By Meredith W. Salisbury
You probably didn’t notice at the time. Some six months ago, wheels started turning in the Manhattan-based GenomeWeb offices. It was time to start work on the second annual GT All-Stars, where readers choose the industry luminaries in several technology categories. This month, GT honors the winners — and recognizes all the nominees — with an awards bash in Boston, and the profiles on the following pages.
We changed our voting process slightly this year by adding a nomination phase. In an online poll, we asked you to tell us whom you think are the best and brightest in each category. Our editorial team composed the final ballot based on your input. In the second round, readers went to our website and voted for their favorite nominees. In a terrific turnout, nearly 3,000 of you deliberated this year and named the eight most accomplished, forward-thinking, and respected experts in their fields (for the complete list of winners and runners-up, see p. 52). Voters also selected GT All-Stars winners in the company and product of the year categories. The following pages profile the winner in each category. After talking extensively with these leaders, it’s easy to see how undeniably each has influenced genomics technology development, and we look forward to watching their progress as they continue to have an impact on the field. 2002 GT All-Stars
Most Outstanding in Sequencing Technology 1 Elaine Mardis, Washington University 2 John Nelson, Amersham Biosciences 3 Lee Hood, Institute for Systems Biology 4 Michael Hunkapiller, Applied Biosystems 5 Daniel Branton, Harvard University Most Masterful in Microarrays 1 Stephen Fodor, Affymetrix 2 Pat Brown, Stanford University 3 Joe DeRisi, University of California, San Francisco 4 John Quackenbush, TIGR 5 Michael Snyder, Yale University High Performance Computing’s Highest Guru 1 Shankar Subramaniam, University of California, San Diego 2 Jeff Augen, IBM Life Sciences 3 Ray Hookway, Compaq 4 Andrea Califano, First Genetic Trust 5 Songnian Zhou, Platform Computing Most Innovative in Bioinformatics 1 Martin Leach, CuraGen 2 Christopher Ahlberg, Spotfire 3 Sean Eddy, Washington University 4 Lincoln Stein, Cold Spring Harbor Laboratory 5 Jill Mesirov, Whitehead Institute Center for Genome Research Most Prolific in Proteomics 1 Ruedi Aebersold, Institute for Systems Biology 2 John Yates, Torrey Mesa Research Institute 3 Scott Patterson, Celera Genomics 4 Samir Hanash, University of Michigan 5 Anne-Claude Gavin, Cellzome Database Doyen/Doyenne 1 Jim Kent, University of California, Santa Cruz 2 Greg Schuler, NCBI 3 Ewan Birney, EBI 4 Janan Eppig, Jackson Laboratory 5 Christopher Hogue, Samuel Lunenfeld Research Institute Most Visionary Dealmaker 1 Jonathan Rothberg, CuraGen 2 Steven Holtzman, Infinity Pharmaceuticals 3 Dale Pfost, Orchid BioSciences 4 Friedrich von Bohlen, Lion Bioscience 5 Carol Kovac, IBM Life Sciences Person of the Year 1 Howard Cash, Gene Codes 2 Sue Siegel, Affymetrix 3 Craig Venter, J. Craig Venter Science Foundation 4 Claire Fraser, TIGR 5 Kathy Ordoñez, Celera Genomics and Celera Diagnostics Product of the Year 1 AcycloPrime-FP, PerkinElmer 2 Human Genome U133 Set, Affymetrix 3 Rolling Circle Amplification, Molecular Staging 4 TempliPhi, Amersham Biosciences 5 Bioperl, bioperl.org Company or Institute of the Year 1 Affymetrix 2 PerkinElmer 3 Strand Genomics 4 Jackson Laboratory 5 CuraGen and Spotfire (tie) Elaine Mardis
Assistant Professor, Washington University Co-Director, Genome Sequencing Center Most Outstanding in Sequencing Technology
When Washington University’s Rick Wilson set out to find a technology development director for the Genome Sequencing Center, he didn’t have to look far. He had known Elaine Mardis from their days together in Bruce Roe’s lab at the University of Oklahoma, where she began in 1984 and got her PhD working out protocols for the new field of automated sequencing. “She’d been working in industry,” recalls Wilson. “It was sequencing from an industrial point of view … exactly what we needed.” Mardis has since made a name for herself as a leader in the field.
