By Julia Karow
COLD SPRING HARBOR, NY — Researchers have long been looking for ways to obtain local sequence at the same time as long-range structural information about DNA. Life Technologies is now developing a method for its quantum dot-based single-molecule sequencing technology that promises to deliver both.
At the Personal Genomes conference that ended here on Sunday, Joe Beechem, head of single-molecule sequencing in the company's genetic-systems division, provided a first glimpse of how the platform, which Life Tech has not yet commercialized, could be used in a "top-down" mode to generate multiple sequence reads with defined locations along a single, horizontally aligned DNA molecule up to 500 kilobases in length.
Beechem first presented Life Tech's real-time single-molecule sequencing technology, dubbed "Starlight," earlier this year at the Advances in Genome Biology and Technology meeting (IS 3/2/2010). The system uses a single quantum dot nanocrystal attached to a DNA polymerase as a photon source for four-color terminally labeled nucleotides and measures light emitted from both the Qdot and the labels in real time as bases get incorporated by the polymerase.
Primed DNA templates tethered to a slide are illuminated with a 405-nanometer laser, and as polymerases synthesize DNA at a rate of between two and 10 nucleotides per second, signals are recorded inside the evanescent field of a total internal reflection fluorescence microscope.
Life Tech's technology faces the same challenge as Pacific Biosciences' in that the laser inactivates the polymerase after a certain time — in Life Tech's case after synthesizing about a kilobase of DNA. But unlike PacBio, Life Tech does not immobilize the polymerase, so it can wash off "dead" enzyme and replace it with a new batch of polymerase that picks up where the old one left off.
Users can determine, Beechem said, how many times they want to exchange the enzyme in what the company calls "relay sequencing," each exchange taking between 5 and 10 minutes. For up to five exchange cycles, he said, the quality of the signal does not change significantly.
In another sequencing mode, called "recursive sequencing," users can strip the newly synthesized DNA off the template and sequence the same molecule five to seven times over, increasing the consensus accuracy, similar to PacBio's circular consensus sequencing application. Five cycles of sequencing, Beechem showed, can increase the accuracy to 99.9 percent if the raw error rate is 5 percent.
What limits the maximum read length in "relay sequencing," Beechem explained, is that the signal drops when the polymerase on the template leaves the evanescent field of the TIRF microscope, typically after about 5,000 bases.
In order to keep long DNA molecules inside the illumination zone, Life Tech researchers have now come up with a way to roll the DNA templates into a horizontal position. This also allows them to record signals from several DNA polymerases on the same DNA template simultaneously.
Beechem, who said he was not at liberty to discuss how the DNA is made to lie down on the slide, showed that this "top-down" sequencing approach has allowed him and his colleagues to analyze DNA as long as 250 kilobases with about 30 polymerases generating sequence information at the same time. The result is a large number of paired-end reads on the same DNA molecule.
"You can get thousands of long molecules lined up and sequence them all," he said, adding later that it is possible to "pile DNA molecules very close to each other" to achieve high throughput.
It is also possible, he said, to direct where on the horizontal DNA template polymerase enzymes latch on and start sequencing. Based on this "fingerprint" of polymerase start points, which can be recorded within the first 5 to 10 minutes of a run, sequence reads can thus be placed directly onto the chromosome, in contrast to the "bottom-up" approach of other sequencing methods, where sequence reads are only placed in context during the assembly.
Beechem likened the "top-down" sequencing approach — which he said provides near real-time mapping of data and structural variation information onto chromosomes — to Google maps, where users start with a high-level view and then zoom in to determine their exact location.
He stressed, however, that the method is currently not close to becoming a product.
While the size limit is currently about 500 kilobases of DNA, the ultimate goal, he said, is to analyze entire chromosomes this way. Researchers could, for example, study DNA from a tumor and its normal control at the same time, or focus in on specific regions of a chromosome.
The idea of analyzing horizontal DNA is not new, of course, and Beechem credited David Schwartz at the University of Wisconsin, Madison, as the pioneer of optical mapping, a technique that has been commercialized by OpGen, where single DNA molecules are loaded into microfluidic channels and then analyzed by restriction enzymes to create an optical map. He said he has discussed the research with Schwartz but the company is not formally collaborating with him on this method.
Schwartz has an ongoing grant from the National Human Genome Research Institute to develop a "Nanocoding" platform to discover structural variation in the human genome. According to the grant abstract, the platform will have "nanoconfinement capabilities that closely parallel those found in sub-100 nanometer devices requiring traditional lithographic approaches" to trap single molecules of DNA that are fluorescently labeled at specific sites.
Also, BioNanomatrix has developed a nanochannel chip platform, called NanoAnalyzer, to analyze long stretches of DNA for genome assembly, structural variation mapping, and other applications. In collaboration with Complete Genomics, it is also working on a sequencing method for its platform under a grant from the National Institute of Standards and Technology (IS 10/20/2009).
In addition, researchers at the University of Oxford led by Kalim Mir have been exploring a method that uses single-molecule microarrays to capture long fragments of DNA and stretch them out using fluidics prior to sequencing (IS 3/27/2007) and have filed several patent applications on this approach.