Flying the flag for bioinformatics

Australia's nascent bioinformatics industry is preparing for a big shot in the arm as the University of Queensland's Institute for Molecular Bioscience (IMB) makes final preparations for the 2003 Intelligent Systems for Molecular Biology (ISMB) conference, to be held in Brisbane from June 29 to July 3.

This year's ISMB, the 11th in a string of annual events, marks the first time the conference has been held outside of North America or Europe. Last year's event attracted some 1800 delegates to Edmonton, Canada, but the war in Iraq, the SARS outbreak, and Australia's distance from North America have driven organisers to anticipate that somewhere around 1100 to 1200 participants will make the trip to Brisbane this year.

The importance of the conference is reflected in the enthusiastic response to the call for papers, says Mark Ragan, co-chair of ISMB and the head of IMB's Division of Genomics and Computational Biology. "We expected something like 250 paper submissions, of which we could accept around 45," he explains. "We actually got 350 submissions. We were flabbergasted. The quality of the papers was breathtaking, and I can't stress how delighted we were; at the end of the day we had to [knock back] good papers."

Leaders in an emerging field

With its focus on computerised methods for biological analysis, ISMB has become a rallying point for the rapidly evolving bioinformatics industry, which has brought together biologists and specialists in computational methods to tackle mutually relevant challenges. One of the overriding goals of the conference has been to keep a balance between biologists and computational experts: the organising committee, for example, intentionally chose four keynote speakers from each side of the fence.

Those speakers include Sydney Brenner of the Salk Institute's Molecular Science Institute; David Haussler with the University of California Santa Cruz (UCSC)'s Howard Hughes Medical Institute; Yoshihide Hayashizaki, project director of RIKEN Genomic Sciences Centre's Genome Exploration Research Group; W James Kent, assistant research scientist with UCSC; IMB's director, John Mattick; David Sankoff, the Canada Research Chair in Mathematical Genomics with the University of Ottawa; Tel-Aviv University professor of computer science Ron Shamir; and Michael Waterman, professor of mathematics, computer science and biological science.

Brenner was a co-recipient of the 2002 Nobel Prize in Physiology or Medicine and is known for his work in establishing the existence of messenger RNA. Kent was the recipient of the 2003 ISCB Overton Prize, and is one of several ISMB keynote speakers involved in the Human Genome Project.

Not surprisingly, recent advances in genome analysis -- made even more pointed by the recent rush to identify the SARS genome, then use that information to understand and combat the epidemic -- will be high on the agenda at ISMB 2003. Waterman, for one, will use his keynote to present some work in classifying heterogeneous chromosomal components by breaking the chromosomes into smaller, homogeneous components that can be used as markers during analysis.

By identifying similarities in the chromosomes, this approach can also be used to help identify genomic variations and their impact on disease and other mechanisms. "It's as if we've been staring at chromosomes through microscopes for a hundred years, and now for the first time we're able to see them," Waterman says.

"The people at this conference are by and large technicians, and [bioinformatics methods] can't fail to become more important because of the volume and quantity of the data they're dealing with. There is so much of it, it's not all deterministic, and some of it has errors associated with it. To do biology in this era, it's very important to have computational and statistical training and skills. It's a long way from the wet lab."

Broad array

ISMB's roster of well-qualified speakers exemplifies what Ragan says is a particularly high standard of content at the conference generally.

The sessions have been arranged into broad categories, including phylogeny and genome rearrangements; expression arrays and networks; predicting clinical outcomes; protein clustering, alignment and patterns; transcription motifs and modules; structure and HMMs; text mining and high-throughput methods; and a baker's dozen of 15-minute short paper presentations covering the whole spectrum of interests.

ISMB's broad array of offerings highlights the many directions bioinformatics research is taking researchers: discussions centre around both new research techniques and the application of those techniques to problems both old and new.

Topics include detection and validation of single gene inversions, Bayesian gene/tree species reconciliation, phylogenetic tree derivation, microarray studies and techniques, methods for optimising antiviral combination therapies, annotation of viral and bacterial genomes, sequence alignment benchmarking, and many more.

Special interests

As well as academic presentations, ISMB will incorporate special interest group (SIG) meetings for the Bioinformatics Open Source Conference, Biopathways, Text Mining (also called BioLink), Bio-Ontologies, and Workshop on Education in Bioinformatics (WEB). The SIGs will meet throughout the first half of the week and report on their findings on July 1.

The ability to hold sub-meetings in conjunction with a major conference such as ISMB 2003 is an excellent opportunity to get together key researchers in a variety of niche areas, says Vincent Schächter, director of bioinformatics with Genoscope, France's National Consortium for Genomics Research, who co-founded the BioPathways Consortium in 2000 and is helping organise the group's SIG meeting at the Brisbane conference. This year will be the consortium's fourth ISMB appearance.

