When it comes to the Windows operating system, there are two schools of thought in the life science informatics community: Computational scientists, who tend to prefer Unix or Linux, view it with indifference or even disdain; while bench scientists, content with Excel and Outlook, embrace the familiar point-and-click interface with open arms. But Microsoft is hoping to bridge the gap between these two points of view, and in doing so, gain market share in the life science computing sector. Recently, the company formed a collaboration with Optive Research — a 15-employee computer-assisted drug discovery software firm based in Austin, Texas — to move toward that goal.
According to Bryan Koontz, vice president of marketing for Optive, the collaboration will work on both ends of the operating system continuum. Microsoft will help the company port its server-based Concord computational chemistry engine to Windows from Irix and Linux in an effort to demonstrate the platform’s suitability for high-performance research computing, and will also help develop a new version of Optive’s Benchware line of desktop software, which already runs on Windows.
Koontz said the collaboration benefits both parties. Optive had originally approached Microsoft with some questions about the .NET web services framework, which the company was planning on using to integrate the Benchware suite with other applications. Microsoft offered to help with usability features and other aspects of the software design, which Koontz said would be key features for keeping end-user medicinal chemists happy. “The goal is to make Benchware as familiar to use as any Microsoft Office product, like Excel or Word, so it’s really usable by a bench chemist who doesn’t think like a computational chemist,” he said. Optive expects the .NET framework to make the suite interoperable with lab instrumentation software, as well as other Microsoft Office software products and third-party products such as informatics databases.
But Microsoft also stands to gain from the partnership. The company has tapped Optive to help Windows gain a foothold in the HPC market for life science research, where Unix and Linux currently reign. “It’s really hard for Microsoft to break in and prove to everyone that Windows is just as good as Linux and people should be using Windows-based servers for high-performance computing,” Koontz said. Microsoft turned to Optive to learn how to “speak the language of the scientist” in an effort to court the life science HPC market, he added. “They would like to demonstrate that Windows can achieve at least platform parity — if not platform superiority — over Linux in particular.”
In Koontz’s view, Microsoft is hoping to convince pharma and biotech that there are total-cost-of-ownership advantages to leveraging the Windows platform throughout the organization: “I think their message is that if you standardize on a single computing platform that’s easy to maintain and easy to use, and may have total-cost-of-ownership advantages, and you get the same scientific results, and it runs just as good, if not better, from a computing performance perspective, then by all means, why wouldn’t you move to Microsoft?”
Dan’l Lewin, corporate vice president of .NET business development, said in a statement that the collaboration was part of Microsoft’s “continuing investment in high-performance computing solutions for life science research.” Microsoft officials were not available to elaborate on Lewin’s statement or to comment on the collaboration prior to BioInform’s publication deadline.
A company spokeswoman told BioInform via e-mail that Microsoft “is committed to the life science market,” and is “taking a grassroots approach to understand the needs of emerging businesses early in their development process.” Although the company is “always interested” in partnerships similar to the one it has formed with Optive, “At this point there aren’t any new partners that we are able to announce,” she said.
Koontz said that Optive and Windows are working together to optimize Concord for Windows 2003 Server, and that the companies plan to benchmark the performance of the code against similar systems running Linux at some of Optive’s client sites.
Although Optive is a small company, Concord and its other discovery engines — such as the conformational analysis program Confort — are installed at a number of large pharmaceutical firms, which would give Microsoft instant street cred in the informatics community if the benchmark results turn out favorably.
Koontz admitted that although Optive sees strong demand for Windows-based desktop systems among medicinal chemists and other end users, “for the computational chemists, where most people have been using Linux and Irix, we don’t get calls from people asking us to move all this stuff to Windows, and we were very up-front with Microsoft about that.” He added, however, that Optive has been pleased with the computational performance and scientific results it has seen so far on the Windows platform in preliminary studies, and would be happy to “evangelize” the platform through co-marketing agreements if the benchmarks are successful.
But even if Optive and Microsoft are able to demonstrate that Windows performs as well as Linux or Unix, computational scientists within pharmaceutical firms will likely take a bit more convincing before they switch platforms. Bruce Ling, director of bioinformatics at Tularik, said that it will still take some time and effort to break the habits of computational chemists “who grew up with Unix” in the event that Windows proves to be an effective HPC operating system.
The head of informatics research at a large pharmaceutical firm, who asked not to be identified by name, told BioInform that moving to Windows would actually increase the number of platforms currently in use at the company because all of its large enterprise applications are already running Unix. For desktop applications, he said, Windows is fine, but when it comes to critical enterprise applications, he said he “wouldn’t feel comfortable” moving to Windows, even if it meant an immediate cost savings on server hardware.
Preliminary results from the benchmarking studies are expected by July, Koontz said. The company has not yet set a target date for the release of the new version of its Benchware suite.