Microsoft Research is supporting the Swiss Institute of Bioinformatics and the University of Geneva in a joint effort that will use mass spectrometry to test blood samples for toxic biomarkers.
The goal of the two-year collaboration, which kicked off last week, is to develop a bioinformatics platform to help identify protein biomarkers that may indicate adverse drug interactions. Microsoft is still determining the overall nature of its software contribution, but will at least contribute its debugging software to the project.
Geneva Bioinformatics, SIB’s commercial arm, is also participating in the project, which will use the company’s Phenyx mass spectrometry-analysis software.
Ron Appel, SIB’s executive director, told BioInform that his group has adapted Phenyx for the project “to not [only] analyze proteins of peptides — which is the main focus — but to search for small molecules, non-protein[s], especially toxic markers for clinical applications.”
SIB’s Frederique Lisacek, who is directing the biomarker project, told BioInform that the group will “tailor” the institute’s existing proteomics software tools for the initiative. In particular, she said, “special emphasis will be put on the study of post-translational modifications and statistical validation.”
Lisacek said the collaborators will focus on lung and liver cancer, with the possibility of working on pancreatic cancer biomarkers in the future. Samples will be drawn from the George Pompidou hospital in Paris as well as an as-yet-unnamed hospital in Geneva.
She said the project will start with 300 samples, but there are no definite plans regarding additional samples that they may need.
Tony Hey, corporate vice president for external research at Microsoft Research, told BioInform that the team will use Microsoft’s Dr. Watson for Windows technology on the project.
Dr. Watson is a program error debugger that gathers information about system errors and reports them to support personnel at Microsoft. The system generates “huge amounts of data” that Microsoft needs to mine efficiently, Hey said.
Microsoft’s collaborators at SIB and the University of Geneva “are going to use the same sort of technologies to deal with the toxic biomarker data we are generating,” Hey said.
“So we, at the beginning of this project, are sitting down with them, going through all their bits and pieces to see what technologies we have to offer. We have the data mining technology from research, but I believe there are other technologies from our products that could be adapted to be useful.”
Microsoft Research also intends to evaluate other technologies it has that may be used in the project. Like Lisacek, Hey said that they will not build a new platform from the ground up.
Rather, Microsoft intends to work with existing technologies from GeneBio, SIB, and the University of Geneva, “and maybe show them how you build a database and put together a system which can much more efficiently examine and manipulate the data,” he said.
“So we, at the beginning of this project, are sitting down with them, going through all their bits and pieces to see what technologies we have to offer,” Hey said. “We have the data-mining technology from research, but I believe there are other technologies from our products that could be adapted to be useful.”
This is not Microsoft Research’s first foray into bioinformatics. For example, in July, BioInform reported that the research group had received $700K in grants to promote genome-wide association studies. And the previous month, the group co-sponsored a three-day workshop at the University of Pennsylvania to promote data integration in the life sciences (BioInform, June 29, 2007).
Hey said Microsoft Research has gotten involved on the SIB collaboration because it wants to work on problems that people care about and make a difference, such as they’ve done in the past with HIV/AIDS and other diseases.
“Or, in this case these are toxic biomarkers and the idea is if we have a really world-class research group and we can have some computer science from our Microsoft research centers around the world, that could help the scientists in dealing with and finding subtle signals in huge amounts of data,” Hey said.
Hey added, “I try to bring research technologies out of the lab and … to use our technologies to help scientists solve problems.”