Engineers and Eggheads Molecular Sciences Institute
by Karen Hopkin(Posted July 10, 1998 · Issue 34)
Abstract
Diverse, but alike in their propensity to take risks and look for the big
picture, these molecular biologists, physicists, and computer programmers
join forces hoping to shape the way biology is done in the 21st century.
"It's crucible time," says Roger Brent, absently running a hand through his hair. Typical Brentian shorthand, the comment describes the intense nature of the birth in progress of the new Molecular Sciences Institute (MSI) in Berkeley, California. As its associate director, Brent is helping to lead a pack of brainy academic adventurers whom he hopes will shape the way biology is done in the 21st century. Although additional reactants - including scientists and staff - are still being added to the mix, Brent is raring to turn up the flame and see what the Institute can produce.
Part of the Institute's job, made more dynamic by its lack of an official
university affiliation, will be to tackle interesting biological problems
that might lie outside the interests of mainstream biologists, says MSI's
director Sydney Brenner, a geneticist who started people working on a little-known nematode called C. elegans back in the 1960s. He believes that scientific enterprise has become too conservative. "Everything now is done by formula: You buy a kit, read the recipe, and bake your cake," says Brenner. "We want to do innovative things - fringe biology." It's exciting and risky, he adds, "but what the hell: If you can't take risks, we don't want you."
To finance this intellectual odyssey, Brenner accepted a $10 million
"one-shot, no-strings-attached" gift from tobacco giant Philip Morris. With that seed money, Brenner established the independent, not-for-profit Molecular Sciences Institute, which sits atop a candy shop across from the BART station that services the University of California at Berkeley campus.
The Genome Crusade
To Brenner, MSI's mission is clear. "We want to learn how genes control the making of an organism - how we go from genes to molecules and from molecular function to physiology."
The task falls under the nouveau rubric of functional genomics - the
systematic collection of information about what genes do. And part of the
challenge, says Brent, will be developing "new gizmos that can generate that information on an industrial scale when we turn the crank." Beyond volumes of DNA sequences, this might include an inventory of mRNAs expressed in a cell or a real-time measurement of the phosphorylation state of every important regulatory kinase or substrate.
To fuel this genomic revolution, the Molecular Sciences Institute is
recruiting scientists from many different fields in addition to biology:
engineering, mathematics, physics, and computer science. Engineers, for
example, will help build the gizmos. And physicists who spend countless
hours pondering the nature of time, space, and matter, says Brent, "will
bring, in addition to their IQ points, an interest in deep, basic questions."
Brenner and his colleagues are now sequencing the DNA of fugu - the Japanese puffer fish whose genome yields "more bang for your sequencing buck" because its DNA is relatively free of the "junk" sequences that usually separate eukaryotic genes. Other projects currently underway or on the MSI drawing board run the gamut from traditional gene sequencing to designing molecular computers.
What ties it all together is a focus on the big questions. "The point is to
talk a lot about the Big Picture and then select the right sort of experiments or organisms or equipment needed to get answers," says Brenner.
Molecular Affairs and the Yeast-a-tron
Larry Lok, the Institute's head programmer, turns his colleagues' concepts into usable code. "Biologists normally don't know what you can or can't do with a computer program," says Lok. "I act as a filter for their fuzzy ideas, because I know which things are doable and which are just cracked."
Brent and Lok, childhood friends from the age of two or three, both did their undergraduate work at the University of Mississippi in Hattiesburg, where their parents were faculty members. Although they went their separate ways in grad school - Lok to study math, Brent to study molecular biology - the pair continued to keep in touch.
Working with funding from National Institutes of Health and data that Brent collected at Harvard University - his stomping grounds before he moved to MSI - Lok has developed a database that researchers can use to search for interesting patterns of protein interactions.
To demonstrate how the program works, Lok grabs his mouse and draws a triangle on the screen. The vertices represent proteins and the lines
connecting them show physical interactions. After running a search for
molecular threesomes, the computer indicates that cyclin D, cyclin-dependent protein kinase 4 (cdk4), and the tumor-suppresser p21 touch one another inside cells. On its next pass, the program shows that p27, cdk4, and an uncharacterized protein dubbed bait1533 also interact.
So? "It's useful to have a nice way to access and exploit this kind of
data," says Lok.
By looking at the patterns that form between proteins that are known to
interact, perhaps researchers will uncover something new and interesting
about their functions.
In a separate project, funded by the Defense Advanced Research Projects Agency, Lok is working with Brent and Brenner on a futuristic-sounding plan to build computational devices that could run inside a cell. It all starts with a simulation: Lok plans to simulate the interactions within a simple network of five or so proteins that are either activated or inhibited by phosphorylation - a chemical reaction that cells use to transmit information. Equipped with the knowledge of how each protein will react to phosphorylation, the researchers should be able to predict how the network will behave in any situation. Mimicking this type of signaling pathway in silicon is useful, says Lok, because "it's easier to do experiments on a computer than on a bench."
