Oxstones Investment Club™Its technology may finally fulfill the dream of gene sequencing
By Ashlee Vance
On Apr. 28, in a factory in Menlo Park, Calif., a few black and white machines were being assembled and prepped to go into shipping crates. The machines looked like fancier-than-usual copier equipment. Each was about half the size of a MINI Cooper and adorned with some design flourishes—oversized, glowing power buttons, a slick touchscreen monitor on the side. Still, to look at them you wouldn’t know they cost $700,000 a pop, that they’re the result of a 14-year, nearly $600 million quest, or that their creators believe these machines may change the scientific understanding of life itself.
These are the first production models of the PacBio RS, gene sequencers made by a startup called Pacific Biosciences, and they’re heading to research laboratories around the U.S., including several national defense labs and the Howard Hughes Medical Institute. Move past its smooth exterior and the RS reveals an interwoven collection of lasers, chemical mixing stations, cameras, robotic arms, and special chips. It’s essentially a superpowerful microscope that records, in real time, biological processes on a molecular scale. That means it can see the creation of the tiniest of things—including, most crucially, DNA—in rapid-fire action. Soon enough, the RS may well do something that’s never been done before: Take an entire strand of human DNA, with its 3 billion bits of information, and map it out in minutes.
The RS has already achieved something of a mythic status in the genetics world, which has had a spectacular run of hype and disappointment ever since a draft of the human genome was first mapped in 2000. The hope then was that the genetic basis for disease would be identified, and cures would pour forth. Reality proved more complicated. The laborious mapping process did identify some telltale gene sequences that cause illnesses. But most maladies like cancer, it turns out, vary as much as their victims—blockbuster drugs work great for people who happen to have a particular genetic code; everyone else needs a treatment tailored to their particular case. Fulfilling the promise of bespoke cures requires a mapping tool that’s accurate, economical, and fast. More than that, it may require a machine that can analyze the incredibly intricate changes that occur not just to DNA but to other mechanisms in the body when disease hits.
Pacific Biosciences claims the RS provides the first lens capable of viewing this type of complexity. “The ability to observe these things in real time has been the goal of modern biology for the last 50 years,” says Eric E. Schadt, the company’s chief scientific officer. “PacBio is the only game in town for that type of work.” Even competitors, who express doubts about the RS’s commercial prospects, allow that the company has created a scientific marvel. Today, it bestows scientists with the power to watch the body’s mechanism for decoding DNA. Tomorrow it could apply similar steps to studying all manner of molecular creatures from RNA to ribosomes, giving pharmaceutical companies, hospitals, and super-food makers fresh insights on the behaviors of viruses and the effects of environmental conditions on organisms over time. “I think it is a tour de force,” says Jonathan M. Rothberg, the chief executive of Ion Torrent Systems, a rival maker of low-cost sequencing systems. “Technically, it’s fantastic.”
PacBio was founded in 2004 by a physicist named Stephen Turner. With his mat of short, curly brown hair, thick-rimmed glasses, and earnest demeanor, he seems most comfortable in front of a white board talking science, making declarations such as, “Here you have E to the minus lambda Z.” Turner is either kind enough or oblivious enough to assume that his audience can keep up.
Turner, 43, grew up near the University of Wisconsin-Madison, where his father taught math and his mother lectured about 13th century Italian literature. His folks hoped he’d follow them into academia, and he did, initially (at Cornell University, where he completed a PhD in physics), but when a graduate student, Jonas Korlach, approached him about inventing a way to eavesdrop on the enzyme polymerase, he became an entrepreneur.
Polymerase is nature’s DNA sequencing machine. As a cell divides, the enzyme travels along strands of DNA, making copies of the cell’s genetic construction. Wouldn’t it be cool, Korlach thought, to get a close-up view of polymerase in action? You’d be able to see how it reads the stream of “bases”—the four types of building blocks strung together in DNA’s helical structure—and makes copies. It’s an efficient process. “Polymerase runs along DNA at 1,000 bases per second,” Korlach says. “That makes it extremely fast, and it can hold on for hundreds of thousands of bases, making extremely few errors.”
