If the demise of Molecular Mining two weeks ago wasn’t enough of a warning for smaller microarray software vendors that their days may be numbered, a recent round of product launches in the sector should drive the point home. In a strategic departure from the shrink-wrapped software model typical of many firms in the space, both Spotfire and statistics veteran Insightful are introducing flexible and extensible platform-based microarray analysis systems that can evolve in step with the quickly changing needs of their users.
Both companies claim that this approach will give them a decisive advantage over niche players as consolidation pressure increases in the sector.
“This is war,” said Spotfire CEO Chris Ahlberg of the heightened competition in the current marketplace. Spotfire’s strategy for the battle ahead is an extension of its technology into every step of the drug discovery and development pipeline. Banking on the promise that biotech and pharmaceutical firms will soon apply microarray technology well past target research and into screening, ADME-tox, and pharmacogenomics, the company is launching a new chemogenomics module for its DecisionSite platform in the next few weeks “that will integrate Affymetrix data with chemistry data in some really exciting ways,” Ahlberg said. The chemogenomics module will soon be followed by at least three other similar modules to drive the company’s gene expression analysis capabilities even further downstream.
“Software that addresses what researchers need to do in drug discovery has to be thought about not just as analyzing the data file that came out of an Affymetrix experiment, but it has to focus on how you integrate that with other data in the drug discovery process,” Ahlberg said. As the market evolves in this direction, he predicted, pre-packaged software that only performs gene expression analysis will soon become obsolete. In that narrow sense, “gene expression software is dead,” he said. However, he added quickly, “Long live gene expression software across discovery.”
Noting that some of Spotfire’s smaller-sized competitors are already “starting to go away,” Ahlberg pointed out that “you have to be able to build big engagements with your customers to be successful, and I don’t think there’s too many who can do that.”
Michael O’Connell, director of biopharmaceutical solutions for Insightful, agreed that the microarray analysis field is ripe for a major shakeout. “There are many players, but it’s a somewhat fragmented market of shrink-wrapped solutions that are consolidating,” he said. “There’s a need for a more flexible data analysis solution that’s strong from a statistical perspective and can evolve with rapid market advancements…So customers can grow with the product and are not trapped by the release cycle.”
Richard Leavitt, Insightful’s vice president of product marketing, offered a similar assessment of the market. “If you don’t have a stable customer base and a breadth of offerings, I would not want to be a young company in this space right now,” he said.
Like Spotfire, Insightful is also hoping to extend its reach across a broader segment of the drug discovery pipeline, but in the opposite direction. The company already boasts a strong pharma customer base for its S-Plus statistics package in clinical trials and biostatistics, “so unlike a lot of discovery companies who are trying to go downstream, we’re actually going upstream by getting into discovery,” O’Connell said.
Last week, Insightful launched S+ArrayAnalyzer, a software package for microarray analysis based on S-Plus. The company first entered the microarray market in September with an initial version of ArrayAnalyzer that Leavitt described as a “services offering.” The new release is an “enhanced” product that allows users to access, prepare, analyze, and visualize microarray data on top of the statistics foundation of S-Plus. In addition, the company collaborated with the developers of the open source Bioconductor project to port some of the project’s R-based components into the software.
Not Just Another Pretty Face
While it may appear that Insightful is entering the gene expression market at an inopportune time, the company is confident that its large installed base and statistical prowess will give it an edge over its competitors. “Some of the other tools will give you pretty pictures but incorrect statistical analysis, but we’re known as having very high-end statistical methods for analyzing data,” said Leavitt.
O’Connell cited the software’s normalization capabilities and ability to accurately identify differentially expressed genes as key advantages over competing technologies. The company uses what it calls a “volcano plot” that charts fold change against P-value in order to determine differential expression with a smaller number of false positives and false negatives than traditional methods, he said. “Our position on some of the shrink-wrapped products for microarrays is that they are primarily visualization and not a lot of analytics. Frankly I’ve seen a lot of junk out there,” he said.
S+ArrayAnalyzer does offer visualization capabilities — such as “graphlets” that allow users to drill down to annotation data for selected genes — but at the request of its customers, the company has also built a connector to Spotfire, which O’Connell described as “a visualization tool that doesn’t really have any depth on the analytics side.”
While not agreeing entirely with this characterization of Spotfire’s technology, Ahlberg conceded that statistics is not his company’s forte. The DecisionSite module for functional genomics does contain a core statistics base, he said, and the company also offers a separate DecisionSite configurationfor statistics, “but no gene expression vendor is going to be able to keep up with all the different things that people are doing [in statistics], so we built really nice hooks into SAS, Matlab, and S-Plus so people can take their own algorithms and hook them into Spotfire.”
However, Ahlberg said, Spotfire does provide a level of analytical capability that many users who rely on the software for its visualization capabilities are not aware of. Over the last year, the company has made some headway in getting its pharmaceutical clients to integrate the software more deeply into their work processes, “but we’re still fighting that old perception,” he admitted.
While Spotfire and Insightful may differ in their core strengths, they do share a similar philosophy for assimilating their software into the working environments of their clients: stay flexible and stay open.
Through partnerships and its own development work, Spotfire has ensured that its software is compatible with a broad set of data and applications. On the data side, the company has developed data models for commercial data sets from MDL, ActivityBase, Affymetrix, and Agilent, as well as public data sources and the proprietary data sets of its pharmaceutical clients. “Instead of attacking this as a gene expression problem, we’ve attacked that part as a data access and data integration problem, and thereby built our technology to make it very flexible [in] being able to go out to new data sources as we go along,” Ahlberg said.
On the application side, Spotfire offers more than 250 API calls, Ahlberg noted, which allows bioinformaticists to easily integrate Spotfire into the databases and software they built themselves. “We usually say that by the time you’ve integrated it, you’re going to think of this as your own system,” Ahlberg said. “We don’t mind if you don’t think about it as Spotfire anymore. The key part is that it is so integrated.”
Insightful also stressed the flexibility of its software. “We keep hearing from our customer base that it’s critical that the application environment be able to move along with advances in the field. This is where a lot of the other vendors seem to be having trouble,” said Leavitt. This consideration ultimately led to the company’s collaboration with Bioconductor, O’Connell said, since customers would have access to the latest cutting-edge algorithms from the academic world with all the usability benefits of a commercial software offering. “It’s an extensible development environment for folks to grow with our product, to incorporate open source code, or to extend the product with their own code,” he said.
In return for Bioconductor’s contribution to S+ArrayAnalyzer, Insightful is sponsoring a graduate student involved in the project and also plans on releasing some of its own code to the Bioconductor package.
In a market notorious for its rapid evolution, Spotfire and Insightful are each positioning themselves for the next phase of microarray gene expression technology — whatever that may prove to be. “Gene expression technology is going to be applied in ways that are hard to predict and follow,” said Ahlberg. “That’s why we’ve built a platform for doing data analysis, and a platform for decision making…We have a configuration that we sell for gene expression, and that’s great, and we’ll keep running with that as long as we can, but if that turns into something else, we’ll run with that.”
Insightful, on the other hand, perceives its newest product as simply a “natural extension” of its experience with downstream pharma applications. Unlike some of its “one-trick pony” competitors, the company isn’t gambling its entire business on the success of the microarray software market. “If you are a company solely focused on microarray analysis, you may be already struggling to find out how you’re going to break your way out of that,” said Leavitt. “We’re really a company that’s established in the market and growing, and we see this as adding to our portfolio of products. I don’t think we see it as the flagship of our life sciences offering.”