QSTR with extended topochemical atom (ETA) indices. 11. Comparative QSAR of acute NSAID cytotoxicity in rat hepatocytes using chemometric tools

Abstract

An attempt was made to develop quantitative structure–toxicity relationships (QSTRs) for acute rat hepatocyte cytotoxicity of a series of non-steroidal anti-inflammatory drugs (NSAIDs) with extended topochemical atom (ETA) indices. The ETA models were compared with those developed with a pool of other topological indices. Finally, attempt was made to develop models from the combined pool of topological (ETA and non-ETA) descriptors. The chemometric tools used for model development were multiple linear regression with factor analysis as the variable selection tool, stepwise regression, principal component regression analysis, partial least squares (PLS), genetic function approximation (GFA) and genetic PLS (G/PLS). Use of ETA descriptors along with non-ETA ones increased the statistical quality of the non-ETA models in different techniques except stepwise regression and GFA. The best three models came from GFA, G/PLS and PLS techniques (leave-one-out Q 2 values of 0.871, 0.854 and 0.834, respectively, using combined set of descriptors except for GFA). Use of the ETA parameters suggests that the toxicity increases with bulk and degree of branching. Moreover, heteroatom count and degree of unsaturation are also important for the toxicity.