Researchers from the Max Planck Institute for Evolutionary Anthropology have published a next-generation sequencing protocol that prepares libraries from single-stranded DNA as opposed to double-stranded DNA, a method that is particularly amenable to sequencing ancient DNA.
The protocol, published last month in Nature Protocols, is well-suited for sequencing ancient DNA because it results in less DNA loss than double-stranded methods, enabling genomes to be sequenced to a higher coverage with limited amounts of poor quality DNA.
The team applied the method to a Denisovan genome last year, increasing genome coverage from two-fold to 30-fold (IS 9/4/2012).
Sequencing ancient DNA has proven challenging because the DNA is often limited, highly degraded, and the fragments are often chemically altered.
Additionally, most next-gen sequencing protocols were designed for experiments starting with sufficient amounts of high-quality DNA, "the exact opposite of ancient DNA," Michael Knapp a professor at the School of Biological Sciences at Bangor University in Wales, who was not affiliated with the study, told In Sequence. He previously led a team at the University of Otago in New Zealand that sequenced mitochondrial genomes from ancient DNA samples of the first groups of Polynesians to settle New Zealand.
In contrast to double-stranded methods, the Max Planck protocol was designed specifically for ancient DNA. While other protocols for sequencing ancient DNA exist, such as a method designed for Roche's 454 platform and another developed by Illumina, those are both still double-stranded methods, Knapp said. "This new protocol is clearly a major step forward."
Preparing a library from single-stranded DNA helps increase the amount of usable DNA. Often, ancient DNA will have breaks in a single strand, explained Marie-Theres Gansuage, lead author of the study. "If we have a single strand break in a double-stranded molecule … the molecule gets lost in the double-stranded method,'' she said.
The method begins by denaturing the DNA to obtain single-stranded DNA. Next, the enzyme CircLigase II is used to add biotinylated arms to the 3' ends, which are then immobilized to streptavidin beads.
The remaining purification reactions take place while the DNA is bound to the beads. The magnetic core of the streptavidin beads is used to collect all the beads in order to change the reaction buffers and mixers.
Keeping the DNA bound to the beads helps avoid DNA loss that would have occurred from protocols that use silica-based column purification or carboxylated bead purification, Gansuage said.
Ancient DNA is degraded, with size distributions around 50 bases or sometimes even less, Gansuage said. "Molecules of this length will get lost with silica purification steps," she said. But, "the biotin system avoids this huge loss."
Once the DNA is bound to the beads, it is amplified and sequencing adapters are added. At the end of the protocol, heat is used to remove the DNA molecules from the beads. The DNA is then copied to make the molecule double-stranded again.
Aside from preventing the loss of short DNA molecules, the protocol also prevents DNA loss that occurs when there are breaks in one strand of the double-stranded molecule. These molecules are entirely lost in double-stranded methods, but in the single-stranded method they are "disassembled into multiple fragments upon heat denaturation, and each fragment has an independent chance of being recovered in the library," the authors wrote.
The method also prevents losses that occur because ancient DNA is often chemically modified, preventing adapter ligation during library prep. By using a single-strand method, the strand opposite to the modified strand can still be recovered.
After the library is prepared, the team uses qPCR to do a quality assessment of the library. At this stage there is between a six-fold and 20-fold improvement in coverage, but following sequencing and filtering, this number decreases a bit.
Nevertheless, when the team applied the method to the Denisovan genome last year, they boosted coverage from around two-fold to 30-fold, she said. Gansuage said the team has since applied the method to other ancient genomes from Neandertals.
The one drawback to the method is its cost, Gansuage said. She estimated that the library preparation for the single-stranded protocol would cost just over twice that for the double-stranded protocol, at around €25 ($33) per sample, compared to €10 ($13) per sample. The streptavidin beads and the CircLigase are responsible for the extra cost, she said.
The protocol is also slightly more time consuming, taking around two days as opposed to one.
However, Knapp said that the added cost of the library prep could be partially offset by the increase in sequencing yield. "If large amounts of sequence data from poor-quality DNA are to be obtained, the increase cost of library preparation may be more than balanced by the increase sequencing efficiency," he said. "More target DNA in the sequencing library means less sequencing runs and less sequencing costs."
Knapp added that his group is now considering employing this protocol for their own ancient DNA sequencing experiments and said that the protocol itself seems relatively straightforward and simple to implement.
The Max-Planck team does not intend to commercialize the protocol. Additionally, Gansuage said that aside from ancient DNA, the protocol may be useful for sequencing DNA from formalin-fixed paraffin-embedded tissue, since DNA from those samples are often degraded and of poor quality.
FFPE samples are "often crosslinked to proteins," she said. "But you can get rid of the crosslinked proteins by degrading it with heat," and creating single-stranded molecules.
Additionally, the protocol could be useful for sequencing oligos, which are single-stranded. Here, sequencing could be used to do quality control of the oligos, she said.