“She’s really been a key player in the throughput increases and cost decreases that we’ve had over the last several years,” Wilson says. “Not only in developing novel methods and technologies in the lab, but going out and finding out what other people were doing.”
Mardis, a tenure-track professor who recently passed the “odometer birthday” of 40, remembers starting at WashU. “Early on in this job, I used to worry about whether there would be enough to do,” she says. “But there’s never a break in the action at all.”
Her earliest work in the field began in Roe’s Oklahoma lab, where she learned radioactive sequencing and got to work with the ABI 370A — “the second instrument ABI ever made” — and the Biomek 1000.
Roe has kept up with Mardis’ work since then. “Elaine was actually the first person to automate the M13 template isolation,” he says. “[She] really has a grasp of the methods for DNA sequencing and how to get them optimized and automated.”
She spent more than four years after grad school in the R&D department at Bio-Rad working on DNA prep technologies and robotics, but always wanted to work on the Human Genome Project. “[Bio-Rad] was a really good training ground,” she says, pointing out that her industry time gives her an edge when bringing new instruments into the sequencing center. “We may be taking in someone else’s product, but it’s rarely handed off to the production team with the same protocols given to us by the manufacturer,” says Mardis, who oversees a group of molecular biologists and engineers responsible for tweaking or souping up new technologies.
Mardis says sequencing is changing, and foresees it taking place “more along the lines of examining clinical specimens.” She’s already in talks with people at the WashU med school about examining diseases on a genomic level. Sequencing is far from over: “There’s still a lot of potential for DNA sequencing,” she says, “but maybe not as we know it now.”
Scale and cost will be two critical factors in the future of sequencing technology. “You can’t spend $100 million to sequence for clinical genomics,” she says — a range of $1,000 to $5,000 is more in line with reality, and that likely won’t happen on current platforms, even if they are optimized. “New instrumentation is going to be the key,” she says, looking forward to one machine that will handle all the sequencing steps.
Eventually, we’ll see “hip-pocket DNA sequencers that [people] can take to a disaster site and do forensics,” Mardis says. Diagnostic sequencing is also on the horizon, but it’ll be a big leap from current technology.
One critical bottleneck lies in integrating sequencing tools with sample prep. Mardis and her group are working with the new ABI 3730, which has improved sensitivity but still doesn’t allow for such integration. Despite state-of-the-art, nanoscale liquid-handling and thermocycling robots, “the irony … is it gets put on a standard microtiter plate and schlepped over to the 3730. That’s the interface that people need to focus on,” she says. “It has to be all hands-off.”
Out of bed New advances in technology development — “that’s what gets me out of bed at 4:30 in the morning.”
Stephen Fodor
CEO, Affymetrix Most Masterful in Microarrays
What Stephen Fodor seems to remember most about those early years of microarrays, aside from how certain he was that academia was the place for him, is the sleepless nights. He blames them not on nerves from the startup environment but on excitement — “just thinking about the really neat things you could do with the technology, that’s what kept me awake at nights,” he says.
Fodor, trained in biology and chemistry, was at the University of California, Berkeley, as a postdoc when Lubert Stryer tracked him down to talk about this new company, Affymax, where Stryer was to be scientific director. Fodor had no interest in heading to industry. But Stryer was persistent, calling Fodor at work, at home, wherever. “I swear I could’ve been walking down the aisle at Kmart and I would’ve been paged by Lubert Stryer,” Fodor, 49, laughs. It worked: in 1989, he agreed to join Affymax for a year and see how things went.
The rest, of course, is history. Fodor, who began working on array technology, was hooked. “We laid out a number of different things: how to make arrays, ideas for applications for them, fluorescent labeling, confocal scanning, combinatorial chemistries for how to actually fabricate them.” The result was a paper covering nucleic acid and peptide arrays and an overview of applications in genotyping, gene expression, and sequence analysis that graced the cover of Science in February 1991.
