amplify DNA in a biological sample. They tag the individual units of DNA with fluorescent markers. And they use lots of chemical reagents and sophisticated cameras to read the flow of those fluorescent tags to record the sequence of each DNA unit of A, C, G & T.
The PacBio machine is different in that it “eavesdrops” on a single DNA polymerase enzyme to watch how sequences are synthesized in cells in real-time. This doesn’t require the expensive chemical reagents that the other tools use. The enzymes pass through tiny holes in consumable chips. The PacBio machine uses a sophisticated camera to watch the DNA go through the process in real-time, and software to capture and help researchers get their data in an analyzable form.
The PacBio instrument, in its current form, has its clear advantages and disadvantages. The instrument itself is high-priced at $700,000, but it can run simple experiments very fast and very cheaply—a basic question can be answered for as little as $99 and in as little as 30 minutes, Martin says. That compares with $3,000 and sometimes a week or more to get basic answers from rival machines, he says. The PacBio instrument can also read much longer stretches of DNA than the other tools, which means less duplication and double-checking of work needs to be done to yield an accurate sequence.
While the PacBio machine can read long stretches of DNA, it can’t do a huge volume of them simultaneously. The first beta version used chips that only had 30,000 holes in them, which is sort of like running DNA samples through a narrow funnel. The chips have recently been upgraded to 150,000 holes, which has improved the throughput, Martin says. Still, the PacBio machine’s strengths are in sequencing short stretches of DNA very fast, like the genomic strains of cholera from Haiti, or certain stretches of DNA in agricultural products, Martin says. Over time, chips with many more holes will go into the PacBio machine, making it a much higher throughput instrument, and enable high-bandwidth applications like whole human genome sequencing, he says.
It would be an understatement to say a lot is riding on how well the market responds to this new approach to sequencing. Although PacBio hasn’t provided a forecast of how many machines it plans to ship this year, analysts expect PacBio to deliver 50 to 55 instruments to customers through the end of 2011. Manufacturing and delivering all those machines is going to be a challenge, Martin says.
The company has had its ups and downs with its beta sites, learning a lot of things about how its machine needs to really function in the real world, Martin says. There were a number of software bugs that had to be fixed before the beta customers could be converted into full-fledged paying customers, Martin says.
“It’s a complicated system,” Martin says. “The ability to build a box capable of monitoring in real-time 150,000 little enzymes that are 15 nanometers on a side isn’t easy. There is an optical system that’s complex and sensitive. There’s fluidics. There’s a robot. A vision system to make sure we know exactly where we are. Getting it all to work and work reliably has been challenging. And there are 3 million lines of [software] code that controls it and makes it run.”
Daniel MacArthur, a U.K.-based geneticist and influential genomics blogger familiar with the new instrument, offered up a mixed review: “The PacBio technology in its current form is supplementary rather than revolutionary,” MacArthur says in an e-mail. “It will be extremely useful for certain niche applications, such as assembling genomes from previously unsequenced organisms. It will also complement existing high-yield sequencers like Illumina’s HiSeq2000 or Life Technologies’ SOLiD in other areas, such as quality control testing and improving the long-range assembly of human genome sequences. However, in its current form the PacBio machine will run alongside these instruments, rather than directly competing with them, and unless there are some absolutely staggering improvements in the technology I can’t see that changing in the near future.”
The Web will surely be buzzing as time goes on, and more researchers get their hands on the PacBio machine to see what it can and can’t do very well. It’s a high-wire act for the company, which remained in stealth mode for its first four years, started building buzz the last three, and is now under pressure to deliver the goods to Wall Street and the scientific community. Martin was clearly well aware of it all—and how perceptions will be altered if PacBio slips even a little on what it has promised.
“This is the biggest year of all seven we’ve had, by far and away,” he says.
