Recent and continuing technological advances are producing large amounts of disparate data about cell structure, function and activity. This is driving the development of tools for storing, mining, analyzing, visualizing and integrating data. This proposal describes the VisANT system: a tool for visual data mining that operates on a local database which includes results from our lab, as well as automatically updated proteomics data from web accessible databases such as MIPS and BIND. In addition to accessing its own database, a name normalization table (i.e. a dictionary of identifiers), permits the system to seamlessly retrieve sequence, disease and other data from sources such as GenBank and OMIM. The visualization tool is able to reversibly group related sets of nodes, and display and duplicate their internal structure, providing an approach to hierarchical representation and modeling. We propose to build further on these unique features by including capabilities for mining and representing chemical reactions, orthologous networks, combinatorially regulated transcriptional networks, splice variants and functional hierarchies. Software is open source, and the system also allows users to exchange and integrate the networks that they discover with those of others.
