When Electronics and engineering giant Siemens launched an internal bioinformatics R&D effort to support its medical systems business recently, it soon found that a critical component for building its bioinformatics competency was lacking: experimental data. In response, Siemens Corporate Research, the company’s US R&D lab in Princeton, NJ, has begun to build a network of partners and collaborators willing to share their biological data sets in exchange for the company’s analytical expertise.
“We need to collaborate and establish relationships with academia, pharmaceutical companies, and biotech companies,” said Lance Ladic, business development manager for biotech and life science IT at SCR. Ladic, speaking at a systems biology symposium at the Institute for Advanced Study in Princeton last week (see story, p. 3), said the company has “a good set of core technologies” that have potential applications in biotech. With a strong background in IT, but no wet labs of its own, Ladic said the company is actively seeking partners to complete the bioinformatics picture.
So far, SCR has active bioinformatics collaborations with two undisclosed academic institutions, and several more are under discussion, according to Claus Neubauer, R&D program manager in SCR’s Intelligent Data Analysis and Modeling group. The biggest challenge for his team so far, Neubauer said, is “getting complete data sets.” Neubauer’s group is one of two research departments at the R&D center that are actively pursuing bioinformatics. The other group, Integrated Data Systems, is also eager to get its hands on some biological data. “We are hungry for data; that’s why we’re here,” said Xiang Zhou, a researcher in SCR’s Integrated Data Systems group, in his talk at the IAS symposium.
Bioinformatics Competency … and Competition
Siemens is a highly diversified global company, involved in everything from power generation to communications. But SCR’s geographical proximity to Siemens Medical Solutions in Malvern, Penn., has given the 250-person R&D group a bit of a healthcare bias, Ladic said. When SCR undertook a recent roadmapping initiative, he said, it became clear that its experience in imaging informatics, data mining, and knowledge management could be applied to biotech and, more importantly, to Siemens’ current medical systems business. Scattered bioinformatics projects that were already underway with SCR were soon formalized under the two departments, and the company plans to expand its bioinformatics research group, Ladic said.
The company’s foray into bioinformatics follows that of GE, Siemens’ competitor in the medical imaging market. GE set up a bioinformatics research group in its Niskayuna, NY, Global Research Center well before it signaled its interest in purchasing life science giant Amersham in October of last year [BioInform 07-04-03]. SCR staffers were unable to comment on any broader plans that Siemens may have in the life science market, and a spokesman from Siemens’ Berlin headquarters told BioInform that the company’s corporate bioinformatics policy is currently in a “confidential status.” Nevertheless, it is clear that the company is expanding its activities in the area. The Irish press reported last week that Siemens plans to build a new research facility in Dublin at the Royal College of Surgeons that will focus on bioinformatics. “One area where Siemens sees a big future is in the area of bioinformatics,” said Lorenz Zimmermann, managing director of Siemens Ireland, according to one report.
So far, SCR has explored a range of applications for its proprietary general-purpose informatics tools. The Intelligent Data Analysis and Modeling group, for example, has developed a data mining and visualization framework that it is applying to protein analysis, gene expression analysis, and RNA structure prediction.
SCR is collaborating with several academic groups on applications for its data-mining framework in mass spectrometry data analysis for cancer diagnostics. SCR has joined a growing list of bioinformatics groups eager to test their algorithms in the area of proteomics-derived biomarkers, but Neubauer said that SCR’s approach is faster than most other methods. So far, SCR has applied its approach to a number of publicly available data sets to evaluate its ability to discriminate between cancerous and non-cancerous patients. While acknowledging the many questions that currently surround the effectiveness of proteomics-based cancer diagnostics, Neubauer noted that Siemens is looking to develop a blood-based diagnostic that would be used in combination with CT, MRI, or other methods to improve the sensitivity and specificity of diagnostic tools. It’s often difficult to determine whether lesions are malignant or benign with imaging technology alone, he said, so an approach that combines imaging with a proteomic biomarker test could be of tremendous benefit in such circumstances.
In another project, SCR is using its data-mining platform to determine how well specific features discriminate between two data sets in microarray experiments. SCR is using a publicly available data set to test the performance of its platform in this application, Neubauer said.
Neubauer’s team is also exploring a concept called “predictive maintenance” derived from Siemens’ work in power generation. While ceding that humans are far more complex than turbines, Neubauer said that the company does see promise in methods it has developed to scan streams of incoming data for changes that may indicate a system has reached a critical state “outside the normal operating range.”
SCR’s Integrated Data Systems group, meanwhile, has a number of database development and data integration projects underway. In one project, the team is integrating imaging data with genotypic data to track clinical outcomes. SCR’s Zhao was quick to point out that bioinformatics is still a very new area for his team, and that several projects have only begun in the last few months. Short term, he said, his group sees the most promise for its methods in molecular imaging analysis, but longer-term goals include large-scale integrated healthcare databases and decision support systems.
While it’s clear that SCR researchers are hungry for data as they continue to validate their IT tools in the life science domain, Zhao suggested that feedback from the biological research community may be the most useful data of all. “Even if you point out that we’re going in the wrong direction,” he told the attendees of the IAS systems biology symposium, “that would be very valuable for us.”