Calling it an unlikely career move would be an understatement.
Labkey, a Seattle-based bioinformatics startup preparing to release the first version of its open-source proteomics analysis system, is the brainchild of several former Microsoft developers a background generally not associated with proteomics or open-source software, let alone both.
The company's six founders all left Microsoft in the mid-90s. Co-founder Mark Igra told BioInform that after a "long road" that included the launch of database software firm Westside and its eventual acquisition by BEA in 2002, the team found its way to the lab of Martin McIntosh at the Fred Hutchinson Cancer Research Center, where it developed the lab's Computational Proteomics Analysis System (CPAS).
Igra said that he met McIntosh through "a friend of a friend," and joined the lab around two years ago. Since then, he said, as the CPAS project evolved, he brought his former colleagues on board one by one.
"Martin wanted professional-quality software, and he built the team here to do that," Igra said.
CPAS includes the X! Tandem open-source protein search engine, data-mining tools and viewers for mass-spec data, and project-management and sample-tracking tools. The system can also be extended beyond mass-spec data, and the team has already built a set of modules for flow cytometry, with another for gel-based experiments in the works. The idea behind the system is to "expose large data sets in ways that are accessible to researchers," Igra said. At the Hutch, CPAS currently manages more than 28 million mass spec runs.
"We're not so interested in being a sales and marketing organization; we learned that about ourselves."
CPAS proved to be a success in the McIntosh lab, and Igra said that he and his colleagues "realized that there is a widespread need for this kind of software a platform for dealing with experiments of any kind."
Working with the Hutch, which owns one-seventh of Labkey, Igra and his partners decided to launch a business around the system's underlying technology, which they have dubbed Labkey Portal. The motivation for the company came from a desire "to solve similar problems for other researchers and expand the system and make it more available, more functional and visible, over time," Igra said.
Several other research labs at the Hutch are using the system, as well as academic collaborators, such as the University of Michigan, but the first public release of the system is scheduled for later this month. The company plans to release all of its source code under the Apache license, with revenues coming from consulting and services based around the platform.
The company expects the software to appeal to labs that are just getting started in proteomics. "Someone can buy a mass spec and download our software and they'll have a soup-to-nuts proteomics system," Igra said.
Labkey envisions the platform as a true open-source project, and welcomes new modules and other improvements from the community, Igra said. "We built the system as a general-purpose platform that could analyze any kind of data, and we haven't even scratched the surface yet for all the different modules" that are possible, he said. One mark of a successful open-source project, he noted, is that it "attracts developers."
Igra said that he and his colleagues had several reasons for adopting an open-source business model. Having already worked at a software giant as well as a startup, the team of self-described techies decided that "we're not so interested in being a sales and marketing organization we learned that about ourselves."
Secondly, he said, the open-source approach was deemed a much easier entry point for consulting engagements, particularly among the academic customer base that the firm is targeting. "We work largely in the world of nonprofits," Igra said, and open-source software "makes investigators more comfortable because it's a resource that can be reused."
An open-source policy "also resonates well with the NCI," Igra added.
As far as the potential market for Labkey's software, Igra acknowledged that "it's not a lot by the standards of our previous software companies, but there's a trend toward more and more data coming available from all of these assays, and these data need to be managed."
And the firm isn't setting any unrealistic expectations. "Nobody is going to get insanely rich, but we'll build good products," Igra said. "We'll forego the big IPO."
Bernadette Toner ([email protected])