That paper fulfilled the potential Stryer recognized when he first met Fodor. “What I saw right away was a person with wonderful technical expertise but much more — he had a certain spirit of science,” Stryer says. “He got it very quickly and went beyond you very quickly.”
Affymax turned toward drugs, so Fodor spun out Affymetrix in 1993. By then, there was plenty of talk about the Human Genome Project, and it struck him that with microarrays, “what we really had was an information storage system for DNA.”
People began to visit Affymetrix to see what Fodor was sitting on. “You wouldn’t believe the number of people that thought we were just nuts,” he says. “‘Even if you could put 100,000 different pieces of DNA [on a chip], why would you want to?’ This is what people would say in the early ’90s. ‘We can do everything we want with 10 probes.’”
Fodor delights in pointing out that in the last quarter alone, Affy shipped nearly 100,000 chips with 500,000 to 600,000 probes on each. But the gratification of being able to say “I told you so” isn’t what keeps him going. “I love technology — technology in general,” he says.
That’s why he set up Affymetrix Laboratories, a division of the company with its own building where scientists are “sequestered” to work on future projects. Affy Labs — which Fodor compares to Bell Labs or any of the science-forward units of Hewlett-Packard, IBM, or Xerox — is responsible for Perlegen’s 60-million-probe genome-scanning chip concept, and recently published its transcription research, which indicates that “there’s at least a factor of 10 more stuff going on at the transcription level in the human genome” than anyone thought, Fodor says.
As for microarrays, Fodor thinks the basic technology is a done deal. He shrugs off debates about cDNAs versus oligos: “Oligos are much more powerful analytically.” Relevant questions are more about what should go on the chip — mRNA or something else? — or how the genome should be laid out.
The really important question, though, is where the microarray applications are going. Chips will obviously continue to play significant roles in expression and SNP studies, Fodor says, “but the area I’m most interested in is where you start to see the crossover of polymorphic nature and its effects on biology.” That, and increased use of chips in genetics, such as for population studies, is what’ll make microarrays particularly useful for healthcare, he says.
Out of bed Between 6:30 and 7:30 to get his kids to school, but he rarely goes to sleep before 1 or 2 in the morning.
Shankar Subramaniam
Professor, University of California, San Diego Professor, San Diego Supercomputing Center High Performance Computing’s Highest Guru
It’s amazing that Shankar Subramaniam has time to be a noted leader in any field, given the number of hats he wears. His post at UCSD covers bioengineering, chemistry and biochemistry, and biology. He’s the director of the school’s bioinformatics graduate program, has joint appointments at the Salk Institute as well as the San Diego Supercomputing Center, and is bioinformatics director for the Alliance for Cellular Signaling.
But GT readers resoundingly chose Subramaniam, born and raised in India, as the HPC All-Star this year. Bernhard Palsson, a colleague who helped recruit the computing guru to UCSD from his former post at the University of Illinois supercomputing center, illustrates the respect so many people have for Subramaniam. When Palsson got a call for this story, he was rushing out the door. But after realizing it was about Subramaniam, he insisted on taking the time to speak about him. “Since Shankar came to campus … he has really had a transforming effect on many programs and activities here. He drove the formation of our bioinformatics degree program [and] the NIH training grant in bioinformatics that we now have,” Palsson says.
Subramaniam, 49, got his start in statistical mechanics in the mid-’80s. But the richness of biological data lured him away from his beloved chemistry and physics problems and he launched a faculty career in computational biology. “The thing that a lot of people are grappling with today is trying to structure all of this very different knowledge and capture it in a computational model,” he says. “The devil is in the details.”
Subramaniam, who, according to Palsson, “makes generous use” of the supercomputer facility at San Diego, is most interested in making sense of the huge volume of biological data without losing the nuances of it all. His lab primarily works on designing computational models for cell signaling networks (including statistical analysis of microarray data) and on data-mining from protein structures. The protein folding in particular “requires fairly high-end computing,” Subramaniam says.