Such simulations may point the way to engineering molecules that can perform logical computations inside a living cell - a biological computer that Lok jokingly refers to as the yeast-a-tron.
Although a living computer might be useful because it could reproduce
itself, Lok says "there's no way it's gonna do anything better than an
electronic computer." Of course, it probably wouldn't have to. "It's like a
dancing bear," says Lok. "It's not important how well it dances, but that it dances at all." He adds, "we're bound to learn stuff about how cells work along the way."
Gadget Man
Rob Carlson met Brenner on a train ride from Princeton to New York. "We talked about traveling in France," says Carlson, "and by the end of the ride he'd offered me a job."
It might not be the traditional way that young scientists advance their careers, but then Carlson's career is not entirely traditional. At Princeton University, he trained as a physicist interested in biological computation - how cells interpret information from their environment and use it to adapt, learn, or evolve.
To address how cells compute, Carlson builds gadgets. Using microfabrication techniques, he constructs arrays of tiny chambers that allow him to control the environment experienced by single cells. At MSI, Carlson will continue his earlier studies examining how white blood cells respond to physical deformation as they squeeze through capillaries. By watching how individual cells interact with the walls of his synthetic channels - noting whether and where they "stick" - Carlson found that white blood cells can remember how many times they've been deformed. The method, he hopes, will allow him to determine how smart single cells are. In other words, says Carlson, "What's the computational capacity of a single cell?"
Carlson would also like to use his fabrication finesse to design swimming
chambers for bacteria.
In an experimental setup he calls "E. coli in Flatland," Carlson would like
to see how the microbes might respond - might even evolve - when forced to navigate in an environment of only two dimensions. If DNA sequencing
techniques improve to the point where researchers would be able to sequence a bacterial genome in about a week, Carlson imagines that he might be able to "watch evolution occur, base by base, across the entire genome."
For now, though, Carlson is still waiting for his equipment to arrive.
"Setting up a lab from scratch is pretty slow," he comments while repairing my eyeglasses with a tiny little screwdriver. Once everything arrives, Carlson says he'll be ready to jump in and "go after the deep truths in biology." He adds quickly, "That's a Roger phrase."
Lingua Franca
Coming from so many different scientific backgrounds, the researchers at the Molecular Sciences Institute have spent a good chunk of time figuring out how to communicate. "I thought I knew how to talk to biologists," says Carlson. "But I'm finding the language barrier is pretty large." Take the word "information," for example. "It has a solid definition in computer science," says Carlson. But because "information" in a biological system can be difficult to define, much less quantify, using the term "puts biologists off." Lok agrees. "It's great to be in this heterogeneous situation: Not everybody thinks the same way," he says. "But understanding what people are talking about can be a big problem." For Lok, understanding what people mean - and what they need - is essential to his being able to generate computer code that will do what they want.
Library of Alexandria
As reams of information on the sequence and function of genes continue to flow from labs around the world, part of MSI's mandate will involve
processing this flood of information, turning the laundry lists of genes and proteins into an understanding of cellular function. "We need to get beyond bioinformatics to real computational biology," says Brent.
An analogy: Imagine what would happen if the Library of Alexandria had been buried rather than burned, he says. When its treasures are unearthed, says Brent, "the first people on the scene will be librarians. They'll work hard to catalog the information, but the trove of knowledge will not really be made valuable - the truths extracted - until individual scholars start to work on it."
In the same way, today's bioinformaticists are still working on developing databases that will organize the volumes of incoming information. "Bioinformatics is really still at the card-catalog level: It's like we've just invented the Dewey decimal system," says Brent. "We need to aim considerably higher."
Definition of Success
With such lofty goals, how will MSI's resident researchers know whether they've achieved success? First, they'll do good science. "We want to contribute to post-genomic biology by developing new experimental tactics or gathering whole new types of functional genomic data," says Brent.
"One fear is that we're a generation too early," he adds. "The whole
proposition that we're now in an era when biological truths can be arrived at by analytical and computational techniques performed on data generated systematically might be premature."
Brenner believes that the success of the Institute - which officially opened its doors in January - rests on the shoulders of the staff. "If, in 15 years, we look at the people who've left here, and they've gone out and are running labs and are important figures in American science, I'd consider that success," he says. For his part, Brenner hopes to get things at the Institute rolling, recruit enough people to maximize the momentum - build the staff from 15 to about 40 - and then retire.
"Maybe I'll take up ballet dancing," he quips with a smile.
