What’s the forecast for bioinformatics in 2003? Unfortunately, the economic climate is likely to remain as stormy as it was in 2002, according to industry observers. However, there is a silver lining within those gathering clouds: The rapid advance of science and technology will present a host of exciting new bioinformatics challenges for those with the guts (and funding) to weather the storm.
“The future of bioinformatics depends on whether you think of it as an asset or a liability,” said independent bioinformatics consultant Jim Freeman. “How much revenue would you lose if you removed the entire bioinformatics department in your company? If you can answer that question, you're in a good spot.” The defining challenge, then, for bioinformatics amid increasing biopharma belt-tightening this year will be identifying parameters that enable managers to quantify the return on investment from their existing bioinformatics installations as well as new technologies they are evaluating from vendors.
Jonathan Edwards, an analyst at Gartner, used the consulting firm’s technology “hype cycle” to describe the current state of the industry, placing bioinformatics “in the bottom of the trough of disillusionment.” The good news may be that there’s nowhere to go but up, but Edwards noted that the sector isn’t likely to ascend the more promising “slope of enlightenment” this year. “I don’t see it taking off again until companies have proved they can get a measurable benefit from these tools,” he said. “That is happening, but very slowly.”
Consolidation, Funding Drought to Continue
Additional trends on everybody’s short list for the year ahead look like 2002 reruns: Consolidation, and plenty of it, along with a bleak funding climate.
Mergers and acquisitions, along with companies closing shop altogether, will be inevitable in the coming year, said Edwards. “There are just so many companies with solutions that vary from being promising, to being average, to being useless. There’s obviously a shakeout going on” — an outcome, he noted, that would ultimately be “a good thing for the market.”
The question, of course, is who will remain standing this time next year. With the ranks of larger, publicly traded bioinformatics firms dwindling either through M&A or strategic makeovers, it’s a good bet that the pure-play bioinformatics model is a thing of the past for companies seeking to grow beyond a few million dollars in revenue a year.
David Stone, managing director of investment firm Flagship Ventures, cited Lion Bioscience and InforMax as examples of how difficult it has proven to generate substantial growth in the sector. “Software companies getting up to around $20 million in revenues are not being able to grow beyond that point.” In the case of InforMax, he noted, “The company was sold because its revenues are actually going backwards.” A crop of new bioinformatics business models has appeared over the last year or so, including strategies that place proprietary software tools within the context of a larger, integrated solution as well as more collaborative and research-based approaches. 2003 may be the year that proves which — if any — succeed.
For new companies planning to launch this year, financing will remain out of reach for all but the most promising startups. “There’s fewer and fewer investors in the space, and those that are staying are making greater demands in terms of how does this really play out in terms of a business and financial model,” said Arthur Pappas, CEO of venture firm A. M. Pappas and Associates. While Pappas said he hasn’t lost his faith in bioinformatics as an investment yet, his firm is evaluating how new companies “are fitting into the bigger picture” — whether that be through a multi-purpose solution that can be applied across several research “silos” in drug discovery, or a more specialized technology that has potential as one component of a broader platform offering. Pappas said the advice he’s giving to bioinformatics entrepreneurs is to “ignore the past” because the period of “pure bioinformatics” — which delivered single-point solutions — is long gone.
Stone advised bioinformatics entrepreneurs to get some customers lined up before approaching a venture capitalist. “The first concern of investors is whether an informatics company can get customers,” he said. Funding based on the promise of a technology alone will be nearly impossible to get in 2003 without some (positive) user feedback.
Same Old Problem, New Crop of Answers
Another familiar face from 2002 has been dredged up as one of 2003’s primary technology challenges. Integration, itself a familiar buzzword in the industry, is now accompanied by an entourage of catchphrases — ontologies, web services, and knowledge management — that promise to alleviate the tedious task of bridging disparate biological data sources.
Largely driven by the success of the Gene Ontology consortium during the course of 2002, ontology-based approaches to data management and analysis are set to flower in 2003. Arthur Thomas, a life science IT consultant with Proteus Associates, said his pharmaceutical clients are “just beginning to make significant investments to train people this way — to take public domain ontology information and integrate it with their own internal ontological models.”
A less established approach, but one likely to create a bit of noise in the coming year, will be web services-based technologies, in which databases, applications, and computers are registered with a central server that can retrieve the appropriate resources when a query is submitted. The I3C is promoting a web service-based approach to interoperability through its proposed LSID (Life Science Identifier) specification, and a number of other in-house and collaborative bioinformatics web service projects sprang up in 2002.