The supercomputer center has IBM machines, various types of Sun enterprise servers, and clusters. Subramaniam wouldn’t give up any of them, and sees a future for the supercomputer as well as the cluster. “There are times when you need to have large amounts of data in memory, and you can’t do it on distributed computing,” he says. “[But] clusters are good because they are cheap, they are fast.” The beauty of programming with languages like Java and C++, he says, is their portability. “You compute almost in a seamless manner” across all the platforms and types of machines.
Subramaniam says developing useful infrastructure and an ontology are critical for computational biologists. “The heterogeneity of data that drives the computation [requires] an infrastructure, a well-defined, structured language” to allow for data mining, he says.
“All the numerical computational elements have been pioneered by physicists,” he says, but the problem is that data in other disciplines tends to be very homogeneous, even if there is a lot of it. Bringing computing into a field as complex as biology demands that users rethink how to handle data and build models. “How to design a computation that’ll use the heterogeneous structure and large volumes of data, this is clearly the largest challenge biologists are facing today.”
Cell signaling networks have become Subramaniam’s pet project, and he expects that it’ll take the next 20 or 30 years before people figure out how to decipher them properly and build models on the large network level. “That’s a field which is really growing and growing every day.”
Out of bed Usually up around 6:00, rarely in bed before 1:00. “This has been my modus operandi for the last 20 years or so.”
Martin Leach
Director of Bioinformatics, CuraGen
Most Innovative in Bioinformatics
In an industry like this one, often it’s about whom you know. For Martin Leach, it’s about knowing everyone. One of the three directors of bioinformatics for informatics-turned-biopharmaceutical company CuraGen, he is responsible for discovery informatics and makes a point of spending as much time as possible with his end users, both inside and outside the company. It’s that exposure to people in the community, he guesses, that won him this award. Indeed, the community concept is familiar to Leach, who from 1995 to 1997 created and maintained the still-popular online Biotech Rumor Mill, which he sold to Biofind. “Hundreds of scientists at various sites around the world, they have used the informatics I built,” says Leach, 34, who has been at CuraGen full-time for the past five years and says his informatics system has a level of integration other packages don’t offer. “I’ve played a very strong role in developing the bioinformatics support between CuraGen and each of our collaborators.” SurroMed’s Keith Joho bears that out. Formerly in the bioinformatics group at Roche, he met Leach during an alliance between the companies. “He’s been in the game for a long time,” Joho says. “A lot of what customers see is from his group.” A biologist by training with his PhD in pharmacology, the London native says he was working on “building collaborative [Web-based] environments for scientists to interact and communicate,” enabling researchers to compare notes or even integrate data from different experiments, when CEO Jonathan Rothberg tracked him down and invited him to bring that kind of integration to CuraGen. At the time, there were five or 10 informatics people at the company, Leach recalls; today there are more than 60. Compared to what he used to do when the company was a target-finding operation and his job was “the strip-mining of the human genome,” he says, “my work is now moving downstream.” His informatics jurisdiction goes from identification of drug targets and continues to the preclinical phase. It’s a trend that Leach sees affecting the whole bioinformatics industry. One goal he envisions for the field is “building the necessary tools to become more predictive in nature, whether it’s predictive markers for toxicity … or predictive in general, about whether a drug is going to succeed or fail at IND, phase 1, phase 2, phase 3.” The ultimate idea is to wind up with prototypical cells and whole model systems, says Leach, who has advocated for an increased focus on systems biology. “Models that allow us to predict what will be a good drug … that is where I see bioinformatics having to move.” But getting to a true systems-biology level will be no small feat, Leach acknowledges. “First you [need] a really systematic approach to generating data of high quality. Once you have these large sets of biological observations, [then you need] the ability to integrate them together.” The big picture, of course, is moving from genes to disease — a process Leach speculates about, though it doesn’t yet exist. He says it’ll come to pass when researchers can use a computational model to reliably integrate genes, physiology, symptoms, diseases, and syndromes, and bring it all back to drugs. A key side-project to that, he notes, will be the ability to mine biological literature and combine it with data. “Integration of that store of information is missing,” he says. “The people who have really touched on it are outside the area of biology. … Lexis-Nexis, the FBI, CIA — they know how to mine literature.” Out of bed Same time as his one-year-old son, usually between 5:00 and 6:00 a.m. Ruedi Aebersold
