Its software has been a technical computing staple for nearly 20 years, and it boasts a number of biotech and pharmaceutical companies among its customers, but The MathWorks — known primarily for its Matlab numerical programming environment — only recently began adjusting its products to meet the needs of biology researchers. With the mid-November launch of its Bioinformatics Toolbox, however, The MathWorks jumped into the bioinformatics software market with both feet.
The company began seeing interest from the bioinformatics community a few years ago, according to Rob Henson, manager of bioinformatics development at the MathWorks. “We started getting phone calls from people who used to work for aerospace and communications companies who were now working for companies with ‘bio’ in their name. They were laid off when the aerospace industry and the communications industry went downhill, getting picked up by the biotech companies, and taking Matlab with them,” Henson said. The MathWorks soon realized that “we were selling an awful lot of software to companies who we didn’t really know what they were doing, so we started to look at what they were doing,” he said.
After gathering feedback from customers who were applying Matlab’s data analysis, numerical analysis, statistics, and visualization capabilities to bioinformatics applications, the company soon took steps to develop its own set of bio-friendly Matlab add-ons. Early adopters in the bioinformatics sector included Applied Biosystems, which used Matlab to create a calibration algorithm for its 3700 sequencer; and Rosetta Inpharmatics, which used Matlab to prototype algorithms and code for Rosetta Resolver. In addition, a number of academic groups have used Matlab to develop bioinformatics software tools (see box, below)
However, these Matlab pioneers found that the package had its limits when it came to biological data. “One of the tricky things that people found was how to get their data into Matlab,” Henson said. “Once they got it in, they could do all this clever math and statistics and visualization, but their stumbling block was getting the data in.”
A key feature of the bioinformatics toolbox, therefore, is support for common genomic, proteomics, and gene expression file formats for developers prototyping algorithms or building applications with the Matlab programming environment. In addition, the toolbox provides an interface to common bioinformatics databases, several genomic and proteomics sequence alignment tools, and microarray normalization and visualization capabilities.
Kristen Amuzzini, who is biotech, pharmaceutical, and medical marketing manager for The MathWorks, said that the toolbox is geared primarily toward “the bioinformaticist developing algorithms, either to support researchers in the lab or to help out some of the software engineers with prototypes,” although the environment is user-friendly enough for some end-user biologists. With developers as its target market, therefore, the company’s primary competition for the product isn’t from bioinformatics software vendors, but from in-house development teams, Henson said. “Our biggest competitor is people doing it themselves in C or Fortran and thinking they can build it themselves as quickly as they get it from our tools,” he said.
Pricing, however, shouldn’t be too much of a barrier for these holdouts. An individual academic license for the bioinformatics toolbox is $200 (although licenses for Matlab and the Matlab statistics toolbox are required at $500 and $200, respectively). For commercial users, the bioinformatics toolbox is $1,000 per seat (and the Matlab and statistics toolbox licenses are $1,900 and $600, respectively).
Bioinformatics certainly isn’t a core component of The MathWorks’ business — out of a total of more than 1,000 employees, the bioinformatics group employs only five right now, Henson said. But the company is confident that the market for its products in this area will continue to grow. The company is hiring for the bioinformatics development team, Henson said, and is currently seeking a developer for systems biology applications using the company’s Simulink modeling environment.
“Bioinformatics is a very, very big subject, so I think we’ll be busy for the next few years,” Henson said. “We’ll follow what the industry does, and where we believe our tools are most applicable, we’ll be building application-specific products.”