By Malorye Branca
May 7, 2002 | Compugen Ltd. is doing target validation, Gene Logic Inc. provides genomic information, and AxCell Biosciences Corp. is starting drug discovery.
Just don't use the word "bioinformatics" when talking about them.
A glance at the plummeting stock prices of former bioinformatics highfliers such as InforMax Inc. and LION Bioscience AG reveals part of the story of the troubled bioinformatics community. Major players have been bought for their cash value or less — Genomica by Exelixis Inc., and NetGenics by LION.
|The Demise of DoubleTwist
|Rumors that bioinformatics company DoubleTwist was closing its doors came as something of a shock, despite the evident difficulties that have been plaguing genomics and bioinformatics companies for the past few months.
Other signs of distress include the recent problems of Entigen Corp. and DoubleTwist Inc., which closed offices in the face of dwindling cash (see "The Demise of DoubleTwist"
). Meanwhile, hordes of other bioinformatics companies are abandoning the field for drug discovery.
Stung by the lack of profits and faced with little prospect for improvement, the bioinformatics industry must reinvent itself or accept diminished expectations. Venture capital interest in bioinformatics has not cooled so much as it has evaporated.
"Can anyone give one example of a successful bioinfor-matics company?" asks Jean-Francois Formela, an investment
principal at Atlas Venture. "Look at DoubleTwist. They couldn't even get a dollar for it."
Such skepticism pervades the venture capital community. "We have yet to see a stand-alone bioinformatics company succeed as a public company, and that has influence on where VCs put their money," says Ken Andersen, editor of VentureWire.
The numbers tell a sad tale. Andersen says his group tracked $4.4 billion in funding for biotechnology in 2001, with only $321 million of that going to companies focused on selling genomic software or databases. In the first quarter of this year, almost $996 million has been raised for biotech, but only $21 million of that went to bioinformatics companies. Those companies were GenMetrics Corp., Target Discovery Inc., and TissueInformatics Inc.
"VC money is going into product stories as opposed to platforms," says Robert Mashal of Boston Millennia Partners — which may explain the rush of companies declaring discovery programs.
No one disputes the tremendous value of bioinformatics as a tool, but a viable business model is proving elusive.
"If you are a Lilly or Celera and doing drug discovery, bioinformatics is crucial to your success," Mashal says. "But providing data, or helping companies to visualize data, has been hard to turn into a sustainable revenue source."
The fact that Celera Genomics Group and Incyte Genomics Inc. couldn't make it by scaling up and abandoned the field underscores the risks. "Anytime now, I bet that LION is going to bite the bullet and become a discovery company," Formela says.
Venture capitalists cite as problems the relatively small number of potential clients for high-priced products, the large number of competitors, and the fact that most pharmaceutical companies have substantial bioinformatics departments themselves.
"It's very hard to sell the expertise," says Andrew Lyall, chief operating officer (COO) and CIO of Oxford GlycoSciences (OGS) and formerly of Glaxo Wellcome (now GlaxoSmithKline). "After all, you meet people, you look at their CVs [curriculum vitaes], and you think you could just hire those people yourself and have them work for you."
Yet another threat comes from IT giants such as IBM, Sun, and Compaq, which have recently staked out the life science market — along with many other IT companies hungrily eyeing the space. These players will inevitably try to capture some of the tasks currently handled by bioinformatics companies, such as database design and application integration.
In addition, some companies focused on discovery, such as Genaissance Pharmaceuticals Inc. and OGS (through its interest in proteomics database developer Confirmant), are adding to the competition by selling their software or data. This trend may increase as companies seek revenue wherever they can find it. "Cash is always a good thing," says Jim Golden, director of bioinformatics at CuraGen Corp., which is considering marketing its bioinformatics platform.
Adapting or Evolving? Most companies have already abandoned bioinformatics as a label, if not as a business platform. Part of this is clearly a reaction to the mood on Wall Street. "They're following the fashion," Lyall says. "The analysts are saying there is nothing left in bioinformatics, so they are grabbing at straws."
But many of these companies argue that they are following a natural path of evolution. They say that genomic data has changed, their customers' needs have changed, and so they are changing, too. The critical question is: How much of this talk is posturing, and how much is a realistic response to changing needs in the marketplace?
Like Entigen and DoubleTwist, Compugen tried to sell products and services through the Web, on its LabOnWeb site. "It's understandable that people want to put us in the same bucket," says Eli Mintz, Compugen's chairman of the board. "There is a distinct reverberation from what has happened to those companies." Now, the Web is playing a much smaller part in Compugen's plans. "It is clear that in this field, the Web is not a sales tool, it's only useful for marketing," Mintz says.
