HONOLULU--The fifth Pacific Symposium on Biocomputing to be held here January 5-9 will feature discussions of some technologies that have not yet been addressed by other bioinformatics meetings, according to conference organizer Larry Hunter. Hunter, who is the section chief for molecular statistics and bioinformatics at the US National Cancer Institute, told BioInform the meeting will introduce attendees not just to new approaches to old problems, but to novel areas of bioinformatics research, such as genetic networks analysis and natural language processing. At BioInform press-time, conference organizers said about 210 attendees were expected.
Wah Chiu, the Alvin Romansky chair in biochemistry at the Baylor College of Medicine, will give the keynote lecture entitled, "Protein Fold Determination of Large Macromolecular Assembly from Electron Cryomicroscopy and Sequence Threading."
As many as nine papers will be presented during each of 10 sessions covering the topics:
• Protein evolution and structural genomics;
• Protein structure prediction in biology and medicine;
• Molecules to maps: tools for visualization and interaction in computational biology;
• Molecular network modeling and data analysis;
• Data mining and knowledge discovery in molecular databases;
• Identification of coordinated gene expression and regulatory sequences;
• Natural language processing for biology;
• Computer-aided combinatorial chemistry and cheminformatics;
• Applications of information theory in biology;
• Human genome variation: analysis, management, and application of SNP data.
The technology of natural language processing has matured to the point that for the first time, Hunter said, the Message Understanding Competition run by the US Defense Advanced Research Projects Agency to test and compare information extraction software can be applied to biology. Hunter and colleagues will present their technology, EDGAR (short for extraction of drugs, genes, and relations), which extracts facts from Medline data with 80 percent accuracy by looking for keywords in every abstract.
Hunter said the meeting is driven by the interests of the scientific community more than most conferences by virtue of the fact that session topics are selected from a list of ideas to which anyone in the field can contribute. Open 90-minute roundtables following each session are also notable, Hunter said, for engendering discussion about new research. He added that discussions on protein structure and the first big conference session on biocomputing methods for single-nucleotide-polymorphism discovery would also be interesting.
Peer-reviewed proceedings and more information about the meeting, which will be preceded by a day of tutorials, are available online at http://www.cgl.ucsf.edu/psb/. Meeting sponsors are Compaq, the International Society for Computational Biology, Partek, PE Biosystems, SGI, Stanford Medical Informatics, the US Department of Energy, and the University of California, San Francisco's Computer Graphics Lab.