Prediction of environmental toxicity based on structure-activity relationships using mechanistic information

Quantitative structure-activity relationships (QSAR) form a powerful tool in predictive toxicology. The application of predictive models, however, needs a very strict definition of the scope and limitations of such models. Therefore, it also of importance that predictive models are understood from a mechanistic point of view. In this paper, QSAR studies for two important parameters in environmental risk assessment will be discussed, namely for bioconcentration and aquatic toxicity. Emphasis will be given to a mechanistic interpretation of the models and to their practical applicability.