Edmonton’s history as a rough-and-tumble northern Alberta outpost made it a fitting — albeit damp and chilly — host city for this year’s Intelligent Systems for Molecular Biology conference. The tenth annual ISMB meeting, held August 3-7, was the largest bioinformatics symposium ever, according to conference organizers, but the rapid expansion of the field was evident in ways far beyond the record-breaking attendance of 1,500. The broad range of subjects covered in the plenary sessions, combined with a proliferating number of pre- and post-ISMB satellite meetings, underscored the field’s continued push into new and uncharted territory.
Several attendees noted the conspicuous absence of a sequence analysis track in the general sessions, proof that the discipline is evolving to answer a new set of biological problems. Protein analysis — motifs and families, structure and function — occupied the first full day of talks, while papers on gene expression analysis filled the entire second day. Network and pathway analysis and gene structure made up the remainder of the dedicated tracks, with two sessions devoted to the catch-all “methods” category. The University of Alberta’s David Wishart, conference chair, said ISMB reviewers had to winnow 207 submitted papers down to the 42 presented at the meeting.
Posters covered a similarly broad range of topics, with microarrays winning the popularity contest — 76 out of the meeting’s 500 posters addressed microarray-related informatics. Data mining came in a close second, with 66 posters.
Keynotes covered a wide swath of topics, from Michael Ashburner’s kickoff talk on the current state of the Gene Ontology, to Overton Prize winner David Baker’s discussion of the intricacies of protein structure prediction. Barry Honig, Isadore Rigoutsos, Ford Doolittle, Stephen Altschul, Terry Gaasterland, and John Reinitz also delivered keynotes.
Special interest group (SIG) meetings on biopathways, open source software, text mining, cellular simulation, education, and ontologies preceded and followed the main conference, drawing the full event out to a total of eight densely packed days.
But it’s very likely that this year’s meeting may be the last to follow this format. Many attendees noted that increasing specialization within the field is quickly making the plenary session obsolete. Indeed, the cavernous meeting hall tended to empty out following the keynotes, leaving the bulk of the talks only one-half to two-thirds full. In response to these concerns, the International Society for Computational Biology is contemplating a new format for future conferences, which will likely include some combination of plenary and parallel sessions. While the notion of parallel sessions was soundly rejected at last year’s ISCB annual meeting at ISMB in Copenhagen, a show of hands at this year’s meeting indicated that most are now in favor of the idea.
ISCB members expressed a number of arguments on both sides of the question. One risk of parallel sessions, according to the University of Colorado’s Larry Hunter, is the “balkanization” that has occurred in other computational disciplines. For example, in the field of artificial intelligence, Hunter noted, researchers are now so specialized that the machine learning community has no way of keeping up with new developments in natural language processing or other AI subdomains. Some attendees, however, noted that “general interest” bioinformatics topics are getting increasingly more difficult to identify, and that a few high-quality invited speakers and keynotes may suffice for community building, while parallel tracks on specialized topics would attract a larger number of quality talks. The ISCB is also looking at ways to more actively involve the SIG meetings in the general conference, perhaps through “SIG days” or other involvement in the conference to keep the general ISMB crowd up to date on emerging areas of interest.
After a decade of organizing the ISMB conferences, the ISCB’s struggle with the meeting format is further evidence that the field of bioinformatics has reached a critical point in its maturation as a scientific discipline. However, in a week that brought news of further layoffs in the commercial tools sector — Genaissance Pharmaceuticals reduced its informatics and sequencing workforce by 20 percent and Pharmacopeia was forced to let approximately 80 employees go, mostly from its Accelrys software division — the strong attendance and enthusiasm (not to mention overflowing job board) at ISMB proved that the art and science of bioinformatics is thriving, despite the field’s periodic growing pains.
Next year’s meeting will be the first ISMB held in the Southern hemisphere. The Institute for Molecular Bioscience and the Advanced Computational Modeling Center of the University of Queensland will play host, June 29 to July 3, in Brisbane, Australia.
— BT