Use of quantitative shape-activity relationships to model the photoinduced toxicity of polycyclic aromatic hydrocarbons: Electron density shape features accurately predict toxicity

The quantitative shape-activity relationship (QShAR) methodology, based on accurate three-dimensional election densities and detailed shape analysis methods, has been applied to a Lemna gibba photoinduced toxicity data set of 16 polycyclic aromatic hydrocarbon (PAH) molecules. In the first phase of our studies, a shape fragment QShAR database of PAHs was developed. The results provide a very good match to toxicity based on a combination of the local shape features of single rings in comparison to the central ring of anthracene and a more global shape feature involving larger molecular fragments. The local shape feature appears as a descriptor of the susceptibility of PAHs to photomodification and the global shape feature is probably related to photosensitization activity.