Co-founder, Institute for Systems Biology Most Prolific in Proteomics When Ruedi Aebersold gave up his position at the University of Washington in 2000 to help found the Institute for Systems Biology with Lee Hood, he knew it was a risky step. But the proteomics pioneer felt confident. “In the proteomics area, in the functional genomics area, there’s good support. If one does reasonable work in that field, certainly right now there is funding around,” he says. One man’s ‘reasonable work’ is another man’s breakthrough. Aebersold, credited with kicking off the proteomics field, has been hard at work on his isotope coded affinity tag technology, or ICAT, which has become widely used throughout the industry. Using a chemical reagent kit licensed to Applied Biosystems, ICAT “makes mass specs into quantitative devices,” Aebersold explains. “We can [tell] not only which proteins are there, but in what abundance.” Anyone who has a mass spec can try ICAT, and many scientists have. “A number of people used the reagent and all of a sudden found out it generates a huge amount of data. Then they run immediately into the next bottleneck, which is how do you process all this data?” says Aebersold. He and his ISB team are busy building software to handle the ICAT data, and the method has become so popular that ABI, Thermo Finnigan, and Micromass all offer software to support it, Aebersold says. He was delighted at a conference this past June to see several posters from people trying to tweak the ICAT chemistries. “Quite a large number of people are now trying to come up with variants of the general idea, which I think is great,” he says. Aebersold, 47, trained as a protein chemist in his native Switzerland before heading to postdoc in Lee Hood’s prolific Caltech lab in 1984. It turned out to be a career-altering move: though Aebersold left in 1988 to start his own group at the University of British Columbia in Vancouver, he and Hood reunited in 1993 when Hood moved his lab to the University of Washington and asked Aebersold to join him in the new department of molecular biotechnology. The two would later launch ISB in an effort “to start a smaller but much more focused research organization” than the university environment allowed for, Aebersold says. Hood calls Aebersold “one of the most outstanding scientists I know in terms of really driving technology development. … With ICAT, Ruedi has just continued to do spectacular pioneering right at the leading edge.” Despite the accolades, Aebersold won’t be taking it easy anytime soon. “There’s a lot of work yet to be done to advance proteomics,” he says. Technology development is really needed in the field, primarily “high-throughput and quantitative, accurate tools that can measure these proteins on a proteome-wide scale,” he says. One ideal tool would solve the dynamic range challenge. “The difference in abundance from the most abundant to the least abundant proteins is huge — maybe up to 10 orders of magnitude. There’s no single technique that has this range.” He predicts that the mass spec will continue to be a crucial instrument for the field, “at least for the foreseeable future.” But he expects that chip-based technology, using antibodies or aptamers, could come to the fore soon, with an eye out for new approaches based on atomic force microscopy as well. “Such methods might emerge which would allow one to identify the protein compounds of a cell based on scanning probe techniques,” he says, but adds, “That’s clearly not something that’s going to be available or even useful in the next two to three years.” Out of bed Without an alarm clock, between 7 and 7:30. “It’s torture to start the day by an alarm clock.” Jim Kent
Research Scientist University of California, Santa Cruz Database Doyen When Jim Kent talks about his role in the Human Genome Project — the weeks of fiendish programming for the genome assembler that got him the recognition of the community and numerous mentions in the press — he might as well be reporting that it’s raining in Seattle. Kent, 42 and so unassuming that he declined to have his photo taken for this story, laughs when his seeming lack of excitement is pointed out to him. “I just tell this story a lot,” he says. It’s a story he’ll have to keep telling, but as the readers who chose him as this year’s database All-Star can attest, it’s a story that keeps growing with Kent’s continued accomplishments. Best known for his GigAssembler program that was used to put together the public side’s draft sequence, Kent has been working since then on the annotated genome browser for UCSC, which now covers human and mouse. His annotations, reputed for their depth, include information on mRNA alignments, EST alignments (with all 10 million public human ESTs), and cross-species