The other measure of success, says Brent, will be educational. "We want to develop a heuristic environment where young people will be trained to think with appropriate breadth about the big problems in biology," he says. Of course, they've pretty much accomplished that already. In that sense, Brent says, "failure is impossible."
Karen Hopkin, a freelance writer and editor, received her Ph.D. in
biochemistry from the Albert Einstein College of Medicine in 1992. She is
the creator of the Studmuffins of Science Calendar.
Send us your comments and ideas for future articles.
Endlinks
Some Biologists Ask "Are Genes Everything?" - a basic article outlining the goals of scientists in the post-genome
sequencing era. From the September 2, 1997 issue of the New York Times.
HGMP-RC FUGU Project - a resource for the
Fugu sequencing project, this site contains genomic and cDNA information,
including a BLAST search feature, as well as facts about puffer fish and
protocols.
DNA as Lego - details how DNA is being turned into complex nanostructures. An HMS Beagle Op-Ed.
Yali's Eclectic Collection of Projects - describes the design of a DNA-based computer. Two clearly written descriptions of this tool are provided in formats for molecular biologists and for everyone else, each at a level of complexity to match its target audience.
Part of the Institute's job, made more dynamic by its lack of an official
university affiliation, will be to tackle interesting biological problems
that might lie outside the interests of mainstream biologists, says MSI's
director Sydney Brenner, a geneticist who started people working on a little-known nematode called C. elegans back in the 1960s. He believes that scientific enterprise has become too conservative. "Everything now is done by formula: You buy a kit, read the recipe, and bake your cake," says Brenner. "We want to do innovative things - fringe biology." It's exciting and risky, he adds, "but what the hell: If you can't take risks, we don't want you."
The task falls under the nouveau rubric of functional genomics - the
systematic collection of information about what genes do. And part of the
challenge, says Brent, will be developing "new gizmos that can generate that information on an industrial scale when we turn the crank." Beyond volumes of DNA sequences, this might include an inventory of mRNAs expressed in a cell or a real-time measurement of the phosphorylation state of every important regulatory kinase or substrate.
Brenner and his colleagues are now sequencing the DNA of fugu - the Japanese puffer fish whose genome yields "more bang for your sequencing buck" because its DNA is relatively free of the "junk" sequences that usually separate eukaryotic genes. Other projects currently underway or on the MSI drawing board run the gamut from traditional gene sequencing to designing molecular computers.
Working with funding from 
In a separate project, funded by the 
To address how cells compute, Carlson builds gadgets. Using microfabrication techniques, he constructs arrays of tiny chambers that allow him to control the environment experienced by single cells. At MSI, Carlson will continue his earlier studies examining how white blood cells respond to physical deformation as they squeeze through capillaries. By watching how individual cells interact with the walls of his synthetic channels - noting whether and where they "stick" - Carlson found that white blood cells can remember how many times they've been deformed. The method, he hopes, will allow him to determine how smart single cells are. In other words, says Carlson, "What's the computational capacity of a single cell?"
In an experimental setup he calls "E. coli in Flatland," Carlson would like
to see how the microbes might respond - might even evolve - when forced to navigate in an environment of only two dimensions. If DNA sequencing
techniques improve to the point where researchers would be able to sequence a bacterial genome in about a week, Carlson imagines that he might be able to "watch evolution occur, base by base, across the entire genome."
Coming from so many different scientific backgrounds, the researchers at the Molecular Sciences Institute have spent a good chunk of time figuring out how to communicate. "I thought I knew how to talk to biologists," says Carlson. "But I'm finding the language barrier is pretty large." Take the word "information," for example. "It has a solid definition in computer science," says Carlson. But because "information" in a biological system can be difficult to define, much less quantify, using the term "puts biologists off." Lok agrees. "It's great to be in this heterogeneous situation: Not everybody thinks the same way," he says. "But understanding what people are talking about can be a big problem." For Lok, understanding what people mean - and what they need - is essential to his being able to generate computer code that will do what they want.
As reams of information on the sequence and function of genes continue to flow from labs around the world, part of MSI's mandate will involve
processing this flood of information, turning the laundry lists of genes and proteins into an understanding of cellular function. "We need to get beyond bioinformatics to real computational biology," says Brent.
In the same way, today's bioinformaticists are still working on developing databases that will organize the volumes of incoming information. "Bioinformatics is really still at the card-catalog level: It's like we've just invented the Dewey decimal system," says Brent. "We need to aim considerably higher."
Brenner believes that the success of the Institute - which officially opened its doors in January - rests on the shoulders of the staff. "If, in 15 years, we look at the people who've left here, and they've gone out and are running labs and are important figures in American science, I'd consider that success," he says. For his part, Brenner hopes to get things at the Institute rolling, recruit enough people to maximize the momentum - build the staff from 15 to about 40 - and then retire.
