Celera Genomics has signed two new agreements to enhance the functionality of the Celera Discovery System. The company entered into a multi-year marketing and distribution agreement for the ModBase protein structure database from Structural GenomiX and also signed a non-exclusive license for Virtual Genetics Virtual Adapt bioinformatics text mining technology.
The agreements kick off a series of third-party relationships that Celera is building to ramp up its CDS literature and proteomics products.
ModBase, developed by Andrej Sali and his lab at Rockefeller University and exclusively licensed by SGX, contains computationally modeled structures for domains in over 310,000 proteins. According to SGX, these proteins correspond to 58 percent of all publicly available protein sequences.
What we have today in CDS is access to the PDB [Protein Data Bank}, but ModBase provides quite a bit beyond that, said Tony Kerlavage, senior director of bioinformatics and product strategy at Celera. The added value here is a computational prediction of structure based upon known structures using Sali''s techniques.
But perhaps of even more value to CDS users will be the literature search capability provided by Virtual Adapt. Jack Robinson, managing director of Stockholm-based Virtual Genetics Laboratory, said that the contract with Celera provides initially for the integration and customization of Virtual Adapt into the CDS for searching through MedLine, and even through other databases in the future.
The value of the order is $400,000, including professional services and support, but Robinson said that the benefits the company will see by having its product integrated into CDS will be far more valuable in the long run. Were a small company only 20 people so this is a very important contract for us, Robinson said. He noted that a number of other companies, such as Lion Bioscience, have really benefited from being Celera providers.
Several companies, including Lion, are working on text mining technology of their own, but Robinson said that Virtual Adapt has a number of advantages over potential competitors.
Virtual Adapt is based on similarity searching, which uses a number of statistical models to pull relevant information out of text. Most of our competitors do not have that kind of advanced similarity searching complemented with classification and annotation tools, and certainly not at the speed that we can do it, said Robinson. I think were ahead of the game.
Celeras director of internet technology, Raymond Cyrus, noted that the company considered other text mining vendors but chose Virtual Genetics because they have a lot of experts on staff that understand the biological space.
Cyrus said the work with Virtual Genetics is the initial step of a three-phase effort to create a complete literature package. the ultimate goal, Cyrus said, will be the creation of derivative products that summarize the most salient facts within relevant journal articles. Celera is currently in the process of integrating each of the new features within CDS.