B.D. Gute, G.D. Grunwald and S.C. Basak
A Hierarchical Approach to Predictive Toxicology using Computed Molecular Descriptors
Natural Resources Research Institute, University of Minnesota, 5013 Miller Trunk Highway, Duluth, MN 55811, USA. E-mail: bgute@wyle.nrri.umn.edu
Recently there has been an increased focus on the use of computed theoretical descriptors of molecular structure to model the activity of a wide variety of compounds. This interest is driven by the fact that these theoretical descriptors can be calculated directly from the molecular structure of a compound. Additionally, the lack of available experimental data necessary for the hazard assessment of many of the compounds in commerce today has driven the need for non-empirical methods to characterize molecular structure for use in toxicity estimation and hazard assessment for existing chemicals and in the rational design of new chemicals. Several of our recent studies [1-5] have utilized a hierarchical approach in developing quantitative structure-activity relationships (QSAR) using non-empirical descriptors for a wide range of physicochemical and biological properties. This approach has been followed for several reasons: