Like many general-purpose IT companies, Tom Sawyer Software has identified the life science sector as a promising market for its graph visualization technology. But the Oakland, Calif.-based firm has a bit of a head start in its bid to build a bioinformatics business — life science sales already make up around 20 percent of the company’s activities, according to CEO and founder Brendan Madden.
Tom Sawyer, which was founded in 1991, built a core base in networking during the nineties, with customers such as Cisco, Nortel, and Lucent Technologies embedding the company’s graph analysis and visualization software in their own applications to represent complex relational data sets. In the late nineties, the company signed GlaxoSmithKline as its first life science client, when the pharmaceutical giant approached the firm looking for a better way to visualize its biological data. It turned out that the company’s technology, developed to identify relational patterns among people, bank accounts, telephone calls, and the like, was perfectly suited for visualizing biochemical pathways, genetic networks, ontologies, and other biological data sets.
At first, Madden said, Tom Sawyer had “just a few, isolated customers” in the life science market, but that business has steadily grown over the past two years. Tom Sawyer currently lists Affymetrix, Amgen, Aventis, Berlex Biosciences, Beyond Genomics, Biogen, InforMax, Merck, and Pioneer- Hi-Bred International among its bioinformatics clients.
Currently, most of these customers are using Tom Sawyer’s software in the same manner as the company’s traditional networking clients — as an embedded visualization front end for in-house applications. But Madden said the firm is looking to move beyond this supporting role. “We’re a little bit behind the scenes now, but that’s changing,” he said. “We don’t want to be like ‘Intel inside.” We’d like to be a little more visible, and reach out more to the end user ourselves.”
To do this, the company is planning new versions of its products targeted to specialized vertical markets. In bioinformatics, Tom Sawyer has already developed an XML interface for the KEGG pathway database, a web interface with “some pathway support,” and is developing a software package for visualizing biochemical pathways that it plans to release by the end of the summer. This pathway software is based on work from several of Tom Sawyer’s life science consulting projects, Madden said.
The company offers three software libraries — Tom Sawyer Analysis, Tom Sawyer Visualization, and Tom Sawyer Layout — that allow developers to build applications for viewing relational data. But the company is currently putting the finishing touches on an enterprise version of its software that will be targeted more toward end users. “Our traditional customers in the old days would always just use our APIs, and we became their visualization,” he said. “But more and more we have everything built so you just configure XML and go, so there’s much less need to program.” The enterprise software — expected to be released this fall — will offer a distributed architecture that will give multiple users access to data and models.
Spotfire is currently the 800-pound gorilla in the life science data visualization market, but Madden doesn’t see the company as a rival. “We don’t even have the same technology,” he said. Spotfire’s technology represents attributes of the individual data points, he said, while Tom Sawyer’s methods represent “relationships in the data as opposed to the data itself,” making the approaches complementary rather than competitive.
The rapid pace of biological discovery may prove to be Tom Sawyer’s biggest hurdle in the bioinformatics market. “Pathways are getting larger and larger,” said Georges Grinstein, a bioinformatics and data visualization expert at the University of Massachusetts, Lowell, “so you really need to be able to take systems, and modify them for specific environments.” Grinstein’s team uses Tom Sawyer’s software as the framework for some of its software packages, but the company’s tools have their limits, he noted. “What I liked about the Tom Sawyer software was that they already had embedded algorithms for layout, and that saved my having to do that,” he said. “But as their environment is not continually evolving, we have had to start building some of those things for specific applications.” One drawback, he noted, is that the company doesn’t offer access to its source code, so any modifications require rewriting entirely new code.
“In the future — unless they continue to evolve quickly to include more advanced algorithms and more applied capabilities — they’re going to find that people are turning to other systems,” Grinstein said.