A contemporary interest in computational toxicology is the prediction of potential toxicity of drugs, industrial chemicals, and xenobiotics using properties of molecules that can be calculated directly from their structure alone without the input of any other experimental data (1-14). This ﬁeld of research is popularly known as quantitative structure-activity relationship (QSAR), where ‘‘activity’’ is sometimes used generally to represent toxicity [quantitative structure-toxicity relationships (QSTR)], physicochemical property [quantitative structure-property relationships (QSPR)], or biological activity [quantitative structure-activity relationships (QSAR)]. QSAR is based on the paradigm expressed by Eq. (1):
P ¼ f ðSÞ ð1Þ where P represents any toxicological, physicochemical, or biological property of interest produced by chemical-biological interactions at any level of biological organization, e.g., molecular (receptor, enzyme), cellular, tissue or organ, whole organism, etc., and S symbolizes quantiﬁer(s) of salient features of molecular structure related to the property of interest P.