Borrowing from the characteristically ambiguous terminology of the National Weather Service, the forecast for the commercial bioinformatics market in 2004 is either “partly cloudy” or “partly sunny,” depending on whom you speak to.
The glass-half-full/half-empty outlook may not be wildly optimistic, but it does indicate that the sector may be on the path to recovery, following a year that was an economic low point for many observers in terms of dwindling financing opportunities, increased biopharma belt-tightening, and rampant consolidation [BioInform 12-22-03].
One sign of hope for the bioinformatics market comes in the form of the technology “hype cycle” compiled by market research firm Gartner. As 2003 dawned, bioinformatics found itself at the bottom of the aptly named “trough of disillusionment,” where failure to meet expectations pushes a technology out of fashion. A year later, however, the field has approached the much more hopeful-sounding “slope of enlightenment,” according to Gartner analyst Carol Rozwell. “We are seeing a revival of interest from many of our clients, and it seems like they are out of the trough,” she said. “They’ve gone through the expectation that bioinformatics is the greatest thing since sliced bread, and then in the trough they say, ‘Oh gosh, this is the worst decision we ever made,’ and now they’re actually starting to be more practiced in the approaches that they’re using; they have more pragmatic expectations of the technology; and they’re beginning to fit it into an overall technology infrastructure.”
For bioinformatics vendors, this mild infusion of enthusiasm is a welcome shot in the arm, but it doesn’t quite signal a boom market for the year ahead. “It’s not going to be a hockey stick,” Rozwell cautioned. “It will be a lot of hard work in 2004, but there is definitely improvement.”
While the capital markets are forecast to improve slightly during 2004, loosening the purse strings of biotech investors, it will still be a challenge for many bioinformatics startups to attract financing. “I think the market in general has been at a pause as they wait to see how the whole thing might sort out,” said Arthur Pappas, CEO of venture firm AM Pappas and Associates. “There’s still a tremendous need on the part of large pharma and biotech for the use of bioinformatics platforms,” he said, but warned that the year ahead is “still going to be tough for the bioinformatics area.” Pappas said that he’s still “keeping his eyes open” for promising bioinformatics ventures, however, and is looking forward to the upcoming JP Morgan Healthcare conference in mid-January to scout out some prospects.
Consolidation, another symbol of the rough economic climate in 2003, is also likely to continue in the bioinformatics sector in 2004, but there is a silver lining, according to Rozwell: “Even though there still are a lot of vendors — and we do still expect that there will be some market consolidation — we are seeing more stability in the offerings,” she said. “It seems like many of the firms that weren’t exactly sure what there business model was two years ago have become more secure; they have access to more funding … and many of them have set themselves a course moving forward where it looks like they will be able to have more productible returns.”
Additionally, the spending patterns of biotech and pharmaceutical companies can be expected to mirror those of 2003, when resources that were previously directed toward early-stage discovery were shifted further downstream toward development. While this trend toward a shorter-term payoff caught many informatics firms by surprise in 2003, it forced some to make the necessary adjustments to give pharma what it wants in 2004: systems that will work along the entire drug discovery continuum; an accurate breakdown of expected return on investment, and effective solutions to previously intractable problems. Those companies who meet these requirements will have the best chance of staying ahead of the rest of the bioinformatics pack in 2004.
Securing a Place in the Pipeline
During 2003, it became clear that biotech and pharmaceutical companies do not want to rely on a single vendor to provide their underlying informatics infrastructures, and instead prefer to stick to home-grown, component-based systems that allow them to plug best-of-breed point solutions into a larger, flexible framework. With integrated systems in — but monolithic platforms out — vendors must gain a better awareness of where their offering fits into the larger context of their customers’ informatics environments to succeed in the coming year.
For those seeking to expand their business in the coming year, the challenge will be to coax growth out of niche solutions. Bioinformatics startups today “recognize that they can’t build a business on just satisfying these individual [research] silos, so we’re seeing a better understanding of what’s required to build an end-to-end solution,” said Pappas. However, he noted, “if the bioinformatics play is to provide low-cost efficiency [into] the drug discovery process, then having a large infrastructure that has a heavy cost component to it just is not going to be favorable.” The trick, he said, “is finding the right way to fit the newer technologies that these smaller companies may be able to develop into what large pharma is building for their large legacy systems.”
