April 14, 2006 | If Greater Boston and Cambridge, Massachusetts, constitute a tantalizing bazaar of biotechnology upstarts and prestigious academics hawking technology and seeking funding, then Phil Vickers is like a little boy with a twinkle in his eye and a pocket full of pennies.
Vickers is VP and site head of Pfizer’s Research Technology Center (RTC), plunked down in Cambridge, on the banks of the Charles River. It’s a six-year-old experiment in identifying, developing, and dispersing useful technologies throughout the largest drug-maker in the world. And the goodies aren’t only outside. Inside RTC, roughly 120 Big Pharma scientists work in a distinctly small biotech environment. That number will soon grow to 150.
Says RTC director of research informatics Giles Day, “Where else can I wander down the street and have a cup of tea with (Sir) Tim Berners-Lee?” — inventor of the World Wide Web, founder of W3C, senior researcher at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL), and a fellow Brit to boot. Day hopes to use Semantic Web technology — Berners-Lee’s latest passion (see Masters of the Semantic Web, Oct. 2005 Bio•IT World, p. 28) — to improve knowledge management inside Pfizer.
“I’ve been here for five years, and it’s never once occurred to me that I want to leave,” says Day.
Can Pfizer, a $50-billion-plus, 115,000-employee behemoth, succeed where other giants have failed and successfully institutionalize the effort to find and develop critical new technology? Should it even try, given that sales dipped last year, that the company promised Wall Street it would slash thousands of jobs, and that its 2005 R&D budget was a staggering $7 billion? And can a man who (proudly) socializes with professional wrestlers cajole a team of quirky researchers to produce concrete results?
Yes, yes, and yes, says Vickers. “It won’t work if we do blue sky research, if we’re the wacky scientists up in Cambridge doing funky stuff that doesn’t have any impact.” It will work, he believes, if the majority of RTC goals are derived from specific problems plaguing Pfizer and if RTC leverages its Massachusetts location to solve those problems. Otherwise, he acknowledges, “We may as well be some place cheaper.”
The early evidence is promising. At least one company — BioTrove — was spawned from an RTC-MIT collaboration (see Breaking the Speed Limit, p. 29) and the high-throughput mass spec technology it produced is now being rolled out to the larger Pfizer organization. A variety of RTC-built informatics applications are now used companywide. Current projects include the Semantic Web initiative, a real-time infrared system for monitoring chemical reactions, and a “smart” chemical structure drawing system, being co-developed with MIT’s CSAIL.
That last technology might soon permit researchers to sketch compounds on an electronic white board, have those compounds instantly recognized, and retrieve pathway data, best synthesis routes, and competitive patent information from databases.
Nick Terrett, RTC senior director and head of chemical services, anticipates that, “four years down the road, we’ll point to chemical technologies that are in common use across the Pfizer world that we initiated or kicked off. It might be a bit of software used to discover a compound; it might be methodology, like parallel synthesis.” Another RTC program, CTAN (chemical tool alert notice), has already helped pushed molecules into specific therapeutic programs (see CTAN: All About the Drugs, p. 28).
Unlike other sites under the Pfizer Global Research and Development (PGRD) umbrella, which are organized along therapeutic lines, the RTC is structured around scientific disciplines. It’s also substantially smaller. Although projects may exist within a single discipline, they most often entail multidiscipline teams doing the work.
Vickers seems a properly eclectic choice to run the RTC. Also a U.K. native, he took his Ph.D. (biochemistry) at the University of Toronto, worked in Scotland Yard’s forensic lab for a time (“great fun”), and had a six-year Merck stint before joining Pfizer where he has been for more than 11 years. He joined RTC as site head some 18 months ago, following three years as head of proteomics and genomics at Pfizer’s Groton, Conn., facility.
Pointing to a nicely framed photo of him shaking the hand of former World Wrestling Entertainment champion Brock Lesner, Vickers says, “You’ll notice I’m smiling because I’m crushing his hands.” Vickers works with the Make a Wish Foundation and says, “We routinely have kids who want to go and see wrestling and to meet Brock.” The combination of quirky bravado and savvy may help keep RTC’s wheels of creativity from screeching to a halt or crashing into a wall.
RTC began life as the Discovery Technology Center (DTC) in 1999, before Pfizer expanded the site’s scope to embrace development and safety sciences, drug metabolism, and pharmaceutical sciences around the same time Vickers became site chief. The basic idea is to mimic the agility of a small biotech environment and leverage ties to the local academic and biotech communities. The other key idea is to pick projects that are important to Pfizer.
Vickers sits on the technology leadership team for all of Pfizer. That group determines what the key companywide scientific problems are, and Vickers tries to match those needs with RTC internal expertise or to find external expertise in the Greater Boston region. Other RTC members sit on Pfizer-wide teams, typically along discipline lines. Vickers’ direct boss is Nicholas Saccomano, senior VP of research technology at Pfizer (see The Platform Diet, April 2005 Bio•IT World).
“The RTC maintains the technical expertise, the scientific breadth, and a spirit of innovation that has allowed this group to engage many of the difficult problems that make drug discovery a challenging and arduous enterprise. We view their inquiry and efforts as central to our future success,” says Saccomano.
RTC gets an annual budget, with annual goals, and is left alone to deliver, says Vickers. “Many big companies come after six months and say, ‘So, how’s it going? What course corrections do we need?’ We don’t have any of that stuff. If we weren’t delivering, we’d be hearing about it. If everything we did had to be vetted by Groton or the larger organization, the place would just grind to a halt.”