BioPathways was created in an effort to take a higher-level view on the study of biological systems by formalising methods for studying biological pathways incorporating a variety of smaller systems. This was a big change from conventional methods of computational biology, which have focused on the study of individual genes and proteins. To enable this shift in research strategy, BioPathways Consortium members have been discussing options for providing a single description language to unify the myriad but distinct languages that have been created for describing biological systems.

"The notion of working with pathways in networks is becoming more and more common," says Schächter. "But pathways data is still very sparse around the world. It is being produced increasingly fast, yet most researchers are frustrated by the lack of data. And many people are trying to develop their own [methods]; some are better than others, and some are just adapted to a specific set of problems. There is a Babel tower effect here, because there are different aims and people don't really understand one another."

Around 150 people attended last year's BioPathways Consortium SIG meeting, although Schächter concedes the triple hit of war, SARS and Australia's location make this year's turnout an unknown. Regardless, the SIG meeting -- which has been scheduled over two intensive, 12-hour days on June 27 and 28 -- will provide a focal point for the group's efforts by encouraging brainstorming amongst researchers with mutual interest in establishing pathways data standards. The meeting will also include plenary sessions on topics including regulation on a systems scale, the evolution of metabolism, and the BioPAX initiative (www.biopax.org) -- one effort at establishing a global pathways modelling standard.

Complementing the niche-focused SIG sessions are software demonstrations from industry (including sponsors like IBM, Sun Microsystems and Hewlett-Packard) and non-profit researchers. There will also be a day of intensive tutorials covering topics from data warehousing and artificial intelligence to bioethics, molecular modelling, and strategies for teaching bioinformatics.

Interest in running tutorial sessions was extremely strong, says Shoba Ranganathan, associate professor with the National University of Singapore's Department of Biochemistry, who organised ISMB's tutorial component. Ranganathan expected around 15 tutorial proposals, but was stunned to have to cull out the finalists from 49 submissions.

"There were even more coming in but they were late," she says. "We tried to cover a range of topics both introductory and advanced, but we had so many tutorials proposed that it was difficult to make a choice."

Regional identity

Ranganathan is also organising the WEB03 SIG, which she is designing to address what she sees as a poorly specified area: how to best integrate bioinformatics into the general biology curriculum. Whereas many of the ISMB sessions are hard-core technology sessions, many WEB03 attendees -- there were 101 last year, and Ranganathan expects similar numbers this year -- are educators interested in how to best equip themselves to teach bioinformatics.

"Teaching bioinformatics requires a lot of computer training," she says. "We get people from all aspects of life: administrators, deans, professors, and scientists who want to learn how to do education. We recommend they teach bioinformatics down to the undergraduate level. In the US, they have just recognised the importance of this, and have recommended the biology curriculum should have a whole lot of IT."

The tone of the WEB03 SIG meeting highlights the learning that is being done by bioinformatics researchers the world over. Although they're meeting in Brisbane to share their findings and build peer networks, attendees at ISMB carry a more fundamental role: as the world's premier conference for bioinformatics research, the agenda and outcome of ISMB will in many ways direct the future of a research area that's evolving as quickly as its participants can.

Bioinformatics' rapid pace of development adds even more weight to IMB's success in securing the conference for Brisbane, says Mattick. "Our view is that the challenge for the next generation of biological sciences is to put Humpty Dumpty back together again," he explains.

"We want to add in silico to in vitro and in vivo as part of the investigative range, and to create a framework for reassembling this information and the levels of transformation. That will become the platform for ultimately being able to accurately predict how subtle changes will affect [structures], and that becomes the platform for really advanced genomic engineering and understanding of genetic traits. This is, then, not simply an arcane exercise but the next stage of the natural evolution."

Mattick's keynote presentation will focus on what he calls the "hidden layer of eukaryotic programming and complexity" -- the control layer, involving the action of introns and non-coding RNA, that regulates the "developmental trajectories" that determine how organisms form.

He has another goal from the conference, however: hosting ISMB will immeasurably raise the profile of the recently created IMB, which this month moved into new premises and is set to take on 40 more full-time researchers and students in bioinformatics in the near future. This pace of growth will position IMB as the pre-eminent bioinformatics centre in Australia, which in turn should boost Australia's profile among Japan, Singapore and the other pillars of Asia-Pacific bioinformatics.

As a representative of the Asia-Pacific Bioinformatics Network, Ranganathan, in particular, is sensitive to the dynamics that have limited bioinformatics' formal study across the region: "we're trying to raise awareness," she says. That includes flying the flag at events like ISMB, as well as at upcoming events like the International Conference in Bioinformatics that began last year in Bangkok, moves to Langawi, Malaysia in September, and will hit New Zealand next year.

"If it's in the region, it's a lot cheaper for [universities] to bring representatives," Ranganathan says. "We hope to bring the bioinformatics to them." The same can be said for ISMB: by concentrating many of the foremost minds in bioinformatics, the conference will bring leading-edge research to a local and regional industry that's still working to secure its place in the world panoply of bioinformatics development.

David Braue is a Melbourne-based freelance writer