Todd Smith, CEO of Geospiza, suggested that the success of the sector may depend on cutting the cost of bioinformatics solutions. “People really get tired of paying a million dollars year in and year out, to only solve half their problems,” he said. A price point “moving down off of six figures into three or four figures is actually supportable” for many systems today, he said, citing first-generation desktop solutions as an example. However, “what they lack is interoperability,” he noted, “so people are going to want interoperable solutions, but at desktop prices.”
At present, some larger informatics vendors are pursuing broad strategies based on this model, but the jury is still out on how well they will succeed. Accelrys launched its Discovery Studio line of integrated desktop solutions in 2002, while InforMax has abandoned its GenoMax enterprise-scale system in favor of a suite of desktop tools under the Vector brand, and Lion introduced its Target Engine component-based platform in June of 2003. The industry will certainly be keeping an eye on the performance of all of these firms in the coming year, particularly in light of InforMax’s continued integration into Invitrogen, and Accelrys’ upcoming launch as a standalone informatics firm following the spin-off of parent company Pharmacopeia’s drug discovery business, which is expected to close in the first quarter of 2004. Following the resignation of Lion CEO Friedrich von Bohlen in late December [BioInform 12-22-03], many in the industry also questioned whether Lion will be sold during 2004. With its landmark $125 million collaboration with Bayer due to expire in June, Lion’s future may hang on a single renewal agreement.
New (and Not so New) Technologies To Watch
During 2003, new technology areas moved into the spotlight, including those related to ontologies, pathways, text mining, and others dwelling under the broad “systems biology” umbrella (see below). These areas are likely to advance during 2004, although questions remain as to whether they will deliver on their promise. For Rainer Fuchs, VP of research informatics at Biogen-IDEC, several of these new technologies were “nice to have” in 2003, “but none of them struck me as an absolute must have.” Fuchs cited text mining as “one of the most exciting technologies” for future research, but deemed current efforts in the area “not quite there yet.” He added that he has “little hope” in using pathways to identify novel targets, but sees great potential for pathway databases and analysis tools, as well as other technology areas, in the realm of biomarker identification. “That’s where technology that has gotten a bit of a bad rap, like transcript profiling and proteomics, will actually be really, really useful,” he said, “So that’s where I really believe there’s going to be — from an informatics perspective — pretty fertile ground.”
For Fuchs and others, the technology focus for 2004 may not be so much on the bleeding edge of bioinformatics, but rather on some “not so sexy” informatics technology areas, like LIMS, sequence analysis, and data integration, where there is still plenty of room for improvement. “Incredible as it seems, there’s still no convincing [LIMS] solution that you can buy anywhere,” Fuchs said. “It’s not the space where you need to develop a new algorithm or any completely new technology, but it’s the fine line between having a theoretical solution and one that actually works in practice.”
For Geospiza’s Smith, even the old workhorse of sequence analysis software will see advances in 2004, particularly as next-generation sequencing technologies from firms such as 454, Visigen, US Genomics, and others come online. Smith said that a year or two ago, “I would get asked the question, ‘Sequencing seems over. Shouldn’t you be doing arrays now?’” But those questions are rarer now, he added, “because the field fundamentally realizes that having a DNA sequence is incredibly valuable” for resequencing, population-based genomics studies, and pharmacogenomics.
New technologies to address data integration — the perennial nuisance of informatics groups — will also appear in 2004, but the days of tackling this challenge through technology alone may be drawing to a close, according to Fuchs. “A few years ago I would have argued that integration technology is a real gap for us. These days, I don’t see that any more.” Today, he noted, “I don’t see it so much as a technology [gap], but more of a cultural hurdle, and organizational boundaries that have to be broken down.” With technologies like web services and XML, Fuchs said, “We’ve been able to get integration done quite effectively” — once the cultural hurdles and business process hurdles are addressed, he said.
