Subhash C. Basak, Gregory D. Grunwald and Brian D. Gute

Predicting mode of action toxicity from structure: A hierarchical approach.

Natural Resources Research Institute, University of Minnesota - Duluth, 5013 Miller Trunk Highway, Duluth, MN 55811, USA. E-mails: sbasak@wyle.nrri.umn.edu, ggrunwal@wyle.nrri.umn.edu, bgute@wyle.nrri.umn.edu

In a recent study [1], molecular similarity and discriminant analysis was used to predict acute toxic modes of action (MOA's) for a set of 283 chemicals. These analyses relied on the use of nonempirical 2-D parameters, such as topological indices and atom pairs, as structural descriptors for use in these models. These models could classify 65% to 95% of the 283 chemicals to the appropriate MOA. This study will present the results of hierarchical analyses in which geometric and quantum chemicals parameters are used in addition to the 2-D parameters. The MOA's, previously determined by toxicodynamic studies in fish, include narcotics, electrophiles / proelectrphiles, uncouplers of oxydative phorphorylation, acetylcholinesterase inhibitors and neurotoxicants.

Back to Program Page

Back to Main Page