There comes a point in the life cycle of every organism when it must change or perish. For bioinformatics, the time for metamorphosis is now. Though computational biology is already an intrinsic part of the drug discovery process, the business models adopted by most bioinformatics firms have failed to produce profits. Competition -- from the IT industry and big pharma itself -- is growing and investors, both public and private, are unimpressed. While some companies are hoping persistence pays off, many are pursuing new business models that should allow them to retain a bigger share of the profits they are helping to create.



Ever since Norman and Leigh Anderson began talking about a human protein index in the late 1970s, the life sciences have been on a collision course with information technology. The inevitable impact occurred in the early 1990s with the advent of the Human Genome Project; the ensuing fusion reaction transformed biotechnology into an information science and touched off an explosion of new data that has since grown far beyond the ability of any one scientist, or team of scientists, to comprehend.

The efforts of genomics have yielded thousands of genes and millions of single nucleotide polymorphisms (SNPs), not to mention the millions of potential proteins coded by those genes. The proteomics revolution has revealed that the fundamental problems of biology are far more complicated than anyone imagined. And that has made bioinformatics a central theme of the post-genomics era.

But it has not made many millionaires.

Despite a tremendous amount of hype, there have been few success stories in the bioinformatics sector. Profits remain an elusive goal for most companies and many are finding it increasingly difficult to secure follow-on funding in today’s anemic economy.

“Right now, in the marketplace, there’s really only two things worse than being a genomics company: being a drug discovery company and – even worse than that – a bioinformatics company,” said Kurt von Emster, a general partner with MPM Capital and manager of MPM’s $300 million BioEquities Fund.

While some bioinformatics companies are still staying their original course others are looking for new pathways to profitability. Some have been gobbled up by big pharma, while others are pursuing the sorts of deals that pave the way to forward integration. Meanwhile, a growing number of information technology (IT) companies are expanding into the bioinformatics space in an effort to tap into what they believe will soon be one of the hottest markets around. All of these changes have made it nearly impossible to define bioinformatics as a business, but one thing has become clear: Bioinformatics, as a sector, is evolving – and not a moment too soon.


“The biggest problem we see with bioinformatics companies is that the economics, traditionally, of their businesses are somewhat unclear,” said Chris Ehrlich, a senior associate with InterWest Partners. “There’s been a limited demonstration, to date, of these companies’ ability to generate a profit.”

Speaking at the Health in the Digital Age 2001 Summit, hosted by Stanford University's School of Medicine and U.C. Berkeley's School of Public Health in September, Ehrlich said the heart of the problem lies in the fundamentally flawed business models adopted by many of the early players in the bioinformatics arena.

Until recently, most of those models revolved around the idea of selling software and consulting services. In the IT industry, such companies are called independent software vendors, or ISVs. The model works great if you are, say, Oracle Corp. and you're peddling enterprise software to Fortune 500 companies. The economics make a lot less sense when your target market consists of a few dozen biopharmaceutical companies. Software offers a scalable, but not dependable, revenue stream. The added consulting services are more profitable than software, but far less scalable.

“You have a lack of scale in the higher revenue-generating component, and an unpredictable revenue model at the other end,” Ehrlich explained.

Other first-generation bioinformatics companies took a different page out of the software industry’s playbook and fashioned themselves as application service providers, or ASPs, offering access to biological databases and software tools via the Internet. Ehrlich said their pharmaceutical customers have been generally unimpressed by these efforts, citing both performance issues and privacy concerns.

Nevertheless, Ehrlich thinks the sector is starting to mature. Some companies, such as Lion bioscience AG and CuraGen Corp., are abandoning the old models in favor of fresh paradigms that add in-house drug discovery or diagnostic components to their core bioinformatics business strategy. These research programs could yield product candidates that could be partnered out, opening the door to potential milestone and royalty payments. Venture capitalists like these models, because they allow bioinformatics companies to demand a bigger piece of the pie they are helping to create.

Ehrlich’s optimism about these new approaches to bioinformatics is tempered by other marketplace realities.

“The customer base – although it is profitable and there is a need for bioinformatics software – is somewhat limited,” he said. “In addition to limited customers, there are relatively low barriers to entering this field and relatively significant competition.”

