One year after throwing its hat in the bioinformatics ring, technical computing mainstay The MathWorks has released the second version of the Bioinformatics Toolkit that it launched last November, and is preparing a new product targeting the computational systems biology market that should launch by the middle of 2005.
“The life science market has taken on a life of its own within our company,” said Kristen Amuzzini, biotech and pharmaceutical industry marketing manager for The MathWorks. The company’s bioinformatics team has grown from five people to 20 people over the last year, and Amuzzini said that around 70 commercial firms and 300 university groups are currently using the first version of the Bioinformatics Toolbox.
“The level of acceptance completely surprised us,” Amuzzini said.
The first version of the Bioinformatics Toolbox essentially served as a sequence and gene-expression analysis plug-in for the company’s flagship Matlab technical computing package. Version 2.0 expands those capabilities with the addition of mass spectrometry data analysis — specifically, the new release includes functions for pre-processing mass spectra before they are fed into peptide matching engines like Mascot or Sequest.
Pre-processing functions in the toolbox include baseline correction, smoothing, alignment, and resampling, along with a GUI that allows users to view multiple spectra simultaneously. There are also functions that allow users to create classifiers and identify potential biomarkers.
Version 2.0 also includes a visualization tool for representing molecular interaction graphs, machine-learning functions for classification and pattern recognition, enhanced microarray normalization tools, and several new sequence analysis features.
The new release is also compatible with The MathWorks’ recently released Distributed Computing Toolbox, which allows users to execute Matlab algorithms across a compute cluster.
The list price for the Bioinformatics Toolkit starts at $1,000 per seat.
Amuzzini said that the first version of the toolbox caught on in bioinformatics groups who were “bogged down with too many tools and spending too much time integrating them.” Matlab offers a flexible command-line development platform for bioinformaticists who need to write and deploy new algorithms quickly, while also providing a point-and-click interface that end-user biologists feel comfortable with. This environment has turned out to be an effective compromise between the inflexibility of off-the-shelf bioinformatics software and the development time associated with writing everything from scratch.
John Keilty, director of informatics at Infinity Pharmaceuticals, told BioInform that his company recently standardized on Matlab for precisely this reason. “A lot of software companies try to solve everything with one fell swoop, but this is unsuccessful,” he said. “It forces the scientists to do things in a certain way to make the software work the way it should.”
Keilty said that his team uses Matlab to develop and deploy curve-fitting models for high-throughput screening data throughout the organization. The system is integrated with other software packages like Spotfire, and new models can be written and pushed out to the researchers on the fly.
But The Mathwoks isn’t the only technical software firm to recognize the need for this type of capability in the life science market. Wolfram Research has also set up an internal bioinformatics team to tweak the company’s Mathematica package for the life science research market. Mohammed AlQuraishi, the company’s lead developer in computational biology, said that Mathematica 5.1, which the company released last week, includes a number of features of interest to the bioinformatics community, particularly for string matching, cluster analysis, and large-scale data analysis.
Wolfram has no plans to launch a targeted bioinformatics toolkit for Mathematica like The MathWorks has for Matlab, but nevertheless, “computational biology is a very important market for us,” AlQuraishi said. Wolfram sees a large opportunity for its general-purpose technical computing package in the life sciences, he said. “One problem we always see in this market is that there are lots of small tools that focus on one area of functionality or another, but linking these tools is a problem. We believe that Mathematica could be a platform for integration in computational biology.”
Whether there is room for both of these players in the relatively small bioinformatics market remains to be seen, but The MathWorks is clearly aiming to claim first-mover advantage with the upcoming launch of its systems biology package. Amuzzini was unable to disclose any specific features for the product, but she did say that it will be “a full platform” — as opposed to a toolbox — that will focus on the simulation and analysis of biochemical systems. A beta version of the system went out to selected users last week, she said.