ISMB 2001 was the first major bioinformatics meeting to be held after the publication of the two human genome draft sequences. The increasingly important role of computational biology was reflected in the 1200 strong participants list, comprised of academics, entrepeneurs, biologists, computer scientists, physicists, publishers, lawyers, and so on. The conference programme was organised around six broad themes - protein structure, sequence analysis, networks and modelling, gene structure/regulation, phylogenetics, and methods. Whilst conveying a wider picture to the more specialised researcher, this also allowed time to enjoy some of the more informal aspects of the conference such as the popular poster sessions, sponsor exhibits and the picturesque Tivoli Gardens.
The audience was eased into each new session with the aid of some enlightening keynote talks. Christopher Dobson elaborated on the link between protein misfolding and disease. David Eisenberg described new tools for identifying functional and physical interactions between proteins, such as the Database for Interacting Proteins which currently contains over 10,000 interactions between several thousand proteins in approximately one hundred organisms. Bernardo Huberman continued the networks theme by exploring the phenomenon of the world wide web and efficient ways of searching this 'ecology of knowledge'. In the memorial Overton Lecture, Christopher Burge - author of the popular gene prediction tool, GenScan - discussed the difficulty of understanding the rules for RNA splicing.
As an informatician in the fast-evolving field of microarray technology, the opportunity to attend ISMB proved vital for assimilating the latest techniques for gene expression analysis. Dana Pe'er of the Hebrew University in Israel presented a promising Bayesian network approach to indentify subnetworks of interacting genes and demonstrated its effectiveness on the Rosetta Compendium dataset. Chen-Hsiang Yeang from the Whitehead Institute tackled the problem of simultaneously classifying multiple tumour types by determining the most effective combintaion of binary classifiers, such as K nearest neighbour and support vector machines. Alexander Zien from the German National Research Centre for Information Technology described a novel normalisation method, called centralisation, which accounts for variation between arrays by rescaling the data in a biologically motivated way. Additional highlights include two informative tutorials ('Genetic network analysis' and 'Integrated analysis of gene expression data'), the International Society for Computational Biology open meeting, and the discussion of the Public Library of Science led by Michael Eisen.
The openness of the bioinformatics field was highlighted by the large turnout at BOSC and Biopathways, two of ISMB's six satellite meetings. Among the lightning talks and reports from the various bio* projects was a thought-provoking talk by Steve Brenner who proposed an open source authors contract for university, government and other workers to use and produce open souce software in the course of their employment. More focused discussion continued well into the evening in the many 'Birds of a Feather' sessions (BOFs) and a new acronym was coined in 'Talk Over Beers' (TOBs).
Conference web sites: