Internet laboratory for predicting harmful effects triggered by drugs and chemicals--a progress report

The main objective of our institution is to establish a virtual laboratory on the Internet to allow for a reliable in silico estimation of harmful effects triggered by drugs, chemicals and their metabolites. In the past two years, we have compiled a pilot system including the 3D models of five receptors known to mediate adverse effects (the Ah, 5HT(2A), cannabinoid, GABA(A), and estrogen receptor, respectively) and tested them against 280 compounds (drugs, chemicals, toxins). Within this set-up we could demonstrate that our concept is able to both recognise toxic compounds substantially different from those used in the training set as well as to classify harmless compounds clearly as being non-toxic at low-level doses. This suggests that our approach can be used for the prediction of adverse effects of drug molecules and chemicals. It is the aim to provide free access to this 3D data base, particularly to universities, hospitals and regulatory bodies as it bears a significant potential to recognise hazardous compounds early in the development process and withdraw them from the evaluation pipeline. Hence, for substances recognised as hazardous in silico, subsequent toxicity tests involving animal models become obsolete.