One technology trend that many observers are tracking is whether open source software licensing models will move beyond the academic and government sectors and into the commercial bioinformatics market. Certainly, 3rd Millennium’s decision to release its software under the GPL in December [BioInform 12-15-03], served as a sign that the open source model is gaining acceptance among commercial vendors, but whether this practice will become more widespread in 2004 is still unclear.
In 2004, observers will also be keeping an eye on standardization efforts in the life science informatics world. While the lack of standards in the sector has long been cited as a barrier to progress, organized efforts to introduce and maintain standards have historically been met with indifference, disdain, or outright hostility. As 2003 came to a close, I3C co-founder Sun Microsystems withdrew from that effort, while David Benton, a co-chair of the OMG’s Life Science Research Domain Task Force, resigned. Meanwhile, community-driven efforts, like BioPAX for pathway data, and PSI (protein standards initiative) for proteomics data, appeared to gather momentum.
“The challenge for the pharmaceutical and biotech industry is that we’re getting even more short-term focused, and standards are not the things that are going to pay a return on investment in six months from the time we get working on it,” said Benton. “The community recognizes that standards are hard, that the science is moving very fast, so it’s hard for standards to keep up.”
According to Fuchs, who stepped down from his position on the I3C’s scientific advisory board at the end of 2003, there are two ways that standards can have an impact on life science informatics in 2004: “Either there’s a strong momentum out of the academic community behind them, and I’m not sure that I’ve seen that, or there’s so much pressure by customers on potential vendors that vendors are forced to adopt it.”
Another challenge for the informatics sector to address in 2004 is the widening chasm between bioinformatics tools and biologists’ capabilities. While many tool developers take the genomic age for granted, most biologists are “still coming to grips with how to do production-based science,” Smith said. The result is that even though the number of available software tools is growing, the potential user base isn’t. “If you talk to people at the genome centers who are at the first level of reducing the technology to practice, they’re way ahead of where most of the scientists really are that want to use it,” Smith said. “But if you go to ASHG, and you’re on the floor and you talk to the scientists that are coming by the booth, they still don’t know what’s available for analyzing DNA sequences.” While most bioinformatics training programs currently focus on churning out people who can develop still more software, Smith said the real challenge — and opportunity for the sector — may lie in training biologists to use the tools they already have access to.
Outlook and Advice
The overall bioinformatics outlook from most observers for 2004 is optimism tempered with a degree of caution. Although Pappas called the bioinformatics market “an exciting space,” he added that “we haven’t seen anything that’s jumped out at us at this point” in terms of new investment opportunities. One of the toughest challenges for bioinformatics companies in the year ahead, Pappas suggested, will be getting “validation within pharma at a higher level.” Pappas’ advice for young companies in 2004 is to stay small, lean, and efficient, and “not add people until they get bigger contracts, not build out their systems support until they have the fundamentals of their technology in place.”
Gartner’s Rozwell noted that a solid ROI analysis will continue to be crucial for companies looking to sell their products to pharma, “but what we also would recommend going forward in 2004 is that it not just be an ROI for acquisition, but a look at the total cost of ownership.” The TCO includes many “soft costs” that are sometimes left out of the analysis, she said, including ongoing system maintenance expenses and the challenges of getting people trained. Fuchs added that ROI is easier to prove the closer you get to the clinic. “A lot of the informatics investments right now are driven by a desire to look at those areas that are the most costly, and the further you move toward the clinic, the more costly it gets. So anything you can do to have even a small impact in those areas will almost pay for itself,” he noted.
In addition, Rozwell said, firms who maintain solid relationships with their customers in order to ensure that they are getting the most value out of their products will help drive the entire field forward in 2004. The bioinformatics sector can learn a valuable lesson from the software industry, she said, where the well-established firms “have very sophisticated processes for staying in touch with their customers — they have user groups, customer advisory boards, really sophisticated support centers. Many of these younger [bioinformatics] firms just didn’t have the time to put those processes in place,” she said, “so the ones who are learning how to operate in the industry and use those techniques to constantly improve the product, to constantly help the users with implementation and using more features, those are the ones who will be successful.”