Oh, And One More Thing: A Wowser Moment in DNA Sequencing
[Updated and corrected, 10:15 am PT] Steve Jobs used to love being on stage, getting the audience in the palm of his hand. Famously, he’d say, “Oh, and one more thing,” to build suspense right before delivering some stunning line about a new Apple product. Last week, that same showmanship was at work in a hotel ballroom off the Gulf coast of Florida, where a buttoned-down Brit delivered a message heard around the world of DNA sequencing.
The scene was at the Advances in Genome Biology & Technology (AGBT) conference, held at the remote site of Marco Island, FL, near the Everglades. This little affair of just more than 1,000 people draws the who’s who of the genomics world, and a mad scrum of technology companies making new machines that are constantly getting better/faster/cheaper at the sequencing of DNA.
The big headline from Marco Island this year came on Friday from Clive Brown, the chief technology officer of U.K.-based Oxford Nanopore Technologies. Just as he was getting toward the end of his presentation of what’s new from Oxford, Brown uttered the familiar “Oh, and one more thing” phrase. He then reached into his pocket, and pulled out a device the size of a USB memory stick. This little gizmo, called MinION, was said to cost less than $900, and can deliver 150 megabases of DNA sequences per hour. A larger Oxford system, GridION, was being designed so when configured a certain way, it could sequence an entire human genome in 15 minutes. [Corrects earlier story, which confused MinION and GridION]. Sequencing has been on an incredible innovation run the past five years, but this takes the game to a new level. Only one month ago, sequencing powerhouses Illumina and Life Technologies said they were on track to make instruments that can sequence whole genomes for $1,000 in one day by year’s end.
George Church, the DNA sequencing pioneer at Harvard University, told Forbes’ Matthew Herper that calling the device a game-changer “is an understatement.” Analysts quickly rushed out reports on how this new technology could potentially disrupt the market again.
“It’s definitely stirring conversation,” says Todd Smith, the founder of Geospiza, a DNA analysis company acquired by PerkinElmer last year, who attended the Marco Island conference. “When someone shows a DNA sequencer the size of a USB drive, it kind of makes you want to run out and buy 1,000 computers to scale up production.”
Part of what’s interesting here is the way the technology has fundamentally changed, a bit like when photography went from film to digital. Today’s DNA sequencing market leaders—Illumina and Life Technologies—use common polymerase chain reaction (PCR) techniques to amplify DNA in a biological sample. They tag the individual units of DNA with fluorescent markers. And they use sophisticated cameras to read the flow of those fluorescent tags. The PCR requires laborious sample preparation, fluorescent tags add some cost per individual DNA unit, and the cameras make for expensive capital equipment.
Oxford Nanopore is approaching DNA sequencing in a different way altogether. As I wrote here in a feature story in February 2010, it doesn’t require any PCR amplification of a sample, any fluorescence, or a camera. Instead, the company’s machine runs the sample through very small (one-nanometer wide) pores. As the DNA passes through these nanopores, the Oxford machine records the electrical charge that’s associated with each individual base pair of DNA, like a signature.
The Oxford machine still isn’t ready for prime time, although the company said it plans to introduce a commercial version of the product in the second half of 2012. PerkinElmer’s Smith noted that Oxford’s chief technologist didn’t actually perform a full genome sequence during a 15-minute stretch on stage, which would have really brought the house down. Still, the speed and portability of the new technology will surely get researchers thinking about all kinds of new experiments. Sequencing microbes during environmental field tests, and point-of-care genomic diagnostic tests, are just a couple of ideas that leap to mind right away, Smith says.
Exciting as the possibilities may be, the Oxford machine, like any other one, isn’t perfect. It has an error rate of about 4 percent, which Smith says is high. “You’ll need extra redundancy and informatics” to help correct the errors, he says.
Smith, who has been coming to the Marco Island conference the past five years, said Oxford is picking up on the buzz that Jonathan Rothberg, the founder of Ion Torrent Systems, stirred up two years ago at this conference. Rothberg, one of the most skilled showmen in the genomics business, famously unveiled his semiconductor-based sequencing machine at Marco Island in 2010. Rothberg memorably stated that the Ion Torrent machine, about the size of a desktop printer, could sequence whole genomes “from the back of a donkey.”
Two years later, we’re talking about a new ballgame, based on dropping a sample of blood onto a device you can stick in your pocket, and which will spit out results in 15 minutes. That’s so cool it would have been sci-fi just three or four years ago. But as Smith rightly points out, a company can only get away with talking about the whiz-bang part of the technology for so long. Quickly, the conversation will shift toward “Hey, what the heck are you going to do with that genomic data?”
“You see this all the time,” Smith says. “The more mature platforms talk about the applications with their machines. The new ones talk about the technology, and the vision it creates. Time will tell in terms of what it delivers.”
Besides all the chatter with Oxford Nanopore, I sought some quick commentary from Marco Island on the news from other major sequencing companies. Here’s what Harvard’s George Church had to say via email.
Illumina: Illumina, the market leader in sequencing, is defending itself from a hostile takeover bid by Switzerland-based Roche. Complicating matters a bit further, Illumina is an equity investor in Oxford Nanopore, which could end up competing with its workhorse instruments. Church was clearly wondering how the Oxford Nanopore presentation might change the Illumina/Roche drama. “Will Roche consider a (Oxford) Nanopore company?” Church says. He also wonders, “if Illumina matches the LifeTech “$1000 genome” (presumably consumable costs), what will their service price be?”
Life Tech: Life Technologies made news earlier this year with its Ion Proton Sequencer, and its purported ability to produce $1,000 genomes by years’ end, but if it gets there, it will be neck-and-neck with other companies, Church says. “Proton II in late 2012/ early 2013 seems very far off now,” Church says. “There will be a few companies capable of $1000 genome (reagent costs) by then and long reads too.”
Complete Genomics: Church notes that he’s impressed with the low error rates from the Mountain View, CA-based company. “The long-fragment read (LFR) technology may routinely deliver multi-100kbp (kilobase pairs) haplotypes and less than 1 error in 10M (1E-7). This could be very attractive for clinical applications (note Mayo Clinic deal).
PacBio: The Menlo Park, CA-based company has hung its hat for a while on its ability to sequence long stretches of DNA, which requires less overlapping and assembly than existing technologies from Illumina and Life Tech. But PacBio has got to be watching Oxford closely as an emerging challenger, based on the claims it made at Marco Island. “Longish hyphenated reads won’t be quite as compelling if the Oxford Nanopore Technologies 100 kbp read-length proves to be robust,” Church says.
Oxford Nanopore: The big question, Church says, is “How many reads does it take to get the 4 percent raw error down to 1E-7 consensus? Any systematic (non-random) errors?”
BGI: The organization formerly known as the Beijing Genomics Institute has taken on some ambitious experiments, but doesn’t appear to be developing its own sequencing instrument technology, Church says. “So far, BGI is a consumer of tech, not a producer. As long as companies like Illumina & Complete Genomics are doing both, and as long as automation keeps increasing, it is not clear how big a niche BGI has in human genomics, but getting better at de novo sequencing may be a good fit for them.”