RESTON, Va.--The Second Annual Conference on Computational Genomics, held here October 31-November 3, attracted more attendees and a stronger industry presence than last year. David Searls, vice-president and director of bioinformatics for SmithKline Beecham, was conference cochair along with Anthony Kerlavage of Celera Genomics. The Institute for Genomic Research (TIGR) Science Education Foundation sponsored the event.
Among significant presentations at the conference, Searls noted, was Celera's Granger Sutton presenting the company's strategy for sequencing and incremental assembly based on the theoretical work of Gene Myers. Sutton also addressed detailed simulations and new approaches to dealing with troublesome repeats.
An address by Pavel Pevzner of the University of Southern California was also notable. Searls recalled, "He spoke on a new approach to using mass spectrometry for de novo protein sequencing, as opposed to a database search for peptides, on a more ambitious scale than previously thought possible."
Searls called "provocative" a talk by SmithKline's Ken Rice on the use of extremely large-scale (greater than one CPU-year) phylogenetic reconstructions of G-protein coupled receptors as a means of predicting associated ligands and signaling pathways. David Haussler of the University of California at Santa Cruz presented what Searls described as "a remarkable new, mathematically well-founded discriminant approach to homology detection that appears to very usefully extend generative methods such as HMM's."
Although the conference attracted only four exhibitors--Gene Logic, Paracel, Silicon Genetics, and the US National Center for Biotechnology Information--attendee interest in product demonstrations was encouraging, according to Victor Markowitz, senior vice-president of Gene Logic's Data Logic division, who said he left the event with several new customer leads. Markowitz added that he was pleased with turnout for a two-hour tutorial on database systems for genomic applications that he and other Gene Logic representatives offered. "It was well attended by academic scientists as well as industry representatives," he noted, adding, "Both at the exhibit and after the tutorial, we were contacted by a number of pharmaceutical companies and academic groups interested in our products."
Pierre Baldi, founder of bioinformatics company Net-ID was pleased with the standing-room-only crowd drawn by his tutorial, an overview of machine-learning techniques and a demonstration of the new version of his software product, HMMpro. Baldi observed that recent rapid advances in genomics are generating increased interest in the field. "New genomes are popping up all the time. There are 20 sequences now, and there will be 50 more in the next several years," he said. That sort of growth will only create more demand for experts "in a field where you compare sequences across genomes," he added. Baldi also pointed to robust sales of his new book, Bioinformatics, the Machine-Learning Approach, as evidence of an expanding profession. More than 2,000 copies have sold since it was published in February, he said. (The book is a bestseller in BioInform's online Bioinformatics Bookstore at http://www.bioinform.com.)
Another tutorial, on tools for target validation, was offered by Pangea Systems' Peter Karp and John Burke.
While its sponsorship by a nonprofit organization might suggest that the conference is more an academic than a commercial meeting, industry interest in the event has challenged that perception. As one attendee noted, "With the continual flow of people from academia to industry, the trend will be toward more attendance from industry." He pointed to conference organizers Searls, who joined SmithKline Beecham three years ago from academia, and Kerlavage, who recently left TIGR for Celera, as examples.
--Adrienne Burke