But Compugen also has unique features that may help distinguish it from competitors. For one thing, it has always had a "wet" biological laboratory, and the ability to generate proprietary sequence, gene expression, and proteomic data. The company has also specialized in areas such as alternative splicing and unique microarray services. It recently forged deals with companies including DiaDexus, Gene Logic, Motorola, and Novartis. "Target validation is the issue we are focused on now," Mintz says.
Other informatics companies are also tracking the "bottleneck bubble" — a generalized term used to describe pinch-points in the drug development process, particularly those steps in which the amount of data is large but the ability to analyze it is slow. Gene Logic, for example, recently announced that it now has toxicity markers and predictive models for its ToxExpress program — a product that moves Gene Logic further into the drug discovery process, and toward the higher end of the value chain.
Demand exists for tools that truly address these bottlenecks, Golden says. "There are opportunities for technology development, as well as software, such as tools for predicting protein folding, ADME/Tox [absorption, distribution, metabolism, excretion, and toxicity], and metabolic profiling," he says. "We need better models of signaling pathways, protein-protein interactions, and structural bioinformatics. Cheminformatics and high-throughput screening are also areas of tremendous opportunity."
But given all of the competing products and other areas in which drug developers can spend money, companies will have to prove their tools can help speed drugs and diagnostics to market.
Viaken Systems Inc. began life as an application service provider (ASP), delivering applications and access to data via the Internet, including the ability to store clients' research data. The basic idea was to relieve its clients of the need to build and maintain costly enterprise IT infrastructures.
"It seemed like a good idea at the time, and prospective clients were saying they wanted it," says Rich Hamer, vice president and chief scientific officer at Viaken. The model offered convenience and efficiency, at least while people were mining public data. But the appeal ended when it came time for researchers to send their data over the Internet and beyond their hallowed firewalls.
Now, Hamer says, "We have evolved our ASP. The fundamental concept is that people want to press a button and get an application, and that hasn't gone away." In addition to providing infrastructure, integration and tools, Viaken is now seeking deals that make it a collaborator in finding targets, validating them, and converting drug leads into hits.
Hamer, who has also held key positions at Hoechst Marion Roussel (now part of Aventis), argues that the value of dry, or in silico, discovery has been established by companies such as GeneData, Tripos, and LION, which have cut deals with pharmaceutical companies. LION's groundbreaking target discovery deal with Bayer Corp. in 1999 is probably the most extensive and dramatic example of bioinformatics-pharma collaboration.
"We can facilitate the research process. We bring knowledge, resources, and technology to bear that they don't have," Hamer says.
Many bioinformatics companies, including AxCell BioSciences, and BioDiscovery Inc., are seeking to add collaborations in which there is profit sharing as well as service fees. Most still offer products and services, but they are also angling for bigger deals. "We began finding interesting things in terms of mining our own data," says John Rodwell, president and chief technical officer of AxCell. "That presented an opportunity for us."
AnVil Informatics Inc., however, does not sell its proprietary software, which includes unique tools for data visualization. AnVil only offers services and collaborations. Like BioDiscovery and others, AnVil's aim is to leverage the growth in high-dimensional data within genomic discovery against the lack of qualified statisticians and mathematicians to draw meaningful results from that data.
In silico discovery, or the use of bioinformatics to mine data to validate targets and optimize drug leads, is one of the most promising trends in the sector. It is fueled primarily by the tremendous growth in the number of novel and poorly understood targets that genomics has delivered. Pharmaceutical companies are under pressure to quickly turn targets into drugs to refill thinning pipelines. Many researchers also express a gnawing fear that some of the targets they haven't had time to scrutinize might indeed turn out to be valuable, and the profits from these could eventually go elsewhere.
"In every case that we've gone in and looked at someone's data, we've found things they missed," says Michael McManus, executive vice president and chief operating officer of AnVil.
Other changes are also occurring that offer opportunities to bioinformatics companies. "Outsourcing is certainly happening," says Alan Williamson, a biotechnology consultant and former vice president of basic research at Merck & Co. "There is a lot of outsourcing of medicinal chemistry today, for instance, which would have never happened 10 years ago. That used to be the crown jewel that you would never outsource."
The two highest-profile players left in the bioinformatics field — InforMax and LION — say that they see little reason to change their strategies.
