PW03 QSAR
Exhibit Hall
8:00 AM - Wednesday
(PW032) The TIMES mutagenicity model: Performance and validation.
Dimitrova, N1, Kotov, S1, Jacob, E2, Mekenyan, O1, 1 Laboratory of Mathematical Chemistry; University "Prof. As. Zlatarov", Bourgas, Bourgas, Bulgaria2 BASF, Department of Product Safety , Regulations, Toxicology and Ecology, D-67056, Ludwigshafen, Germany
ABSTRACT- A mutagenicity model for the Ames TA100 and TA1535 was recently derived taking into account metabolic activation of chemicals and clarifying the anticipated interaction mechanisms. The model developed as a decision tree combined alerting groups with 3D structural requirements assessing their activation witin specific molecular environment. The database with 1200 structurally-diverse chemicals tested by the National Toxicology Program has been used to build the model including chemicals that tested positive without rat liver enzymes (S-9), chemicals that tested positive with rat liver enzymes and inactive chemicals. The model was developed by first comparing the reactivity profile of chemicals that were positive without rat liver enzymes to the reactivity profile of the remaining chemicals. A series of hierarchically ordered metabolic transformations were used to simulate a rat liver metabolism. The simulator was applied to chemicals in the training set that are positive only in the presence of S-9 and the rest of inactive chemicals. The potential metabolites were then screened through the mutagenicity model to identify reactive mutagens. The model explained the variation of mutagenicity data with accuracy comparable with experimental error. The performance of the model has been evaluated in an external validation exercise using 49 chemicals from the BASF proprietary database. Two chemicals only were found to be out of the model applicability domain. Four chemicals, representing azo compounds, had been tested negative without the Prival modification of the S-9 mix and therefore were not included in the score. For the rest of 43 chemicals the model showed 81.4 % concordance (with rates of false negatives and positives each 9.3 %).
Key words: metabolic activation, mutagenicity, TIMES, validation