By Julia Karow
This story was originally published Dec. 16.
As large-scale sequencing projects involving hundreds of samples are getting underway, a transposase-based library prep method sold by Epicentre Biotechnologies that promises to be faster and simpler and use less starting material than standard methods is beginning to gain traction.
Earlier this month, three early adopters from the University of Washington, the Marine Biological Laboratory in Woods Hole, and BGI, in collaboration with the company, described their use of the method in Genome Biology, including a comparison with standard library prep protocols, sub-nanogram library construction, exome capture from nanograms of DNA, PCR-free libraries, colony PCR libraries, and 96-plex barcoding.
"This has the potential to completely change how people approach library construction," said Jay Shendure, assistant professor of genome sciences at the University of Washington and the senior author of the paper. His lab has completely switched over to the transposase method — sold by Epicentre under the name "Nextera" for use with the Illumina and 454 sequencing platforms — to prepare libraries for all their sequencing applications except production-scale exome sequencing, for which they plan to use it in the future, too.
At the heart of the method is a hyperactive version of the Tn5 transposase, which fragments DNA and adds adaptors at the same time. This is followed by an optional PCR step to add platform-specific primers and, if desired, barcodes, though it is possible to skip this step and introduce Illumina primers as part of the adaptors. The method usually requires nanograms of input DNA, although the researchers have gone down to as little as 10 picograms.
"If you take out the PCR, you can start from your genomic DNA and have a library that's ready to go on the sequencer in 30 minutes, easily. Or a few hours with the PCR," said Andrew Adey, a graduate student in Shendure's lab and one of the lead authors of the paper. The lab routinely makes 96 libraries in parallel now "with zero automation," Shendure said.
By comparison, standard methods for library construction require many more steps — including DNA fragmentation, end-polishing, ligation of adaptors, gel-based size selection, and PCR amplification — that take six to 10 hours and need liquid handling robotics to automate, according to the authors.
In their paper, the scientists compared the transposase method with two standard library construction methods, one using mechanical fragmentation of DNA, the other endonuclease-based fragmentation. Using the three methods, they generated human and bacterial libraries and sequenced them on the Illumina GAIIx. They also made bacteriophage libraries, which they sequenced on the Roche 454 Genome Sequencer FLX.
The transposase method, they found, is slightly more biased than the others with regard to where the genome gets fragmented, but the overall coverage distribution across the sequenced genome was "almost identical across the methods," said Adey. However, the method's bias was greater for sequencing PCR amplicons, compared to genome sequencing, Shendure said.
The authors noted in the paper that the bias of the transposase method was "offset by large advantages with respect to speed, simplicity, and low input requirements."
One downside of the method is that the distribution of insert sizes is wider than with some shearing methods, but the researchers have also coupled it with a microfluidic chip-based automated size selection platform from Caliper to get more defined insert sizes.
One problem the method does not solve is how to make long-range mate pair libraries more easily. "That's an ongoing challenge in the field broadly," Shendure said.
[ pagebreak ]
The method's low input DNA requirements will be especially useful for human resequencing projects, he said, where the amount of sample is limited, for example tumor samples or buccal swabs. In the paper, he and his colleagues showed that they could capture and sequence exome DNA from libraries made from 50 nanograms of genomic DNA, using the Nimblegen SeqCap EZ exome method and the Illumina GAIIx sequencer. They achieved similar percentages of on-target reads and similar uniformity but lower specificity than with a standard library.
To test how little DNA they could get away with, they also made bacterial libraries from 500 picograms and 100 picograms of DNA and a human library from 10 picograms of DNA and sequenced them, and only found reduced complexity for the latter.
Shendure said he guesses the method — relying on an enzyme — will be more robust to variations in DNA quality and degradation than standard methods, although they have not yet tested this. "In cancer, this is a huge issue," he said.
In another extension of the method, the scientists used it to make genomic libraries directly from bacterial colonies, which might be of interest to groups wanting to sequence many bacterial genomes.
In general, the method might be less relevant for large genome centers that have already automated the sample-prep process and have robotics equipment in place.
BGI, for example, currently still uses mainly the standard library prep method, according to Asan, a BGI researcher and one of the lead authors of the paper, because most of its collaborators require that method to keep their data consistent.
But for users wanting to scale up their library production, or for new users of next-gen sequencing, "this is a lot simpler than the standard way of doing it," Shendure said.
Epicentre, which developed the transposase library prep method in house, introduced its first Nextera kits about a year ago, for use with the 454 and Illumina sequencing platforms.
Nick Caruccio, Epicentre's director of market development, said the company has seen the most uptake from "new places that are just now addressing the sample prep bottleneck," rather than established genome centers.
Right now, the company has no plans to make Nextera compatible with the SOLiD platform. "The SOLiD chemistry makes it tricky to adapt Nextera," said Caruccio. "It's much better suited for sequencing-by-synthesis chemistries."
However, the firm does plan to develop kits for new sequencing platforms entering the market, such as the recently launched Ion Torrent PGM, as well as the Pacific Biosciences sequencer.
During the first quarter of next year, Epicentre plans to introduce a PCR-free version of the Nextera kit, a low-input version for sub-nanogram amounts of DNA, and a single-tube kit. Sometime during the first half of 2011, it also plans to launch a mate-pair Nextera kit for use with the Illumina platform.
In addition, Epicentre is testing a Methyl-seq Nextera kit for bisulfite sequencing on the Illumina platform, which could easily be adapted for the 454 as well, Caruccio said.
Developing Nextera kits for ChIP-seq and exome sequencing, on the other hand, has been more challenging, and the company has no timeline for introducing these, he added.
Regarding pricing for the Nextera kits, Caruccio said that "the material cost will be higher, but the time savings outweighs that, and the ability to sequence rare samples is just something you cannot do with the standard prep."
"We realize that library prep is going to be a commodity and to be scalable, so we do need to address the cost issue there," he said, "and that's something we will be doing with the single-tube and low-input kits." High-volume users also receive discounts on the list price, and for those users, "it's affordable and comparable to their other methods on a per-library prep basis."
Epicentre currently does not have comarketing agreements with sequencing platform vendors for Nextera but is evaluating those options, Caruccio said.
Have topics you'd like to see covered in In Sequence? Email the editor at jkarow [at] genomeweb [.] com.