Text extraction: Agilent releases open-source literature search tool for Cytoscape

PALO ALTO, Calif.—Pharmaceutical researchers and others found themselves with a new tool in late March when Agilent Technologies—a provider of technologies for life sciences, chemical, communications and electronics analysis—released a free automated literature search plug-in for Cytoscape version 2.1.

Cytoscape itself is an open-source bioinformatics platform that allows researchers to form a visual map of complex biological networks. The Agilent Literature Search plug-in is designed to automatically search for and extract information from multiple textual databases, such as those of the U.S. Patent Office and the National Center for Biotechnology Information (PubMed).

"Agilent is committed to accelerating the productivity of biologists," says Jim Hollenhorst, director of the Molecular Technologies Laboratory for Agilent Laboratories. "This literature search tool was developed to facilitate our own research collaborations and has proved very successful for us."

That success is something that Agilent wanted to make available to others but wasn't comfortable selling.

"Agilent believes in standards rather than just proprietary formations. We wanted to share it with other researchers, and Cytoscape seemed to be the emerging standard for biological network representation," says Annette Adler, project manager for the systems biology effort in the Molecular Technology Laboratory.

The plug-in pulls information from the various database sources and represents it within Cytoscape as a map of the relationships between the biological entities, such as genes and proteins.

So, for example, a researcher could enter the names of all genes of interest, along with keywords that describe the types of relationships being looked for. The plug-in would search accessible databases and generate a computational representation of gene/protein associations grouped into a network that can be viewed and manipulated within Cytoscape.

"PubMed alone has something like 15 million abstracts," observes Aditya Vailaya, a researcher involved with the plug-in project. "In the past, researchers might be dealing with tens of molecules or genes—now they have to deal with tens of thousands. You can know a lot about tens of molecules, but you cannot know details of 20,000 or 30,000 of them. So, to find what is of interest to your research, you need a tool like this to bring into context what all these molecules you don't know about can do."