The baroque Grand Ballroom of New York’s Plaza hotel served as the backdrop last week for a re-evaluation of the role of bioinformatics within an economic and biological landscape that is proving to be much more complex than many had envisioned previously. With far more uncertainty and a bit less optimism than last year’s conference at the same venue, speakers and attendees of Scientific American’s BioSilico 2002 questioned the current capabilities of in silico biology to meet the demands of biological research and attempted to assess its future within the broader health care landscape.
The opening keynote by former NCI director Richard Klausner set the mood for the remainder of the meeting by addressing the limitations of bioinformatics, rather than the promise of the technology that often characterizes such talks. While noting that information technology will indeed be a key enabling tool for biological research, Klausner pointed out that the “underside of interest in IT” is the complexity of genome-based biology, which “represents a much larger challenge than is often talked about.”
A key proponent of the use of microarrays and other genomic-based approaches in cancer research while at the NCI, Klausner, who now serves as special advisor to the presidents for counterterrorism at the National Academies and liaison to the White House Office of Science and Technology Policy, said that the biological knowledge gained from gene expression patterns has so far been “dissatisfying.” In the absence of further annotation or context-based data, clustering and pattern-recognition techniques have only “replaced microscopic phenomenology with molecular phenomenology” and have not provided sufficient biological knowledge or, more importantly, a better understanding of the relationship between a drug and a target or disease.
Additionally, Klausner called for a re-evaluation of the industry’s nomenclature for “validated targets,” a phrase that he said is “thrown around way too much in this community.”
Said Klausner, “We talk a lot about validated targets, but the reality is that it’s extremely difficult to validate a target short of clinical trials.” Instead, he proposed the alternative term “credentialed targets” to describe “levels of evidence” for the target’s validity in the pre-clinical stage.
Show Me the Drugs, Show Me the Money
Over the course of the two-day conference, speakers from pharma, biotech, genomics, bioinformatics, and IT firms discussed their varying strategies to address some of the issues Klausner raised. While hopes for the power of IT to answer many of biology’s questions remain high, it’s clear that doubts linger for some.
One recurring theme was the bottleneck in the drug discovery pipeline, which has only shifted downstream from the glut of targets identified through genomic techniques. Another was the devaluation of genomics and bioinformatics companies in the capital markets. This perceived “failure” to either generate drugs faster or make anyone rich quickly led to a number of pointed questions from conference attendees: “How are we going to sustain the investment in genomics and bioinformatics?” asked one participant. “Have any drugs or targets been discovered using genomics or proteomics yet?” queried another. Perhaps echoing the doubts of many in the room, one attendee asked simply, “How are you going to make any money at this?”
But as most participants agreed, practitioners in the field have always known that the drug development cycle time would not decrease dramatically as a direct result of genomics. “People are just realizing now that it’s not access to the data, it’s moving it down the pipeline,” said Brian Moldover, US head of bioinformatics at Aventis. “The bolus of data [from the Human Genome Project] needs to be picked over for the next few years,” he added. Sudhir Sahasrabudhe, CSO of Myriad Proteomics, agreed that it’s far too soon for anyone to see a payoff from genomic tools or data. “The drug development cycle is nine to eleven years,” he noted, “and most of the technologies we’re talking about here have been around for far less time than that.”
Some companies proved that they are rising above the doubt and seeing a light at the end of the tunnel: Gene Logic discussed how it’s building a successful business out of its gene expression database offerings; Genaissance Pharmaceuticals said it expects to begin re-marketing off-patent drugs for haplotype-specific populations by 2005; and Quest Diagnostics is partnering with Roche Genetics to bring genomic-based diagnostic tests to market in three years.
All agreed that a better dialogue with the public would be necessary to articulate the amount of work that remains to be done before the promise of genomics becomes a reality. Public perception of the risks and benefits of pharmacogenomics, government spending on new research and technology development, and continued support from the investment community all hang on the clear message that while in silico techniques are only just beginning to tackle the inherent complexity of biology, they remain the only way that complexity may ever be understood at all.