement for all following sequences
For the purposes of their study, the researchers developed a new method for extracting specific DNA sequences from a DNA library called primer-extension capture (PEC). It combines the specificity of high-throughput sequencing primers and the benefits of library sequencing. 5’-biotinylated oligonucleotide primers are added to a 454 library. (Biotinylated means that biotin (also known as vitamin H) is attached to the primers for later identification. 454-libraries are made by breaking potential target into pieces 500-800 bp long and attaching identification “barcodes” to the ends of the DNA fragments, shown in Figure one as “A” and “B”.)  They bind to a target sequence and a single round of taq-polymerase mediated extension is allowed to occur. This extension creates a bond between the primer and the target sequence. Excess DNA is then removed by spin column and the biotinylated primer (with target sequence attached) is captured by streptavidin-coated magnetic beads. The beads are then washed at a temperature above the melting temperature of the primers in order to assure that there is a strong association between the primers and the extended sequence. Streptavidin is known to have a strong affinity for biotin and the biotinylated primers remain firmly attached to the beads despite the washing process. The A and B adapters added to the ends of the sequences can then be used to amplify or sequence the fragments that were captured by PEC.
This process is shown in Figure 1. 1i shows the target and background DNA being mixed with the biotinylated primer.. 1ii shows the primer  binding to the target sequence and extension of the complimentary sequence. 1iii shows the biotinylated primer binding to a streptavidin-coated magnetic bead and bringing with it the target sequence. 1iv shows the final sequence extracted from the mix of target and non-target DNA.
Of the many sequences extracted and processed by this method, 18.2-40.2% of the sequences were from mtDNA. Researchers were able to determine the entire mtDNA from all five Neandertal specimens. They found that coverage was positively correlated with GC content and “bases differing from the consensus were dominated by G/C-A/T substitutions”. According to the paper, this is likely due to the deamination of cytosine over time, causing their sequencing methods to partner an adenine with a deaminated cytosine (uracil). The more accurate sections that best fit with the consensus sequence would have a higher GC content and a stronger association with the primers in order to produce better coverage.
According to the paper, their alignment of the five finished Neandertal mtDNA genomes showed that their sequence results were essentially reliable. All but 43 nucleotide positions showed high support with the consensus sequence with an average of 98.5%. These inaccuracies in these 43 positions were seemingly consistent with changes made by the deamination of a cytosine. If there had been any major sequencing errors or other base damage errors, researchers would have expected there to be much weaker support and many more sequence output differences than they found. As a final test of their sequencing and extraction methods, they compared PEC-obtained mtDNA from each individual with its own PCR-amplified HVR sequences to ensure identical results. In 4 out of the 5 cases PCR and PEC-derived sequences were identical. The only non-identical sequence was likely due to previously-noted PCR error and, according to the authors, sheds no unfavorable light in the PEC process.
Upon comparing the six Neandertal sequences, they found 55 supported variable positions out of 16,565 aligned nucleotides. Compared to modern human mtDNA this shows a significant comparative lack of genetic diversity. Specifically, Neandertals appear to have only 30% of the genetic diversity shown in human mtDNA. Since the temporal distance between specimens may have played a part in determining genetic diversity, researchers ran simulation to determine the differences they may have seen if the specimens had all come from the same time period. According to their simulations, they could have overestimated Neandertal genetic diversity by about 20%. If they have correctly anticipated the alterations in diversity attributable only to time, then Neandertal genetic diversity is significantly lower than their findings originally predicted.
As shown in Figure 2, this study created a phylogenetic tree to show the relatedness of the six Neandertal mtDNA sequences a long with human and ape mtDNA. They estimate the Most recent common ancestor (MRCA) for the Neandertals to be around 109,800 years BP. Five of the Neandertal sequences showed high similarity with only one showing a difference significant enough to place it on a separate branch. The outlier sequence was from the oldest time period and the easternmost location. The fact that—for the most part—Neandertal groups in such different geographic and temporal locations would have such similar sequences suggests that there was a rather small population size including no more than 3500 females.
Figure 2 shows the geographic position and distribution of the Neandertal DNA used for this study. Beneath the map is a phylogenetic tree of Neandertals and their close relatives based on mtDNA. It shows the clearly closely-connected relationship between humans and Neandertals and the comparatively large distance between chimpanzees and bonobos and Neandertals. To the far right is an expanded view of the branch that contains humans and Neandertals. A single branch point leads to both Neandertals and humans. There are multiple branch points within both the Neandertal and human sequences showing significant diversity both between and within species. In the human branches, significant differences are found and labeled between Africans and non-Africans.
Table 1 compares lengths of mtDNA in Neandertals and several groups of modern humans. It compares on the basis of (from left to right) number of sequences, number of distinct mtDNA sequences, number of variable sites, mean number of pairwise differences and average percentage of pairwise differences per site. The “European (expanded)” row included extra data published in another paper on European mtDNA sequences.

The researchers concede that, due to the small sampling size, their findings may only be relevant to the more recent Neandertals. Expansion of the human population may have reduced the Neandertal gene pool in later years. In order to further explore this question they looked at protein evolution and found that it was higher in the first sequenced Neandertal mtDNA genome. They measured protein evolution by the ratio of per site amino acid replacement replacement to per site silent substitution (dN/dS). The higher this ratio, the more amino acid changes were cause by mutations compared to the number of silent of non-amino-acid-changing mutations. In larger populations it would be expected that evolution would preferentially ‘select’ those without mutations. However, in a smaller population this kind of “purifying evolution” has less effect. Even though they were working with small amounts of evidence and fractured DNA (both of which they concede to be true) the writers of this paper take the dN/dS to suggest that Neandertals did experience a reduced population size throughout most of their history.

(All information was taken from the original paper with the exception of information on 454-sequencing libraries. I didn’t use any citable information from Roche but their description of the process bettered my understanding of a 454 library. (


Back to Main Page