Predictive Toxicology

Predictive toxicology is concerned with the development of new approaches to forecast the toxic effects of chemicals without the need for animal testing. Within the group we are interested in devising so called in silco models, i.e. models that predict potential toxic effects purely based on computer models. The field of predictive toxicology is closely related to graph mining (link 2.2), pattern mining (link 1.3), and cheminformatics (link 3.1).

Within the group, we have participated in a number of project like BmBF REACH (link) and the EU project OpenTox (link). The overall objective of OpenTox was to develop a framework that provides a unified access to toxicityOpenTox logodata, (Q)SAR models, procedures supporting validation, and additional information that helps with the interpretation of (Q)SAR predictions. The OpenTox framework has been developed as an open source project to optimize the dissemination and impact, to allow the inspection and review of algorithms and to attract external contributors. We closely collaborated with related projects and authorities to agree on common standards and to avoid duplicated and redundant work. The project established as an international research collaboration between groups from Switzerland, Bulgaria, Italy, Greece, Russia, India, USA, and Germany.