But the jury is out on just how successful this approach will be, and whether any short-term benefits will be realized from web services in 2003. While many organizations are evaluating web services technology to varying degrees, few are fully sold on it yet, largely due to the enormous hype that it has already generated. “Everyone is labeling everything as web services,” said Thomas.
Joe Landman, founder and CEO of informatics consulting firm Scalable Informatics, was equally cautious. “Web services is one of the things that looks like it has a tremendous amount of promise, but associated with that level of promise is a significant amount of hype. The difficulty with this is this is setting expectations I consider to be unrealistically high for what it can do and what problems it can solve. This may prove problematic for web services in the not too distant future.” As more companies give web services a try in the coming year, some of these doubts are likely to be borne out or resolved.
Another candidate for overhype in 2003 will be knowledge management — an often vague term that 3rd Millennium's Dweck defines as “more about storing the findings than the raw data itself.” Dweck predicted that knowledge management methods would be driven by cost as well as by research demands in 2003. “Before, companies could throw money at whatever problem they saw, so they could afford for each individual group to rediscover the same thing, but now they can't; they have to be much more efficient. But every researcher in an organization wants to know, ‘What do we know about this thing I'm studying? What is all the knowledge in this company?’”
IT Shifts from Big Iron to Distributed Solutions
On the IT side, the biggest challenges for 2003 probably won't be in processing power or storage, according to some observers. “People are learning that throwing raw computing power and information storage and data mining at the problem isn't the answer for discovery or clinical trials,” said Keith Batchelder, an independent consultant who advises pharmaceutical companies on clinical trials.
Landman observed that one likely IT driver would occur in data transfer. “People who've been building out infrastructures to gather tremendous amounts of data also now see that they have tremendous amounts of data motion that they have to worry about … We're starting to see a need for true distributed computing resource in not just a cluster or grid sense, but having some kind of method to place the computing near the data because that may be cheaper than actually moving the data.”
Distributed computing solutions, which entered the bioinformatics mainstream in 2001 and 2002, will continue to gain ground in 2003, expanding their reach into data storage. This trend will be driven largely by cost, as end users seek more effective methods to utilize the compute resources they already own, and also by improved technology that promises to organize computational as well as data resources using the grid model.
This trend could trigger a slowdown in large-scale hardware spending, negatively impacting vendors such as HP and Sun. One player everybody will be keeping their eyes on will be IBM, which has broadened its offering beyond hardware and has aggressively pursued a partnership strategy that could likely make it the clear leader in life science IT. Said Edwards, “The big pharmas will probably look more to the companies like IBM to say: ‘Help us choose the right vendors. There are so many vendors out there and we don’t really know the right ones to work with.’” If this is the case, those companies who have already secured partnerships with Big Blue, including Lion and Accelrys, should be in fairly good shape for the coming year.
Bioinformatics Expands its Boundaries
By the end of 2003, it’s likely we’ll be hearing less and less talk of bioinformatics as its own discipline, and more use of general phrases like “research informatics” or “life science informatics” as integration with chemi-, clinical, and medical informatics becomes more commonplace. “People have realized that the boundary between different disciplines is becoming blurrier and blurrier,” said Thomas.
Even as bioinformatics itself grows to address the challenges of gene-expression analysis, proteomics, pathway analysis, systems biology, cellular simulation and modeling, and other post-sequencing aspects of high-throughput biology, the field is also expanding beyond its former barriers to merge with research domains along the drug discovery and development pipeline.
This trend has raised one important question on which there is still no consensus for the coming year: The impact of regulatory issues on discovery data. As more and more clinical and medical data are being brought back into the discovery end of the pipeline, questions of compliance with FDA regulations — particularly the 21 CFR Part 11 guidelines requiring electronic submissions of drug applications — are beginning to arise, and there are no easy answers available.
Within one to two years, according to Jim Freeman, “21 CFR Part 11 will change how companies store and collect data and how they secure their data. That will have tremendous effect in clinical and pre-clinical, and that effect will reach back into discovery and people will have to standardize the use of resources across the entire organization.” Dweck, however, said his company is only now “hearing more talk about it, but we don’t quite see the movement yet.” Most companies and vendors alike are taking a wait-and-see attitude until a real demand, driven directly by the FDA, is evident.
Finally, prospects are still good for those hiring in bioinformatics, but that’s a trend that may not last forever. Bioinformatics departments gained from the downturn in the technology sector as many laid-off IT specialists turned to biotech for employment. However, if the tech sector experiences an upswing in 2003, look for the talent pool to dwindle considerably as IT employees realize their future prospects in bioinformatics are limited without a PhD in biology.
But this talent may not wind up going that far. As Freeman pointed out, many IT professionals may take a familiar path: “They could leave or they could become consultants.”