the kind of pathogen they were up against in a matter of hours. Mark Stevenson, Life Technologies’ president, said on a recent investor conference call that scientists love the machine’s speed, because speedily obtained raw data can pave the way to fast publishing in top peer-reviewed journals. And, he added, Life considers itself fortunate to be selling a $50,000 machine in a period of budget austerity, at least compared to, say, a $500,000 machine.
This being a brand new way of doing science, skeptics are voicing all kinds of hesitation and criticism about the new machine, which creates risks in the commercial rollout. Scientists’ ability to generate raw sequence data far exceeds their current ability to make sense of all the data on their computers. And scientists have raised questions about the accuracy of the early genomes produced from the new instrument, and pointed out that while Ion Torrent is fast at reading short genomes like those of bacteria, it isn’t yet able to compete in one area of intense interest—the sequencing of whole human genomes. (Harvard’s George Church told the New York Times in July that it probably cost $2 million to sequence Intel co-founder Gordon Moore’s genome on the Ion Torrent machine, while Illumina and Complete Genomics can do that kind of sequencing job for less than $5,000.)
Daniel MacArthur, a genomics researcher and prominent science blogger for Wired, says he is impressed that Life Technologies has been willing to promote the new Ion Torrent machines over its workhorse legacy instruments known as SOLiD. But as someone with experience using various sequencing machines, he’s nowhere ready to declare Ion Torrent the best thing since sliced bread. Here’s what he said in an e-mail yesterday:
“I’m typically skeptical that a company the size of LIFE can successfully take disruptive technology to fruition, but I admit there are some promising signs here: LIFE has clearly been willing to set aside (at least to some extent) its ailing SOLiD platform in favour of the new technology, and the public competitions it’s been holding are an intriguing way of harvesting innovation from its customer community.
Still, I’m not yet convinced that Ion Torrent is sufficiently different from the competition to be a truly disruptive innovation. Right now my money’s on the real disruptive technology being one of the third-gen single-molecule platforms currently in development, such as Oxford Nanopore. Something like PacBio, but with a lower error rate, higher yield and cheaper instrumentation/reagent costs, would really change the field.”
Rothberg has heard all the critiques before, and he insists that his team is all over it—improving accuracy, throughput, and ease of sample preparation steps to feed his instrument. And he says all these improvements are happening while maintaining profit margins. The genomic research community, he says, is in the midst of figuring out the new kinds of experiments the machine can enable. It’s a lot like how hackers and open-source programmers have figured out new applications for computers that the original inventors never imagined. “There will be killer apps,” Rothberg says.
No one really knows yet what all those killer apps will be, but Rothberg says he can foresee the machine being used for spotting differences between healthy tissue and tumor samples from an individual with cancer. There’s long-range potential for screening of newborns for all kinds of genetic abnormalities and future predispositions for disease. And while the FDA may want to weigh in on this question in the future, an automated machine that’s cheap, requires minimal skill to run, and can provide basic answers in a couple of hours also sounds like something that could work as a diagnostic tool in a doctor’s office.
While the Ion Torrent is officially a research-only tool, some users apparently have diagnostic applications in mind. “We are winning diagnostic accounts,” Rothberg says. “We lose if someone wants complete human genome sequencing at low cost. If that’s what they want, they should go to Complete Genomics. But if you have an outbreak and want to know what it is in 2 hours? You use ours.” He adds that most diagnostic applications don’t require the full genome to be sequenced, but really need to focus on a particular region.
All of this work is still in the very earliest stages, which Rothberg says is what excites him so much. He used a computer analogy, saying sequencing technology is where computing was when the Apple II series came out. And while he admits Ion Torrent isn’t yet in position to take away anybody’s market share lead just yet, and while he says he respects his competitors, he isn’t shy about predicting victory.
“Nobody has ever competed with semiconductors and won. Semiconductors are the winning hand right now,” Rothberg says. “Nobody wants to be on the wrong side of history.”
