NEW YORK (GenomeWeb News) - As 454 Life Sciences works to improve its GS 20 DNA sequencer and offer more applications, some of its customers have taken matters into their own hands.
Currently, 454 specifies in its manual that the Genome Sequencer 20, marketed by Roche Applied Science, can sequence at least 20 million bases in a 5-hour run with an average read length of 100 bases. 454 is working on an improved version — scheduled for launch in 2007 — that will provide 100 million bases per run.
454 has also been developing a method for paired-end sequencing, as well as applications for exon sequencing and microRNA sequencing, that are currently “in the hands of early-access collaborators as we prepare … for a general market launch,” 454 CEO Chris McLeod said in July.
But several customers have made — and published — their own improvements, sometimes ahead of the company’s schedule.
At the University of Oklahoma’s Advanced Center for Genome Technology, 454 users have figured out how to obtain up to about 280 bases per read using a beta release of a new version of 454’s software, and recommend a protocol resulting in 150-base reads using the current software. “It’s fairly routine,” said Bruce Roe, the center’s director. His group is one of several testing 454’s upcoming software that correctly calls the bases for longer reads.
To be sure, 454 has presented data at conferences showing that it, too, can get read lengths of 200 bases and beyond, even though it does not formally support this capability yet for the GS 20. The new version of its instrument will provide longer as well as more reads, “the exact mix of which has yet to be finalized,” Bill Spencer, 454’s director of worldwide systems sales, told GenomeWeb News.
In order to up the read length on the GS 20, the Oklahoma scientists increased the number of flow cycles, chilled the apyrase enzyme and other reagents, and made sure there was sufficient computer memory to run the 454 assembly software for large projects.
454 currently sets the machine to run 42 flow cycles, according to the researchers, where the four nucleotides are added in sequence in each cycle. Since the same base often repeats in a given piece of DNA, this translates into an average read length of 100 bases, according to Roe.
By increasing the flows to 63, Roe and his colleagues have achieved read lengths of around 150 bases. “After 63 flows, the 454 software does not do a good job,” he told GenomeWeb News last week. The next version of 454’s software, however — currently in beta-testing — “fixes that problem and can do upwards to 126 flows,” he said. According to 454’s Spencer, the company has not set a date yet for releasing this software more widely.
Using 126 flows, Roe and his team have been able to obtain high-quality reads of 280 bases, he said.
454 “told us that we could not reuse these pico plates, and we figured out how to do that.”
But this represents the current limit, he said, both due the limited size of the reagent bottles inside the instrument and the new software. To remedy this, his lab is considering modifying the reagent holder or moving the reagents outside the machine.
In addition to its protocols for increasing read length, Roe’s center has published several other improvements to 454’s technology on its website, including one for recycling the PicoTiterPlates — the disposable plates on which the sequencing reactions take place that cost up to $1,000 a pop.
“They told us that we could not reuse these pico plates, and we figured out how to do that,” Roe said.
Other 454 users have developed new applications for the GS 20.
Yijun Ruan, for example, a senior group leader at the Genome Institute of Singapore, has developed a paired-end sequencing strategy for the GS 20, a capability that many 454 users have craved. He just published his strategy in July online in Nucleic Acids Research.
Ruan and his group created paired-end ditags, shrunken versions of large DNA fragments in which short stretches from the 5’ and 3’ ends of a cDNA are linked. In the past, he had analyzed these constructs, called PETs, by conventional Sanger capillary sequencing. By linking two PETs — each approximately 40 bases in size — to so-called diPETs, he was able to obtain two PET sequences from each 100-base read on the 454 instrument.
As a result, the team obtained 100 times more PETs from 454’s instrument than from standard Sanger sequencing in the same time. They also estimated at least a tenfold cost-saving and time savings of at least several weeks for one experiment.
About 90 percent of the PET sequences can be mapped specifically to a reference genome, according to Ruan.
Although Ruan collaborated with researchers at 454 for his NAS publication, “the idea of sequencing paired end ditags and the construction of the diPET template is entirely ours,” he told GenomeWeb News, adding that he has filed a patent application for the method.
Right now, Ruan’s method for generating the ditags is based on cloning, he said, but he has been working on a colony-free based method for DNA fragments ranging from five to 10 kilobases and more. “This kind of application can be very useful for genome sequence assembly,” he said.
A detailed protocol of his paired-end strategy will be published in Current Protocols in Molecular Biology soon, Ruan told GenomeWeb News. In the meantime, he is providing the protocols to other scientists upon request.
He has also “been working with Roche to diffuse this technology to 454 users,” he said.
Spencer told GenomeWeb News this week that 454 currently offers paired-end sequencing through its sequencing center and through Roche, using a paired end library that requires no cloning, “similar to that which we currently use for genomic DNA.”
What does 454 think of its customers pushing the capabilities of its technology prior to official upgrades?
“Roche and 454 have always supported an open platform for development,” Spencer said. However, he noted that the companies cannot formally support these improvements immediately.
Bruce Roe, Spencer said, “understands that what we support is specifically stated in the user manuals. I think it’s very interesting what he is doing, but we cannot provide support for anything other than what we have validated,” he said.
According to Roe, the company has been very supportive.
“These companies are not going to make money by selling the machines,” he noted. “They make money by selling the reagents.” In order to get longer reads, he has to use more than one reagent kit per experiment, he said. And although he saves some money by recycling the plates, “we have to buy twice the amount of reagents from them for each run,” he added.
“They have a very positive and receptive attitude, which is quite refreshing” when compared to other companies selling DNA sequencers, he told GenomeWeb News in an e-mail message.
Julia Karow covers the next-generation genome-sequencing market for GenomeWeb News. E-mail her at [email protected]