alignments from research at UCSC, the public sector, and third-party information as well. Everything is filtered to show the best data as quickly as possible, and his BLAT tool enables users to “put in a little of the thing [they’ve] sequenced and get back all of the places it hits in the genome within a second,” Kent says. David Haussler, who works with Kent at UCSC, recalls being “totally floored at how fast he picked up new algorithms and ideas, how creative he was” when Kent headed back to UCSC for a PhD in biology in 1995. Haussler adds, “There must be other people that do extraordinary work with software, although I’ve seen none better in my entire career.” A major difference between Kent, who was born in Hawaii and raised in San Francisco, and many of his fellow database peers is his background: with a master’s degree in math, he spent 12 years writing computer art and animation programs, mostly working for his own company. That gave him a perspective that was sorely needed in the bioinformatics field. “Coming from a consumer software background,” Kent says, he approached the genome browser and database “expecting the user to have a nice, interactive experience.” To that end, he’s focused on making the free browser fast and simple. What began with 100 Linux boxes is now a 1,000-machine cluster, still using Linux, relying on a number of Kent’s job-control and other algorithms to keep it running smoothly. While many database-builders invest in a costly supercomputer and put their data in RAM, Kent says “we were able to get the same thing by a combination of clever indexing and laying out the data on the disk in the same order it is in the genome.” Some features remain on the fly, such as sequence or BLAT searches. “Behind the scenes,” Kent says, “we’re aligning 10 million sequences.” The UCSC browser and database are fairly well established, but Kent, whose own goal is to figure out regulatory pathways, sees no shortage of work in the offing. A significant problem for databases is the difficulty in tracking down the last quarter of genes, most of which are so rarely expressed that he says the only way to find them may be with solid gene-prediction programs that incorporate mouse-human alignments. Alternative splicing and incorporating disparate sets of expression data represent even more hurdles coming down the pike, he adds. “Maybe the biggest challenge … with the human genome in particular is going to be integrating the literature in a very good way,” he says. “I’d like to see it become a real focus.” Out of bed Prefers to wake up around 10, goes to bed at 2 or 3 in the morning. “Mammals started out nocturnal, maybe I’m reverting.”
Jonathan Rothberg
President and CEO, CuraGen
Most Visionary Dealmaker
It’s somehow fitting that this business-oriented award goes to Jonathan Rothberg, who, in addition to studying chemical and biomedical engineering at Carnegie Mellon and biology at Yale, had 20 years of business training. Not possible for a 39-year-old? “Dad was an entrepreneur, so I got to hear [his lessons] every day at the dinner table,” Rothberg explains.
CuraGen started out as a genomics company in Rothberg’s basement in 1993, and the New Haven, Conn.-based company has rapidly evolved into a biopharma. Key to that progression, Rothberg says, were the deals he and his team brokered with big pharmas and biotechs along the way. It didn’t begin easily. He remembers being criticized over the years, continually questioned by potential investors about why he didn’t just set up shop as a database company. “I said, ‘Well, that’s going to blow up,’” he recalls. He likens the database business to seeing a movie: “I go to the movies, then it’s on pay TV, then it’s on free TV.” The lesson for genomics companies: “What you have to own is the drugs.”
The main problem he sees today for competitors is the early model of selling targets. “Any of the biotechs that sold their targets are going to be out of business,” he says. “You have to think about the long term.”
Thanks to its strategic deals, according to Rothberg, CuraGen is working on more than 100 monoclonal antibody projects and some 40 small-molecule targets — without ever making an antibody or spending a penny on screening. The trick is getting other companies to pay for his crew’s work. Abgenix invested $65 million and scales up and characterizes all the antibodies, leaving CuraGen “to focus on what we know best — the human genome,” Rothberg says.
Likewise, Bayer invested $85 million in the company to develop diabetes- and obesity-related targets for small-molecule development. Bayer pays for the first $375 million in CuraGen’s expenses for the work. The upshot is that CuraGen doesn’t have to dip into the $450 million it has in the bank, and it owns half of any drugs developed with Abgenix and 44 percent of the profits from any drugs with Bayer. In the meantime, the company has its own work in protein therapeutics, all told giving Rothberg a fat pipeline to delve into. “It’s not one deal, it’s the way we’ve strategically placed ourselves” that has spelled success for his company, he says.
