Data management and processing in toxicoinformatics - from chemical databases to automatic extraction of unstructured resources

Toxicoinformatics needs electronic access to all available information about toxicological substances and their influences on biological systems. Compound information such as physicochemical properties and pharmacological data is of utmost importance. In the past, mainly commercial information providers supplied access to this chemical information. Now, increasing amounts of this type of information are becoming available in public databases. A number of representation standards have been developed for the unique representation of chemical compounds, and an increasing number of databases have collected different chemical and toxicological information. Despite these efforts, most information about toxic substances and their relationships is electronically available only in unstructured text resources. This makes it necessary to extract the knowledge from those resources to make them available in an electronic readable format for further processing. Thus, in addition to an overview of chemical representation and resources, short overviews of the current status of information extraction techniques are given and a suitable network representation for compound-target interactions is described. Examples are provided to illustrate the capability of these methods.