Based on Ehrlich’s calculations, there are now more than 100 privately held bioinformatics companies. The number of publicly traded companies depends on whom you ask, since notions of what constitutes a bioinformatics company vary widely. Add to this mix a host of IT companies and the internal IT development efforts being pursued by many pharmas and it starts getting pretty claustrophobic.

With so many different companies competing for already limited pools of cash and talent, it is hardly surprising that bioinformatics has failed to ignite the passions of both private and public investors. MPM’s von Emster, also speaking at the Stanford conference, said bioinformatics companies have done little to generate sparks.

“Investors really don’t understand bioinformatics,” he said.

In part, this is the result of sheer ignorance on the part of investors, most of whom know nothing about the bioinformatics sector. After all, said MPM Capital's von Emster, since most investors do not have the vaguest idea what goes on inside a biopharmaceutical company, it is not likely they will appreciate the needs being addressed by bioinformatics. Moreover, he said the few investors who do have a clue are usually more interested in intellectual property behind the genes themselves than the algorithms used to convert them into targets or therapeutic products.

Of course, it does not help that most bioinformatics companies do a lousy job of telling their own stories. Many are having even less success on the revenue front, creating much more serious perception problems for the sector.

“There’s a clear lack of earnings visibility, and investors in this environment will not invest in these types of companies because of that,” according to von Emster. “There’s this close association with an Internet-like model and everybody who’s been burned by an Internet company is not going to burn themselves twice.”

Then there is the issue of the technology itself. While it may be getting better all the time, much of it remains unvalidated. That makes many pharmas nervous – and understandably so, said von Emster. Nagging questions remain about the quality of the data available from public databases. If the data cloud – the collective mass of biological and chemical data – is not 100 percent accurate, then even the best technologies are bound to produce bad results, at least occasionally. As the old hacker mantra goes: garbage in, garbage out.

“Wall Street’s perspective is highly negative right now,” von Emster warned. “You’re fighting a very large uphill battle.”

The result: companies like InforMax Inc. and Genomica Corp., both of which are trading below book value.

However, Wall Street’s attitude may not be entirely justified. Even skeptical observers like von Emster recognize that there is a high demand for bioinformatics solutions – a demand fueled by the simple fact there is just no other way to transform the promises of genomics and proteomics into products.

“You do need this software and algorithm basis to do it,” von Emster said. “If you try to triangulate where bioinformatics fits, it’s really a fairly narrow window. It comes post gene expression, post gene database, post array output and post the core biological information that’s already widely known.”

The solution lies in rethinking business strategies and charting new courses to profitability.

“There is a trend towards acquisitions,” van Emster said, pointing to Merck & Co. Inc.’s acquisition of Rosetta Inpharmatics Inc. in July by way of example. “I think that’s a valid model.”

The other valid model, he said, is forward integration. Like Ehrlich, von Emster said Lion bioscience is a good example, but he also pointed to companies like Compugen Ltd. that are moving toward the wet lab and setting up their own target-discovery programs.

“Moving towards the target is the true answer,” von Emster said, adding that deal volume needs to increase in order to attract investor attention. “There are not too many (deals) being signed right now.”


Mark Edwards, managing director of Recombinant Capital and publisher of Signals, said that situation is starting to change, albeit slowly.

Though he called it “a tempest in a teapot” compared to the deal volume in the biotech industry as a whole, Edwards said bioinformatics companies are starting to make some waves. While he believes the bioinformatics sector is still at an early stage in terms of alliances and forward integration, Edwards said some companies have managed to sign some genuinely interesting deals – deals which point to an evolution of business models consistent with the trends identified by Ehrlich and von Emster.

“There has been some very positive segmentation of the field,” Edwards explained. “If you view bioinformatics from the deals and from the progress that the deals have made over the last year to year-and-a-half, there are up-ticks.”

A good example is the deal that Bethesda, MD-based InforMax signed with Amersham Biosciences in August 2000. Essentially a software and services deal, Edwards said this partnership was aimed at developing an enterprise systems approach to bioinformatics, bundling the former’s informatics software with the latter’s drug discovery platform.