Intrigued, Turner set to thinking about how to realize Korlach’s dream. His breakthrough: really small holes. He coated a piece of glass with aluminum, hit it with an electron beam, and made a series of holes small enough to hold one molecule each. The technology gets complicated here, but essentially Turner and Korlach realized that they could fasten a polymerase molecule near the bottom of one of these holes and fill the well around it with a liquid made up of the four bases of DNA, known in scientific shorthand as A, T, G, and C. The researchers could then attach fluorescent markers to each type of base, shine a light into the holes, run the DNA through like thread through the eye of a needle, and pick up the colors as the bases react with the polymerase.
The crude lab technology worked well enough to convince Turner it could be the basis of a new tool for scientific discovery. “I’m like, ‘Guys, We have to patent this,'” he says. “The reaction from everyone was, ‘Hah, hah. Steve wants to patent a hole.'” Turner thought about his colleagues’ reaction and figured he had a marketing problem. Instead of holes, he started calling them zero-mode waveguides. The name change eventually got his lab partners thinking patents, too. “It took about three weeks,” he says.
Turner decided to quit academia to make a business out of his zero-mode waveguides. From about 2001 to 2003 he worked alone, making refinements and trying to raise money. (Korlach kept helping on the side, but didn’t commit full time until years later.) Turner and his wife crisscrossed the country, visiting venture capitalists in Boston, New York, Baltimore, and Silicon Valley. “It doesn’t look very good if one person shows up to these meetings, so my wife came along to act as the business development executive,” Turner says. None of the venture capitalists bit, and Turner was burning through a meager stash of grant money. He decided to make a final stab at turning the holes into a business and hired a handful of people on the promise that he could pay them for four months. Luckily, that’s when Mohr Davidow Ventures showed up.
The prestigious Silicon Valley VC firm had been keeping a close eye on the sequencing market. They saw investors bail out of the first wave of sequencing companies after the early promise of the freshly mapped human genome proved too optimistic. The conventional wisdom was that a modest collection of genomes would reveal a great deal about the root causes of illnesses. Researchers simply needed some time to look over the genetic code and figure out which A, T, G, and C combinations were unusual and disease-causing. Bill Ericson, a Mohr Davidow general partner, rightly calculated that people were underestimating the complexity of the genome and the amount of sequencing work still left to do. In 2004, after performing a worldwide hunt for the next big thing in sequencing, Mohr Davidow settled on Turner’s technology. “We were blown away by the elegance of the solution,” Ericson says. “We didn’t see any laws-of-physics showstoppers that would prevent it from working. We saw a lot of hard development.” PacBio was born.
In exchange for its investment, Mohr Davidow called for Turner and his team to head to Silicon Valley. Turner then started hiring, beginning with a chief executive officer. He tapped Hugh Martin, a Valley veteran who’d run an engineering group at Apple (AAPL) and served as president of the 1990s-era gaming company 3DO. His big hit came via ONI Systems, an optical networking company he founded in 1998, took public, and sold to Ciena (CIEN) for close to $1 billion. ONI produced hundreds of millions of dollars in returns for Mohr Davidow, which had invested in the company, buying Martin a lot of goodwill.
Equipped with game-show-host looks and a fierce competitive streak, Martin has developed a legendary skill for wooing workers away from rivals. That came in handy at PacBio, which required a Manhattan Project-worthy menagerie of technical and scientific talent. He needed experts in nanotechnology, electrical engineering, thermal engineering, industrial design, organic chemistry, enzymology, signal processing, software, and more. In one case, Martin got in a bidding war over a superstar chemist with General Electric (GE) chief Jeff Immelt—and won.
The company had hoped to buy a lot of the technology it needed off the shelf. But the PacBio team came to realize they had to craft most of their own tools from scratch to meet their unusual requirements. The mishmash of such diverse skills added to the problems. At times they simply could not understand each other. “I have sat through half an hour of a meeting and realized that we’re not making progress because of these interdisciplinary vocabularies,” Turner says. “To an enzymologist, a substrate is part of a chemical process. To a nanochemist, a substrate is a flat thing you build machines on.” The staff resorted to inventing their own words to cover complex topics. The Monster, for example, is PacBionese for “the photo- physical phenomenon in which, if a fluorophore gets excited, and while in that state gets hit with the right wavelengths of light, can get further excited. Thus, instead of being in the S1 state, it will be in the S2 state.” And so forth.