Disciplines represented at RTC include molecular informatics, research informatics, chemistry, biology, and more recently systems biology. RTC is also a Kinase Center of Emphasis for Pfizer. The site has its own “multimillion-dollar” budget for strategic alliances. Drug Pfinder is an active program with staff dedicated to identifying academics with promising technology or compounds and forging partnerships. RTC regularly brings in Harvard Business School folks to talk on how to innovate, and it’s the Pfizer point of contact for MIT’s Industrial Liason program.
Of course, there’s also some gee-whiz technology at the center. One conference room is outfitted for 3-D viewing of molecules. Researchers bespectacled with special glasses can look for promising or problematic nooks and crannies in putative drug targets.
It’s great work if you can get, and last year many tried — 2,500 applicants vied for 25 positions. Now, what will keep the Pfizer funding spigot open is proven success.
“The collaboration with BioTrove is the best example of every aspect of what I want this site to do,” says Vickers. “It was a good link into the academic community. It’s leveraging our location. A company was spawned from it. It’s a new technology, and we’ve now developed a second generation of it, which we’re now using routinely, and we’re rolling the technology out to other parts of the organization. What more could you want?” says Vickers.
It was a case of serendipity and need. RTC’s head of biology, Steve Faraci, knew Ian Hunter, a prominent MIT professor working in nanotechnology and bioinstrumentation. The two got talking about mass spectrometry’s limitation and how useful high-throughput mass spec would be. The result was a collaboration in which members of Hunter’s lab, and later BioTrove employees, moved into the RTC building.
Over a few years, this group developed a “rapid-fire” plate-based system for feeding samples into a mass spec that was 30 times faster than existing tape-based systems. The original prototype still resides at the RTC. BioTrove has since moved out, and actually changed its business emphasis somewhat, but Pfizer is aggressively using the new technology to speed screening.
Not surprisingly, RTC has a large informatics contingent, including 11 in research informatics under Day and 15 under director of molecular informatics Enoch Huang. The collaboration between the two can make it hard for an outsider to distinguish between them.
“The research informatics line is a global line that runs through [all] sites,” explains Huang. “Molecular informatics is called different things on different sites. Sometimes it’s called bioinformatics; sometimes it’s called computational chemistry. It really depends. Suffice to say, they are scientists who know how to do some computing, whether it’s PERL scripting, genome analysis, or cheminformatics analysis, or high-throughput analysis.”
Day adds, “The creation of tools tends to be partnership between the two groups. So what you tend to see is prototyping of algorithms in molecular informatics and then partnership with research informatics to take those and make them robust standardized tools. So they’ve got a problem, they know their problem and how to fix it, but they might not always write the best code, and they certainly don’t document the best code.”
Whenever possible, commercial software is used, and there is no easy recurring ratio of commercial-to-homegrown software in use. It all depends on the need.
“We’re big into structure-based modeling,” says Rob Stanton, associate director of molecular informatics and head of computational chemistry. “I can bring some type of technology to bear on the problem, and by having all the skills here in house, we don’t have to go out and talk to a vendor to try it out. Other sites have to concentrate on drugs.”
Huang adds, “Sometimes we get into partnership with vendors, and there’s customization that happens with the two parties working together, such as with Spotfire. We pay Spotfire for those Pfizer-specific modifications. We have funded FTEs at Spotfire for guides or modifications so they integrate better with our systems.”
Research informatics has a dual role: a small group runs the site — desktop support and networks, etcetera — while a larger group is devoted to development. “Our role is to take new informatics technologies and push them out,” says Day.
Not surprisingly, workflow and collaboration tools — critical for enterprisewide efficiency — are high on Day’s development list. RTC, for example, is Pfizer’s Center for Emphasis for Kinases, and Day’s team is putting the finishing touches on a Kinase Knowledgebase that will be accessible throughout Pfizer. Occasionally, Day receives a specific request from a therapeutics site, as in 2002, when he received a transatlantic request from Andrew Hopkins at the Sandwich, United Kingdom, site for an application to relate targets to compounds through literature data.
“They needed a whole bunch of code written to do that. We did the prototype. As soon as it became apparent it was going to have value, then global research informatics group picked it up and hopefully threw out most of the code because I wrote quite a lot of it. In fact, they only got rid of the last bits of my code a few weeks ago,” says Day. Pharma Matrix has been rolled out globally.
Another success was the development of lab automation workflow software able to control steps from chemistry synthesis to compound purification. It’s now used globally throughout Pfizer and led to an ongoing and strong relationship with Teranode, which supplied the basic platform.
“One reason we like the Teranode [solution] is you can customize it so easily,” says Day. “If you take the solution to another site, maybe they’ve got a slightly different piece of machinery or something, but you don’t have to phone up the vendor and get into a change to solution. All you’ve got to do is create a new file report of components, and you’re off.” A Teranode consultant is “on site all the time because we’ve changed the shape of that product fairly dramatically.”
Exploring Semantic Web technology is a big part of Day’s 2006 agenda. (Coincidently, Teranode recently hired Eric Neumann, who was leading the Semantic Web charge at W3C before joining Teranode.)
“We intend to use the Kinase Knowledge base to test-drive a whole bunch of Semantic Web technologies,” says Day. “I’ve very much bought into the vision of the Semantic Web. I have significant concerns about how it’s going to perform in an environment like ours. Obviously I’ve seen the demos. To be blunt, the demos that they’ve done are of little or no value to us. Integrating genomic data is, as far as we’re concerned, a solved problem.”
Semantic Web’s value for Pfizer, says Day, will be in enabling data to cross organization boundaries more easily. He has concerns over security and trust but adds wryly, “Eric and Tim Berners-Lee assure me that they are going to be able to solve those relatively quickly. We’ll see.”