KAPALUA, Hawaii--The relationship between structure and function emerged as the key theme of the 1998 Pacific Symposium on Biocomputing, held here earlier this month, with speakers examining the capability of current computational techniques and technologies to model the complex relationships adequately. Attendees and organizers alike also commented on the unique diversity present at the meeting, which attracted nearly 200 scientists from both the computational and biological camps, with a large contingent from Pacific Rim nations.
"I felt the meeting was very, very successful," organizer Teri Klein, of the University of California San Francisco, told BioInform. "It was a real chance for the different communities to come together."
"More collaboration comes out of this meeting than from any other meeting I've ever heard of," added Larry Hunter of the US National Institutes of Health's National Library of Medicine, another member of the conference's organizing committee. He observed that attendees often follow up on contacts and collaborations originated at the symposium, and reiterated the point that the event brings together a unique mix of experts from a wide variety of disciplines, "for example, the enzymologists, who have previously been more involved in molecular modeling."
Attendee Joanne Luciano, a researcher at Harvard, also noted, "what impressed me was the diverse group of disciplines that was represented. It made for a very efficient and effective scientific meeting. And all the discussions were very open and frank. You have a lot of people with different opinions taking a very mature approach to a young science."
"It was very enlightening and energizing being exposed to all these new ideas," Luciano continued, noting that the exotic venue that is the symposium's trademark was, far from being a distraction, actually a plus, since the "relaxed setting fostered a lot of interaction and creativity."
Among the sessions that drew special interest was a panel discussion chaired by the University of California San Francisco's Patsy Babbitt on "The Relationship between Protein Structure and Function: How Have Proteins over Time Diverged in Function?" Other panelists included Klein, Russ Altman of Stanford University; John Gerlt of the University of Illinois, Urbana; Adam Godzik of the Scripps Research Institute; and Gregory Petsko of Brandeis University. "It's hard to describe function from a computational perspective," Babbitt explained. "My session was talking about the issues involved in that, such as gathering the right kind of data."
"The biological and computational communities don't talk enough. They need to become more interdisciplinary," she commented. "We didn't solve any big questions, but we came to the realization that the tools are at a certain level of maturity. We made some progress."
Babbitt, too, noted, that in general the symposium featured "a lot of people, both computer scientists and biologists, who were interested in talking to each other. There are a lot of people here from really disparate backgrounds, a lot of different philosophical backgrounds," she elaborated. "And it's a very good mix of academics and industry. The goals of the two groups are often divergent, but they're also complementary. There's been a lot of good dialogue. It's a very valuable place to make contacts and build bridges with industry."
"Next year we'll do a session that expands the role of the biologists," Babbitt told BioInform. "We'll get a bunch of enzymologists to come and maybe do a workshop that gives the computer people an idea of what we're working with."
The panel's Petsko--who is director of the Rosenstiel Basic Medical Sciences Research Center at Brandeis and a founder of the combinatorial chemistry company ArQule--was also the symposium's keynote speaker. His talk, "Structural Biology and Biocom puting: A Marriage Made in Heav en or in the Other Place?" focused on the issues involved in "getting to structure and going from there to drug development."
"The genomics people are expecting you to give them a 3-D structure. That's much harder to do than they think," Petsko observed, predicting, "Building models of homologous structures will be a major growth industry."
"Function is where the action is going to be," he continued, claiming, "Until we get a complete catalogue of folds and sequences to go with them, we need to get better at deducing function from structure." Before scientists can solve all the function/structure problems, "many experimental approaches need to be coded into computational approaches," Petsko emphasized. "There is a desperate need to get away from doing this experimentally and start doing it in a computer." Fortunately, he continued, the data needed to computationally define the problems "are rapidly being accumulated" and should now be "used to make better programs."
"The pharmaceutical industry is rapidly approaching the point where it's going to drown in data," according to Petsko. "Rules are imbedded in all these data and need to be teased out."
Next year's symposium will change islands, moving from Maui to Hawaii. Klein said she hoped that as the event gathers steam--this was its third year--it will attract more corporate support, both financially and in terms of bringing more of an industry perspective to the meeting. Sponsors this year included Zymogenetics, Pharmacia & Upjohn, and Millennium Pharmaceuticals. Proceed ings and other information about the 1998 symposium are available online at http://www.smi.stanford.edu/projects/helix/psb98