National Editor

ATLANTIC CITY, N.J. - Rounding off with more talk about moderating expectations for the sector (the phrase "reality check" came up often during the three-day event), the World Genomics Symposium & Exhibition closed Friday after a panel of experts spoke directly with attendees.

The panel used a "town meeting" forum to deal with the usual questions about how biotechnology firms, and those struggling in the field of genomics particularly, can get their ideas to would-be deal makers in the larger companies.

But it also included more arcane debates regarding statistical analysis of microarray findings and a caution about going further than the technology may allow, into such areas as toxicogenomics.

There's not an across-the-board wariness about genomics technologies among pharmaceutical firms, said Joseph DeGeorge, vice president of preclinical safety evaluation for Novartis Pharmaceuticals Corp., of East Hanover, N.J.,

"I don't think there's any wall up that says, Don't bring your ideas,'" he told an audience member, but talking with the right person in the right department of the pharmaceutical firm is one of the genomics company's more important matters in such an approach.

And, he added, "Don't bring us bad ideas."

Sandra Glucksmann, senior director of target advancement for Millennium Pharmaceuticals Inc., of Cambridge, Mass. - which has gotten big enough to make its own sizeable deals with other biotechnology firms - said that "if I were a small company right now, it would not be sufficient just to get in the door. You want to make sure there's a buy-in to implement and work with you on expanding that technology across the company, not just in some niche. You have to think long term. Is it just getting the money today, or making [yours] the standard operating technique that everybody's going to use?"

The "ballgame" is different today, she said, with higher demands. "In biotech, even large biotechs like Millennium, we're putting our efforts in clinical trials, since that's what the market is demanding from us," Glucksmann said. "Everybody's asking, Where are your genomics targets?'"

Still, there are collaborative opportunities if the genomics company has something broadly useful, and doesn't try to oversell a technology that might not be - unless you "go to maybe an old-fashioned group that has been doing preclinical the same old way, maybe they don't even understand the difference between DNA and RNA," she said, adding that Millennium has "hired some of those from large pharma."

Mark Cockett, director of functional genomics for New York-based Bristol Myers Squibb Co., warned genomics firms against catching the ear of someone high up in the pharmaceutical company and finding themselves in positions that turn out to be less than viable.

"It's often very difficult to find a champion who's going to make [the deal] happen," he said. "A lot of big deals are top-down deals," in which an executive sold on the technology tells researchers, "You're going to have this whether you like it or not."

Cockett said that "in general, they don't work, and they're not successful. There might be a big splash with bigger dollars at the outset, but I would argue that for a biotech company, that could cause so many downstream problems" that it's not cost-effective for either side.

As the genomic data pile in, validating RNA expression data remains a challenge for everybody in the industry, noted Ronald Salerno, director of worldwide regulatory affairs for Wyeth Research, a division of Wyeth, of Madison, N.J.

Researchers bring the issue up often, he added, but "the way to solve the problem [has been] to say, We'll discuss it next time.'"

DeGeorge said much of the trouble relates to biology and not testing methods. "Even if we have inbred strains, we don't get the same results" from sample to sample, he said.

Given such imprecision, he added, establishing a discipline such as toxicogenomics - using genomics data to determine the likely toxicity of a compound in certain patients - is particularly hazardous, DeGeorge added.

"The toxicity of one site is the pharmacology of another," he said. "I'm really worried about this field. Using that kind of label on [a compound's] function is a real problem for the immediate future," especially given the way regulators may view such prospective drugs.

Salerno noted that "if you're a regulatory reviewer, you're paid to be conservative." Because companies are nervous about what such toxicogenomics tests may show - and because they realize the results do not make for an altogether reliable indication of bad effects - the data are "being generated in almost a secret lab," he said.

The FDA is said to be working on some kind of "safe harbor" provision to handle the tricky subject, but "no one knows what that term means" in the agency's view.

"Over the next few years and months, we're going to see what the FDA intends to do with safe harbor," Salerno said.

The symposium and exhibition drew about 500 pre-registered attendees, and about 1,500 took part either with booths or as visitors, said John Golicz, CEO of Unicomm LLC, of Milford, Conn., which organized the event.