Unlike some mainstream IT companies who have recently entered the bioinformatics market, Endeca isn’t aggressively seeking business in the sector. In fact, the Cambridge, Mass.-based search engine developer said it’s got plenty of work providing the navigation architecture that supports the online catalogs of Tower Records and other clients.
But Pete Bell, product marketing manager for Endeca’s Guided Navigation solution, said the company’s location “in the heart of bioinformatics land” led a number of local universities and biotech companies familiar with Endeca’s technology directly to its doorstep.
Endeca recently signed a deal with the chemical engineering department at Massachusetts Institute of Technology to deploy a custom solution for microarray analysis. The company’s Guided Navigation technology will be modified to combine data from microarray experiments with data from SwissProt and GenBank. Endeca has also been working with “a Harvard-affiliated cancer research center” on microarray analysis solutions.
Acknowledging that the online catalog business is a “far cry” from bioinformatics, Bell noted that each field presents the same challenge of extracting valuable information from large amounts of structured and unstructured data. “One approach is data mining — clustering and statistical analysis. Another option is to give ad hoc access to the data. Our product is in the latter category,” said Bell.
Vinay Mohta, a senior scientist at Endeca, said the company’s technology is especially well-suited for the high-dimensional data encountered in bioinformatics. Rather than providing a list of keyword-based hits like most search engines, Mohta said the Guided Navigation approach is able to compute all the dimensions in a data set in real time in order to index all possible relationships within the data. Once the data is indexed once, he said, querying across massive data sets is faster and more productive.
Endeca is also developing data mining tools to complement its navigation technology. Daniel Tunkelang, a senior scientist at the company, recently delivered a paper on a clustering algorithm using the relationship-based approach that drives Endeca’s navigation technology at the SIAM International Conference on Data Mining in Arlington, Va., April 13. The data mining tools are “exploratory work,” Tunkelang said, but noted that his talk generated a great deal of interest from biotech companies in attendance.
While Endeca has no plans to get out of the online catalog business, Bell said that “based on the number of queries that we’re getting,” bioinformatics will most likely make up a growing segment of its business in the future.