Modeling the toxicity of chemicals to Tetrahymena pyriformis using molecular fragment descriptors and probabilistic neural networks

The results of an investigation into the use of a probabilistic neural netw ork (PNN)-based methodology to model the 48-h ICG50 (inhibitory concentrati on for population growth) sublethal toxicity of 825 chemicals to the ciliat e Tetrahymena pyriformis are presented. The information fed into the neural networks is solely based on simple molecular descriptors as can be derived from the chemical structure. In contrast to most other toxicological model s, the octanol/water partition coefficient is not used as an input paramete r, and no rules of thumb or other substance selection criteria are employed . The cross-validation and external validation experiments confirmed excell ent recognitive and predictive capabilities of the resulting models and rec ommend their future use in evaluating the potential of most organic molecul es to be toxic to Tetrahymena.