• Toxicology data;
  • Public resources;
  • Data quality;
  • Computational models;
  • Predictive models


(1000–1500 characters) In spite of the increasing amount of public access resources that offer original data related to drug toxicology, the successful exploitation of such data for the development of in silico predictive models is still limited by the quality of the data available, its integrability and its coverage for each toxicity endpoint. This work describes the strategy developed by the IMI eTOX consortium for identifying and compiling data and other related resources from the biomedical literature and a wide spectrum of public on-line sources. The main result of this effort is a large web-based structured library containing links to articles of toxicological relevance (data that can be used for modeling purposes, computational models, and toxicity mechanisms), public databases, standardized vocabularies and modeling tools. All this material has been manually reviewed, systematically evaluated and grouped into different categories. The library has been made public at the eTOX website (, where it is updated on a monthly basis, constituting a useful resource for affording the in silico toxicity prediction of novel drug candidates.