SAN DIEGO, Jan. 14 (GenomeWeb news) - A workshop at the Plant and Animal Genome conference held here this week discussing microbial genome sequencing concluded that 454 Life Science's Genome Sequencer 20 is well-suited for resequencing and de novo sequencing microbial genomes.
But the session also highlighted some rough edges of the technology 454 must still iron out.
In his introductory talk, Stephen Kingsmore of the
Kingsmore, who predicted that the new breed of sequencers will "transform how sequencing will be performed over the next three to five years," also touched on some rough edges that 454 might want to address. For example, he cited the need to refine base-calling algorithms. "454 is aware of this," Kingsmore said, adding that quality scores need further validation. The sequencer tends to insert a base following a run of identical bases, he said.
Other improvements on 454's radar, according to Kingsmore, include improving the de novo assembly algorithm, currently limited to 50 megabases, and developing a paired end-read process.
Kingsmore also pointed out the bottleneck in moving from raw sequence generation to bioinformatics. "We are moving from having behemoth sequencing centers to needing behemoth bioinformatics centers," he said. Aside from that, Kingsmore said 454's technology is "certainly ready for bacterial genomes."
After Kingsmore's overview, several attendees volunteered their own experiences with 454's instrument.
George Weinstock of Baylor College of Medicine has used 454's sequencer on genomes ranging from bacteria to insects. For the sequencing and assembly of bacteria, Weinstock said the overall quality of 454's technology, not counting homopolymer regions, does a "good-enough job." The instrument is particularly useful for finishing or adding data to already-completed sequences, he said.
But Weinstock, who has been evaluating the machine's performance via high usage, has found it to be "very uneven" in terms of producing consistent read lengths. He estimated that, on average, the device delivers reads between 90 and 105 base pairs. Weinstock said the sequencer appears to be "set at the factory to only do enough runs to get 100 base pairs," and his attempts to tinker with this capacity have not been successful.
Kingsmore, who has worked with 454's service business in the past, noted that 454 does not intend to extend its instrument's read length to 200 base pairs until late next year.
Stephan Schuster, associate professor at
Schuster has also departed from using 454's assembly application, and instead relies on software he developed for the job.