Oct. 10, 2007 | CAMBRIDGE — Four years ago, biotech entrepreneur and venture capitalist Stanley Lapidus was browsing an issue of the Proceedings of the National Academy of Sciences when a paper by Stephen Quake caught his eye. “It took my breath away!” Lapidus recalls enthusiastically as we sit in a conference room at Helicos’ headquarters in Kendall Square, Cambridge.
“Viewed through one facet of the prism, he’d done very little. He’d gotten one strand [of DNA] to go to 6 bases. But it was addition of single bases to a single molecule. It didn’t escape my notice that this was worth pursuing.”
Quake’s reaction when Lapidus and Flagship Ventures CEO Noubar Afeyan suggested they form a company? “Steve thought we were nuts!” says Lapidus. “He’d started Fluidigm, and Fluidigm technology was farther along than the technology that became Helicos.” But in July 2003, Lapidus, Afeyan, and Eric Lander flew to Los Angeles to meet Quake and sketch out the first patents. By the end of 2003, Helicos had raised $27 million and was ready to begin operations.
An engineer by training, Lapidus’ business background is the world of diagnostics. He started a pair of companies, Cytyc Corp. and Exact Sciences. The idea at Exact was to extract DNA from stool samples and look for mutations associated with colon cancer. But such indirect methods made little sense to him. “One does indirect science because one can’t do the direct experiment... I kept wondering, Why aren’t [oncologists] just sequencing 1,000 tumors? Then I learned the idea of doing very high-throughput sequencing ... what [COO] Steve [Lombardi] calls the DNA microscope.”
Lapidus speaks justifiably proudly of his managerial team. COO Steve Lombardi worked at Applied Biosystems (AB), building its sequencing business to $800 million annually, and was formerly at Affymetrix. Leading the R&D effort is Bill Efcavitch, credited with building much of AB’s product line, including all its sequencers except for the 3730. The advisory board includes luminaries such as Lee Hood, Gene Myers, and John Quackenbush.
In June, Helicos raised $46 million in its IPO — below target, but “we raised the money we needed,” says Lapidus. Indeed, when viewed in the context of this being the first life science tool company to go public in some six years (Lapidus cites Third Wave in 2001), it’s no small achievement.
CFO Louise Mawhinney, who was formerly at ArQule, adds: “We hadn’t published any papers, [investors] had to take it on faith we could actually sequence single molecules. We didn’t have a single order or a commercial product. It’s amazing and a great tribute to the VCs and to Stan that the market would even listen to us.”
Within weeks of closing its IPO, Helicos had two prototype HeliScope instruments running, edging toward the 90 million bases/hour benchmark slated for commercial release at the end of 2007. The HeliScope is a single-molecule imaging instrument that measures the fluorescence of single dye molecules. The main components are an imaging head with four lasers; a cooled CCD camera optimized for low image acquisition time; and an XY stage that moves swiftly at 15 frames/second. Because image stability is critical, the machines are housed in a 500-pound slab of granite shipped from New Hampshire.
The prototype HeliScope differs from Quake’s lab-made prototype in several key respects. Quake’s 2003 paper used fluorescence energy transfer (FRET), but Helicos found that FRET was limited to short read lengths and added extra cost and complexity to the instrument. Breakthroughs in surface chemistry obviated the need for FRET.
That was “a big deal for us,” says Lapidus. “We had to develop surfaces that rinse well and without using ring structures in the solvents. The experiments take days, so that stability throughout a multiday period is not easy to do.” The flow cells have now been upgraded to minimize the problems caused by background fluorescence (see “A Frightening Computational Problem”).
Another key development was engineering new “virtual terminators,” thanks to Helicos’ 7-person organic chemistry team. The virtual terminators help solve the issue of repeat sequences (homopolymers). One way to read them (as Roche/ 454 does) is to measure intensity differences. “The problem with those is you can tell 1 from 2, but 3 from 2 is only 50% as bright, 15 from 14 is very difficult. With single molecules it’s even worse, because of quenching,” Lapidus says.
The virtual terminators interact with DNA polymerase to slow down subsequent base incorporation. “You don’t wait for the reaction to run to completion,” he says. “You wait a minute or two and wash it out. The chance of a second base being added with the first base there is very, very low.”
Lombardi fleshes out the details: “We look at 600,000 growing DNA strands/second [40,000 growing strands/image, moving 15x/second], 90 million bases/hour. There are two flow cells — one does chemistry, one does imaging. When [the camera] takes an image, we’re at a density of 1 strand/square micron, or 100 million/cm2. Three billion strands in both flow cells.” (See “Clusters and Coordinates.”)
The early access program was essentially a collaboration with members of Lee Hood’s group (Hood is a Helicos SAB member). The next phase, Lombardi explains, is to select “labs that we think best represent the market opportunity, who want to be early adopters. We’ll place the instrument in the labs and confirm what we’ve done on the manufacturing floor.” Those instruments will go through a rigorous validation verification, then have a Field Service Engineer and a Field Application Scientist accompany the instrument once it is installed in the customer’s lab to “confirm the same results in their lab as we got on the factory floor.”
As for cost, Lapidus simply says the instrument will be “in the ballpark of mass specs.”
Helicos will debut its first instrument two years after the first machines from Roche and Solexa (now Illumina). Doesn’t that set Helicos back, given its competitors have already booked orders for 200 instruments? Lombardi turns on the question with the aplomb of a clean-up hitter. From protein sequencers to oligonucleotide synthesizers, DNA sequencers to microarrays, being first to market is no guarantee of ultimate success he says. “This feels like 1997,” says Lombardi. “AB had the 377, [then] up popped the MegaBACE. Oh my god, AB’s dead! Darned if the 3700’s arrived a year later. The MegaBACE was a major advance but it wasn’t right. We think this is the right product for the market.”
“It’s fine we’re not the first in the market,” agrees Lapidus. “We believe we have the best platform and the best trajectory. I didn’t start this company to do the Neanderthal [genome]... It’s really about medicine. It’s about reducing morbidity and reducing mortality, living longer and better lives.” He adds: “It’s inconceivable ten years from now that anything other than single-molecule technology will be used for nucleic acid measurements. Maybe it isn’t Helicos, but I believe it will be.”
As excited as he is about the technology and the business opportunity, Lapidus relishes the bigger picture. “This is my shot — and Helicos’ shot — at being more than a footnote in the history of science,” he says. Recalling the importance and excitement of the era of microbe hunting, Lapidus sees the coming era of next-generation sequencing as “an age whose opportunity to change the quality of life of mankind for the better is just as profound as it was then.”
Subscribe to Bio-IT World magazine.