LONDON--Significant challenges in data integration must be addressed and large pharmaceutical companies must reengineer their drug discovery processes around bioinformatics for the pharmaceutical industry to continue enjoying the 10 percent growth in sales revenues seen in past years. This, along with suggested solutions to the growing data dilemma, were common themes across presentations at a two-day seminar here last week. In addition, a major protein database supplier selected the venue to announce significant changes in the organization and in distribution of its core database.
Arranged by IIR Ltd. and sponsored by Silicon Graphics, "Har nessing Innovation to Develop Effective Bioinformatics Programs" brought together trendsetters from universities, national research centers, pharmaceutical companies, and bioinformatics companies to present innovative solutions to challenges facing the bioinformatics industry. The formidable list of speakers provided a comprehensive representation of those tackling large-scale problems in bioinformatics: a 50/50 split between commercial and nonprofit concerns and between protein and nucleic acid experts; and broad geographical representation, with 11 speakers from Europe, five from the US, one from Israel and one from South Africa.
Early in the conference Steve Gardner, CEO of Synomics, cited a 1997 Andersen Consulting study to show that pharmaceutical firms must increase the number of drug development candidates sixfold each year if they hope to maintain the current level of growth in sales to the year 2000. The urgency of this goal was reflected by pharmaceutical representatives who demonstrated that, although the dream of a bioinformatics-driven drug discovery process is fast becoming a reality, pharmaceutical companies' bioinformatics departments continue to be staffed by a fraction of 1 percent of the total number of staff found in research and development.
The exploding number and type of databases currently in the bioinformatics realm was a focus for both academic and commercial researchers who spend significant efforts to clean, integrate, and annotate data in a way that brings maximum value to end users. Large-scale bioinformatics technologies are shifting rapidly, out of necessity, from high-throughput data generation and collection to algorithms and methods for increasing understanding of the data, conference discussions indicated.
Integrating genomic, proteomic, and EST data is no small task, but add to this expression profiles, population genetics, patents, literature, and clinical outcomes and one begins to approximate the data load pharmaceutical companies contend with. Bioinformaticists, challenged with simultaneously increasing and integrating the amount of information they work with, are looking for standards to adopt. Visualization becomes increasingly important as complexity and visibility of the data increases, and Java is widely used for this purpose. CORBA, despite some avid supporters, was not as wholeheartedly embraced as a standard for seamless integration access to outside information. Despite these much-needed efforts, very little standardization exists in the industry.
While commercial entities focused on data integration issues, academic researchers discussed novel developments for gleaning additional knowledge from the data. These ranged from extensive cross indexing between heterogeneous data resources, to enhanced EST clustering systems, to discovery techniques based on Hidden Markov Model algorithms. Visualization remained a key component throughout, since the complex data produced by these algorithms is not easily understood in the textual outputs so popular in the past.
Amos Bairoch, widely known for his SWISS_PROT and PROSITE databases, used the conference as an opportunity to announce the founding of a new nonprofit research organization, the Swiss Institute of Bioinformatics (SIB). SIB brings together five research groups and 50 researchers from various institutions across Switzerland (see BioInform, March 16, 1998). Bairoch also announced that, as of the next release, the SWISS_PROT protein database will no longer be free to commercial companies. Additionally Geneva Bioinformatics (GeneBio), a for-profit company, has been created to commercialize and develop products based on those developed at SIB.
Conference-goers generally agreed that bioinformatics is the new driving force for pharmaceutical companies meeting intense pressure to increase the number of potential drug targets. They indicated that the definition of bioinformatics is broadening rapidly and now includes information from throughout the drug development process, causing data integration nightmares for relatively small bioinformatics staff found in most pharmaceutical companies.
One difference revealed at the meeting: While commercial bioinformaticists focus on data integration issues, leading academic researchers are devising intelligent systems for clustering, linking, characterizing, visualizing, and annotating biological information. These innovative methods are migrating to the commercial sector and are rapidly affecting and improving the understanding of human disease.