Rothberg never misses the opportunity to work out a deal. In fact, he’s perplexed by the recent trend of acquisitions in the field — if you bring a technology or whole company in-house, he reasons, you’ve missed the chance at a strategic deal.
That philosophy reflects much of the advice Rothberg has gotten over the past 10 years from Phil Whitcome, a startup veteran and an advisor to CuraGen. “My advice to him in the early days was don''t sell anything before you know its true value in the marketplace,” says Whitcome, who believes Rothberg has been successful doing that. “He''s tried to do things that are strategic that dovetail into the long-term interests of CuraGen.”
The idea for CuraGen, now with almost 500 employees, came to Rothberg when he decided that he probably came too late to realize his dream of changing the computer world. “As an engineer, you want to make a huge difference,” he says. When the Human Genome Project came around, it occurred to Rothberg that he could string together his experience in chemical and biomedical engineering and molecular biology and “change that genome into products.”
But the outlook wasn’t always as good as it is now. He recalls reading interviews where Bill Gates said if he had to start a company in the early ’90s, he would’ve done it at the intersection of biology and IT. “I used to cut out those quotes and show them to employees and say, ‘Don’t worry what the VCs say.’”
Out of bed “Whenever my kids get me up.”
Howard Cash
President, Gene Codes
Person of the Year
In a way, it’s remarkable that Howard Cash was even nominated for GT’s Person of the Year. Cash spent the better part of the past year mired in nondisclosure agreements that prevented him from telling anyone what his company, the Ann Arbor, Mich.-based Gene Codes, was up to. From late September 2001 to now, the 30-strong crew of Gene Codes has been fiendishly working on building a software program to track and genetically identify, with STR and forensic SNP analysis, those who died at the World Trade Center site on September 11, 2001.
“There have probably been four weeks since September 27 when I haven’t worked all seven days of every week,” says Cash, 42, who typically spends two or three days each week at the medical examiner’s office in New York City. “I’ve always put in long weeks, but I’ve never worked this hard on anything in my life. … I plan my week around trying to get five hours of sleep a night.”
For Cash, the bioinformatics field has provided plenty of opportunities to make a difference. The one-time musical composer has been working in the industry since 1984, starting at IntelliGenetics, and has operated on projects including writing software to sequence the AIDS virus; helping a Dana-Farber group win its race to identify the MSH2 colon cancer gene; supporting the preservation of endangered species; and “[returning] the remains of loved ones to families from tragedies as distant as the second world war and as recent as the World Trade Center disaster,” he says.
Because remains identification was a niche for Sequencher, Gene Codes’ main program, Cash was known to customers such as the FBI and AFDIL and was an obvious contact when New York’s chief medical examiner’s office was faced with the Trade Center wreckage. Bob Shaler, director of the forensics biology unit there, met with Cash. “There were no software vendors out there who could do exactly what he could do,” Shaler recalls. “They bring new updates of this very special program that they’re writing for us every week.” And to Shaler, the importance of the new software goes even beyond the 9/11 tragedy. “This product, when it’s finished, [will be] something that should be applicable to any mass disaster anywhere in the world.”
And when it’s finished, life might return to normal for Cash and the Gene Codes crew. There’s a building backlog of feature requests from Sequencher customers that have to be addressed, for starters. And he’ll be able to pay attention to the ups and downs of the bioinformatics industry full-time again. Cash, who kept his company profitable for 35 consecutive quarters, predicts that the current merger mania will begin to die down. “There will be fewer mergers where bioinformatics companies join to shore up their own weaknesses,” he says, pointing to eBioinformatics and Oxford Molecular as examples of that in the past.
He issues a warning to the industry based on a lesson he’s learned over the years: “It is critical not to be seduced into pursuing what’s fascinating at the expense of what is useful.” He seems to lump mapping out the whole proteome under this umbrella. “I’m less bullish on proteomics than most people,” he says. Though it’ll uncover masses of interesting information, he likens it to “an academic exercise similar to determining all the grammatical structures in a spoken language.”
For now, though, Cash isn’t looking too far past the work at hand. “I honestly do not believe I will ever do anything in my professional life as important, as significant, as demanding, or as fulfilling as this.”
