Bakhtiyor Rasulev - QSAR modeling of acute toxicity on mammals caused by aromatic compounds: The case study using oral LD50 for rats

Version 1

      Publication Details (including relevant citation   information):

      Rasulev B., Kusić, H., Leszczynska D., Leszczynski J., Koprivanac   N., Journal of Environmental Monitoring, 12 (5), pp. 1037-1044


      The goal of the study was to predict  toxicity in vivo   caused by aromatic compounds structured with a single    benzene ring and the presence or absence of different substituent   groups  such as hydroxyl-, nitro-, amino-, methyl-,   methoxy-, etc., by using  QSAR/QSPR tools. A Genetic   Algorithm and multiple regression analysis  were applied to   select the descriptors and to generate the correlation    models. The most predictive model is shown to be the 3-variable   model  which also has a good ratio of the number of   descriptors and their  predictive ability to avoid   overfitting. The main contributions to the  toxicity were   shown to be the polarizability weighted MATS2p and the    number of certain groups C-026 descriptors. The GA-MLRA approach   showed  good results in this study, which allows the   building of a simple,  interpretable and transparent model   that can be used for future studies  of predicting toxicity   of organic compounds to mammals. © 2010 The Royal  Society   of Chemistry.

      Address (URL):