InforMax is betting on having a large, established customer base, $60 million in the bank, and the expertise to develop the software researchers want.
"We have over 31,000 scientists using our software, and we have sold to 23 of the 25 largest pharmaceutical companies," says John Green, InforMax's chief financial officer and COO. But some parts of the market have been more welcoming than others. "Large biotechs and pharmaceutical companies still have not found enough utility for our products," he says, "but as our products get more sophisticated, and as we drive down the cost of ownership and make the architecture more flexible and friendly, we expect to see the market expand."
LION concurs. "There is a lot of talk," says Rudy Potenzone, president and CEO of the company's U.S. subsidiary, "but we haven't seen any lack of interest or downturn in the market."
Besides having sufficient money in the bank and one of the industry's most popular products in its SRS data integration platform, LION has been steadily working to grow the proportion of revenue derived from services versus software sales. That revenue split is now about 50-50, Potenzone says, although he concedes much of the service revenue comes from the Bayer deal.
Potenzone makes it clear that LION is more than a software company. "We don't see ourselves as selling boxes of software. We believe the fact that we have expertise in drug discovery offers something more [to clients]."
Landing another big deal would be an important step for LION, but big deals don't come easy. "These discussions take a very long time; we've had some that have gone on for more than 18 months," says Potenzone, echoing a common frustration among biotechs dealing with large pharmaceutical companies. "It's tough trying to figure out the right relationship."
New deals capitalizing on the bioinformatics exodus are likely to be plentiful, however. At the start of April, LION announced an arrangement with Incyte to provide its SRS platform to Incyte's customers, which include most of the top 20 pharmaceutical companies in the world.
Deciding Factors Will climbing up the drug discovery value chain and forging collaborations save bioinformatics?
New opportunities are clearly emerging as genomics matures, and there is interest, even from VCs, in software tools that target downstream drug development activities, such as ADME/Tox prediction and clinical trial management. But the cost of developing these tools will remain high, and staffing will continue to be difficult.
Collaborations offer another set of challenges. As anyone who has knocked on the doors of Big Pharma knows, they are not easily opened. "The plans for collaborations were always there," Atlas Venture's Formela says. "Companies haven't been able to get them."
Bayer is a rare example of a large pharmaceutical company that needed a lot of targets in a hurry, and so it cut many big deals. It's difficult to imagine that scenario being repeated on a regular basis, despite the belief that large pharmaceutical companies do indeed need these services. "The people in charge of bioinformatics at pharmas are the gatekeepers of outsourcing," Williamson says. "They have to realize they need to outsource and have the competence to do it, or senior management has to realize there is a [bioinformatics capacity] shortfall and act."
That's not all. If striking a deal is hard, striking a deal that includes royalty revenue or investment is even more difficult. "If companies are looking for equity positions, that will be a really uncomfortable position for large pharmas," says one Big Pharma insider, speaking on condition of anonymity. "I'm not saying they won't do it, but it will require more attention from management, and many of them still need to be educated on the importance of informatics."
One hope is that doing business with Big Pharma won't be necessary for success, at least in the short term. If enough money flows into genomics discovery companies and biotechs, they may provide a decent market for bioinformatics companies. "I will be the first to say that our initial targets are not the large pharmaceutical companies," Hamer says. "There is far greater innovation in small and midsize companies, and we are working on making higher-value deals with them."
Such deals are unlikely to look much like the $100 million-plus deals that went, with some regularity, to genomic discovery companies like Millennium, Vertex, and CuraGen. However, if the theory holds true and such deals represent ideal unions of innovation and expertise, they could become an important factor in the overall pharmaceutical market. It is also important to note that most of today's successful genomics discovery
|CuraGen's Quest for Real-Time Results
|CuraGen has more than $800 million in business booked, $500 million in the bank, 25 active development projects, and plans to file two investigational new drug (IND) applications this year. Now, the New Haven, Conn.-based company must show it can go the distance.
companies, including the three just mentioned, made bioinformatics a major aspect of their platforms from the beginning (see "CuraGen's Quest for Real-Time Results"
Bioinformatics is also playing an increasingly important role in next-wave platform companies, such as those doing systems biology. Tool companies are also eyeing bioinformatics players as possible partners or acquisitions as the flood of complex data resulting from high-throughput technologies grows. Without the right bioinformatics, these tools will not give useful results.
As Compugen's Mintz points out, "Bioinformatics should always be there, in the background like math is to physics. Physicists don't say, 'Look, here is the new science of physicomathematics.' We should recognize bioinformatics as an integral part of biology today."