Out of bed At 5, though he’d rather it were 10.
Affymetrix
Company of the Year
With its microarray chips, scanners, and readers that have become the industry’s gold-standard, Affymetrix is one of the biggest success stories in the genomics technology market. After nine-and-a-half years in business, its products are prolific — Affy shipped 97,000 chips just in the last quarter.
But instead of Affy’s products, GT readers more often cited the company’s attitude and leadership when naming it Company of the Year. Voters credited Affy for having “the best attitude in the industry,” for being “the most innovative company out there,” and for being “the ultimate genomics leader [with] innovative technology, excellent vision, smart people, and products that make money.”
In fact, the company has been pro forma profitable for the past three quarters and is projecting GAAP profitability in Q4. And while other genomics tools companies are making cutbacks, Affy has the nice problem now of figuring out into which areas it will expand.
Anne Bowdidge, director of investor relations, says Affymetrix is looking, among other things, at genotyping: “We think our technology will be uniquely suited to assay tens to hundreds of thousands of SNPs to help find the genetic causes of complex diseases and help maximize the information gained from affected and control patient samples for disease research and clinical trials.” And for DNA and RNA analysis, she says, “we are less focused on the custom market for a few SNPs at a time … Instead, we will make large, economical batches of very small standard chips for diagnostic and clinical applications. We believe that just as whole genome RNA analysis has become the standard, the same thing will be said about whole genome DNA analysis just a few years from now.”
Affy’s popularity with its customers hasn’t come without effort. The company has been working steadfastly lately to keep its hold on the microarray market and to polish its image. Since this time last year (when Affymetrix president Sue Siegel graced GT’s cover for a story about the company’s new affability) Affy has announced 40 percent more chip shipments, 50 percent growth in installed instruments, and twice as many academic and three times as many biotech access customers.
And it seems that the image campaign has been working: both Siegel and CEO Stephen Fodor were nominated as All-Stars themselves. Siegel was a runner-up for Person of the Year, and Fodor won his category for Most Masterful in Microarrays.
What’s been the key to success for the nearly 1,000-employee company? Says Bowdidge, “We mobilize our resources very efficiently.”
— Adrienne Burke
PerkinElmer’s AcycloPrime-FP SNP Detection System
Product of the Year
In the interest of full disclosure, we’ll straightaway note that PerkinElmer is a sponsor of the GT All-Stars. But no, the company didn’t buy the Product of the Year award for its AcycloPrime-FP SNP detection system. It won votes the old fashioned way: campaigning. PerkinElmer used its own website to urge AcycloPrime fans to vote, and made it easy for them by providing a hyperlink to the GT All-Stars ballot page. The company’s efforts paid off. AcycloPrime got nearly 50 percent of all votes cast for the category, beating out Affymetrix’s U133 human arrays, Molecular Staging’s Rolling Circle Amplification, Amersham’s TempliPhi, BioPerl, and ABI’s new 3730 DNA sequencer, which split the remaining votes almost equally.
PerkinElmer developed the SNP-detection assay based on work by Pui-Yan Kwok, then at WashU but now at UCSF.
AcycloPrime’s major contribution to the market, says Walter Tian, PerkinElmer’s business director in molecular biology, is that it brings genotyping to the masses, both in terms of cost and ease of use. “Every single research lab can do SNPs without having to commit huge amounts of capital investment or technical expertise,” he says.
The savings come in two areas: primers and detection. “A lot of competing assays use labeled primers, which are much more expensive than the unlabeled, unpurified primers you can use with AcycloPrime,” says Richard Greene, who headed up the product’s R&D team. And while competitors such as Sequenom require several hundred thousand dollar investments in detection instruments, the fluorescence polarization detector required for the AcycloPrime assay lists at a more affordable $25,000.
The SNP kits come in three sizes: 1,000, 10,000, or 100,000 assays and list at $600, $4,250, and $22,500, respectively. They each include PCR cleanup reagents and buffer and a pair of dye terminators.
Kwok’s response to hearing that AcycloPrime may be awarded product of year status: “I hope it deserves it.”
—Aaron J. Sender