By Julia Karow
Scientists at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, have used a new DNA capture method that relies on primer extension to sequence and compare the mitochondrial genomes of five Neandertals from across Europe.
The method, called primer extension capture, or PEC, which was published last week in Science, is especially suited for capturing short DNA fragments in order to sequence a few small genomic regions in large numbers of individuals. Originally developed to analyze areas of interest in the Neandertal nuclear genome, it might also be useful for other types of targeted sequencing applications.
The Leipzig team, led by Svante Pääbo, has been sequencing the Neandertal genome since 2006 and published an analysis of its mitochondrial genome last year, based on 454 whole-genome shotgun sequence data. At the Biology of Genomes meeting in May, Pääbo said that his group had generated a draft assembly of the Neandertal nuclear genome sequenced at low depth, using both the 454 and Illumina platforms, and planned to increase the coverage over the next few years (see In Sequence 5/12/2009).
Ultimately, the scientists want to identify genomic regions where humans differ from both Neandertals and chimpanzees, meaning these changes happened fairly late during human evolution, according to Adrian Briggs, a graduate student in Pääbo’s lab and the lead author on the recent Science paper.
In order to double-check the sequences of these regions in the Neandertal genome, and to study them in other Neandertal individuals, the researchers needed a targeted resequencing method, since the cost of shotgun sequencing entire Neandertal genomes is prohibitive, he said.
Direct PCR from a DNA extract is not suitable for ancient DNA, according to Briggs, because it is highly fragmented, and many of the pieces are too small for PCR amplification, which requires two primers to anneal to the ends of the desired region.
[ pagebreak ]
At the time he developed PEC — more than a year ago — the only other DNA capture methods he was aware of were hybridization-based. Many of the ancient DNA fragments, though, are shorter than the probes used in these hybridization approaches, he said, so those methods seemed not ideally matched. In addition, DNA extracted from a fossil bone is heavily contaminated with microbial DNA, and less than 1 percent usually comes from Neandertal. "It's a massively complex metagenomic library to select DNA fragments from, so I wanted something that is rather specific, and I didn't know whether hybridization was specific enough," he said.
Brigg's method combines the high specificity of PCR primers with the strong association between a long probe and its target. It uses short 5'-biotinylated primers that bind specifically to targets in a pre-amplified sequencing library, where each fragment carries adaptors for next-generation sequencing.
The primers produce "quite a high specificity of the reaction," he said, but do not create a strong enough bond with the target to be able to wash away unwanted DNA. Thus, in a second step, the researchers extend the primer using DNA polymerase, thus increasing the length of the double-stranded region between probe and target.
The concept of using primer extension to capture DNA of interest for analysis is not entirely new – researchers at Generation Biotech, a New Jersey-based firm, and the Children's Hospital of Philadelphia published a related method last year. But the Leipzig researchers used the approach for the first time to capture DNA from second-generation sequencing libraries.
In order to test their PEC method, they employed it to capture and sequence the mitochondrial genomes of five Neandertal individuals from four sites, located in Croatia, the Neander Valley in Germany, Spain, and Russia.
Covering the 16.5-kilobase mitochondrial genome required almost 600 capture primers and four separate reactions with about 150 primers each. In order to increase the enrichment, the researchers applied the capture procedure twice, leading to a maximum enrichment of about 80,000 fold. Up to 40 percent of the sequence reads — generated on the 454 platform — derived from mitochondrial DNA.
Briggs believes that the enrichment factor could be higher with smaller target sizes and fewer primers. "As you increase the number of primers, you increase the amount of background at the end," he said.
The researchers used the 454 platform to sequence the mitochondrial genomes, but have since also combined PEC with sequencing on Illumina's Genome Analyzer. PEC is also compatible with other second-generation sequencing platforms that generate DNA libraries with adaptors, according to Briggs.
[ pagebreak ]
Briggs reckons that the PEC method, which is not patented, is best suited for projects that aim to analyze a few relatively small regions in many individuals in a short period of time. In contrast to hybridization-based methods, where the hybridization step can take several days, he said, primer extension takes only minutes, and the entire method about half a day.
However, like PCR, the PEC method is limited by the number of primers that can be multiplexed in a single reaction. "For applications like capturing [all] exon regions, or hundreds of kilobases, this is probably not the most useful method," he said.
In fact, the Pääbo group is now also using Agilent capture microarrays as a complementary method for isolating specific regions of Neandertal DNA for sequencing. Although array capture experiments take one to two weeks, according to Briggs, "we find this [method] very useful for covering many regions." For example, the researchers have used arrays to capture 14,000 amino acid positions in the Neandertal genome. "If you have a big project, like a 500-kilobase genomic segment or a few thousand different exons you want to sequence, it's probably the best method for ancient DNA," he said.
Other groups are already using the PEC method to study ancient DNA, according to Briggs, but applications for the method might be more general. "I can envisage it being used for things where you know there is a target sequence somewhere, but you don't know what's flanking this region," he said.
This could be, for example, a repetitive element, such as a transposon, that has inserted itself in multiple sites of the genome. A primer could target the end of this element and be extended into the insertion region, which could then be identified by sequencing.
Another potential application could be the taxonomic identification of species in metagenomic samples. This requires a conserved region — for example, 16S RNA — which is currently usually amplified by PCR. According to Briggs, it would be "much easier" to use primer extension to extract and sequence the conserved sequence.
The Leipzig researchers are employing both primer extension capture and Agilent arrays "to confirm interesting things that we find in the big 1-fold shotgun scan," Briggs said.
They also want to use the methods to confirm results from the five mitochondrial genomes. Their analysis of the diversity between these genomes suggested that the total population of Neandertals was probably small, but "to be really sure of the story, we need to do this for many loci across the nuclear genome," he said.
In addition, he and his colleagues plan to use the methods for other ancient DNA projects, such as identifying mammalian species from bones found in Neandertal caves.
Most of these analyses will involve Illumina's Genome Analyzer, of which the MPI in Leipzig currently has five in house. With a read length of 100 bases – anticipated to be available later this year – and a larger number of reads per run, the platform will be better suited for sequencing fragmented ancient DNA than 454's, which has been focusing on longer reads. "For ancient DNA, the perfect match that we had originally [with 454] diverged a bit, and now we have a more perfect match to the Illumina platform," Briggs said.