Related Experiment Videos
Physicochemical and graph theoretical descriptors in developmental toxicity SAR: a comparative study.
O T Macina1, N B Sussman, H G Claycamp
1Department of Environmental and Occupational Health, Graduate School of Public Health, University of Pittsburgh, 260 Kappa Drive, Pittsburgh, PA 15238 USA. macina+@pitt.edu
SAR and QSAR in Environmental Research
|May 1, 2001
Summary
Physicochemical and graph theoretical parameters can predict developmental toxicity in mice. Consistent features were identified, showing equal utility for both parameter types in structure-activity relationship modeling.
Area of Science:
- Toxicology
- Computational Chemistry
- Cheminformatics
Background:
- Chemicals can harm developing fetuses, causing various adverse outcomes.
- Understanding structure-activity relationships (SAR) is crucial for predicting and preventing developmental toxicity.
Purpose of the Study:
- To investigate the utility of physicochemical and graph theoretical parameters for SAR modeling of developmental toxicity.
- To assess the consistency of features identified across different modeling approaches.
Main Methods:
- Utilized data from the Chernoff-Kavlock in vivo mouse assay for developmental toxicity.
- Calculated physicochemical (electronic, steric, transport) and graph theoretical (indices) parameters.
- Employed linear discriminant analysis, decision trees, and neural networks with random sampling for model building and validation.
Main Results:
- Identified consistent features predictive of developmental toxicity.
- Demonstrated that both physicochemical and graph theoretical parameters offer comparable utility in SAR modeling.
- Found no significant advantage of specific model-building procedures in predictive performance.
Conclusions:
- Physicochemical and graph theoretical parameters are valuable tools for predicting developmental toxicity.
- Consistent SAR features can be identified using computational approaches.
- Both parameter types exhibit equal potential for developing predictive models of developmental toxicity.