The workers toiled away for years, getting hints that all of this effort would result in a working product but few proofs. To keep them going, Martin rewarded them with perks—a $100,000 gym, lobster feasts, and a custom 24-foot, twin-axle barbecue. The rewards worked. Computer scientists and biologists slogged on side-by-side through six-day weeks and holidays. One Thanksgiving they manned their work stations for the day and then stayed up all night cooking turkeys on the barbecue.
One limitation of sequencing machines has been their inability to sequence an entire 3 billion-base strand of DNA in one go. Companies in the field apply various techniques to carve up and analyze DNA, but they all tend to stumble over long, repetitive strings of bases. They see a bunch of T’s or C’s in a row and struggle to figure out how they weave into the whole. That drawback means hundreds of millions of bases in the genome are left unmapped. And it turns out that cancer cells and plants often exhibit repetitive and complex patterns of bases that give sequencing machines fits.
PacBio’s use of polymerase means its technology can handle much longer strings of bases, known in the trade as “long reads.” That breakthrough, plus the raw speed of PacBio’s systems, which can analyze a SARS or cholera variant in minutes instead of the usual hours or days, has the early RS customers excited. “Our hope is that in areas like tumors, it will provide us with better information about which therapeutics to provide to patients,” Steve McPhail, the CEO at Expression Analysis, which performs genomics services for pharmaceutical and biotech companies. “It will allow us to go places where we have never been before, and do things we have never been able to do before.”
Years into this adventure, PacBio still has a tough road ahead. The consensus view among rivals and some outsiders is that the company is unrealistic about its commercial prospects. Three years ago, PacBio executives were sending competitors into panic mode with claims about shattering speed and cost barriers. “Everyone had been thinking that the true next-generation machine would come from PacBio, but people seem to have backed off those claims now,” says Raymond McCauley, the chief science officer at the startup Genomera and a former executive at the sequencing company Illumina (ILMN). PacBio went public last October with its shares hitting $17.47. Today they trade at about $12.
Just as damning are the suggestions that PacBio will be relegated to niche status. It’s unclear how big an advantage they will get in the cancer and agriculture fields from the long reads. “We’ll have to see if they can come up with a killer app to sell lots of machines,” says Ion Torrent’s Rothberg.
The competitive jabs underscore the sequencing market’s resurrection. Illumina, based in San Diego, has set a torrid pace for all rivals, regularly pushing down prices. Where it took 13 years and $2.7 billion to churn through the human genome the first time, conventional sequencers can now do the job in days for a few thousand dollars. Jay T. Flatley, Illumina’s CEO, predicts the price of sequencing a human genome will fall below $1,000 in the next three to five years. Next stop: mass sequencing. “Ultimately, it will take a system where you put a drop of blood in one end, press a button, and get the answer out the other side.” If that happens, patients will be able to get sequencing done in their doctor’s office, and every man and woman will get their genomes mapped routinely throughout their lives. That would enable customized medicine and more detailed checks on how individuals’ body chemistry changes over time.
George M. Church, a professor of genetics at Harvard Medical School, says there are two dozen to three dozen companies in the process of building sequencing machines. He serves as an adviser to many, including PacBio. “They would all like to believe there is one market, and that they’re the ones who own it,” he says. “In reality there are about eight niches, and no one has them.” In other words, there’s room for a PacBio that’s good with cancer, an Illumina that caters to the mass market, and an Ion Torrent tempting new types of customers with low-cost gear. “This is the most rapidly changing technology I have ever seen,” says Church.
For now, Martin says PacBio will focus on its long-read technology. The company is working to add more holes to its chips and speed up its chemistry. “Everyone has to bitch about the new guy,” Martin says.
If Martin sounds irked, it’s because his stake in PacBio’s success goes beyond the financial. In September 2009 he was diagnosed with multiple myeloma, a devastating form of blood cancer. The disease had already devoured one vertebra and started to spread throughout his body. Martin had surgery to fix his back and started chemotherapy, all while PacBio sought to close another funding round and prepare for its initial public offering. “I looked around and thought about what I could do to get a cure as fast as possible,” he says. “There’s no CEO more motivated than me.”
http://www.businessweek.com/magazine/content/11_19/b4227056590447_page_4.htm
Tags: cancer, cure, decoding DNA, DNA, gene sequencers, genetics, illness, Illumina, PacBio RS, sequencing market, silicon valley, VC firms