ETOX ProjectETOX ProjectETOX ProjectETOX ProjectETOX ProjectETOX ProjectETOX Project
  Home > Project > Summary and Objectives
ETOX Project Print page ETOX Project Send page


General Information


Summary & Objectives


Background & Concept


Potential Impact of Project Results


Project Activities


ETOX Project line

ETOX Project


ETOX Project

ETOX Project


ETOX Project



Project Summary & Objectives


In vivo toxicological studies carried out by means of non-GLP and GLP animal experiments represent a key step in drug development, since they often unveil straight chemistry-linked toxicities or mechanisms such as cross-target activations or inhibitions related to side effects of drugs. Early in silico prediction of in vivo toxicological outcomes would increase the quality of drug candidates and ensure a lower attrition rate during subsequent phases of the drug development pipeline. This would also reduce the number of animals (3Rs) to be used in preclinical studies.


The in silico prediction of biological phenomena on the basis of structural information of the drug candidates shows attractive results in the case of some molecular pharmacology parameters (e.g., affinities for particular biological targets), in some pharmacokinetics-related properties, or for a few toxicological endpoints such as mutagenicity. Nevertheless, the in silico prediction of most of the important in vivo outcomes on the basis of information available in early stages of the drug development is still far from being a reality.


In order to move towards a predictive and reliable modelling of the complex relationships existing between in vivo observations of the toxicity and safety of drug candidates, the results of their in vitro testing, and their molecular structures, we need to address the following challenges:



  • An integrated exploitation of data and knowledge accumulated in public (biomedical literature, freely-accessible databases) and private (in-house databases and legacy reports of pharmaceutical industry) resources. Moreover, the integration of the information supplied by pharmaceutical companies poses the further challenge of the secure coding of sensitive data.

  • The development of multi-scale and multi-level modelling techniques able to deal with the complex relationships existing between the biomolecular processes and the resulting in vivo outcomes.

  • The joint and well coordinated efforts of a multi-disciplinary team of experts.



This project aims to develop innovative methodological strategies and novel software tools able to predict the in vivo toxicology of new molecular entities by means of information that is available in early stages of the drug development pipeline. This will be achieved by jointly storing and exploiting private data from the participating European Federation of Pharmaceutical Industries (EFPIA) and Associations EFPIA companies, as well as publicly available data, and by coordinating the efforts of specialists from EFPIA pharmaceutical companies, relevant SMEs and academic institutions. The proposed strategy includes a synergetic integration of innovative approaches in the following areas:



  • Database building and management, including procedures and tools for protecting sensitive data.

  • Ontologies and text mining techniques, with the purpose of facilitating knowledge extraction from legacy preclinical reports and biomedical literature.

  • Chemistry and structure-based approaches for the molecular description of the studied compounds, as well as of their interactions with the anti-targets responsible for the secondary pharmacologies.

  • Prediction of DMPK features since they are often related to the toxicological events.

  • Systems biology approaches in order to cope with the complex biological mechanisms which govern in vivo toxicological problems.

  • Computational genomics to afford the inter-species and inter-individual variabilities that complicate the interpretation of experimental and clinical outcomes.

  • Sophisticated statistical analysis tools required to derive the inevitably highly-multivariate QSAR models.

  • Development and validation (according to the OECD principles) of QSARs, integrative models, expert systems and meta-tools.



The project will facilitate as much as possible the accessibility to the resulting non-confidential part of the database and the predictive systems by the broader scientific community.

To carry out the present project we propose a consortium that includes top experts in toxicology, database management, bioinformatics, chemoinformatics, biostatistics and software development, all of them showing a broad experience in industry-academia collaborations and international partnerships.