insilco study: modeling parameters for prediction of activity of iminochromene derivatives

A quantitative structure activity relationship analysis has been applied to a data set of 34 derivatives of 8-hydroxy-2-iminochromene with inhibitory activities for carbonyl reductase 1. semi-empirical quantum chemical calculations at the am1 level were used to find the geometry of the studied molecules. whole numbers of descriptors were calculated with dragon software, and a subset of calculated descriptors was selected from 407 dragon descriptors with the multiple linear regression (mlr), partial least square and principal component analysis methods. results displayed that the mlr method predicted of activity good enough. the best model of mlr, with seven descriptors was selected. also it indicates very good consistency towards data variations for the validation methods. the predicted values of activities are in suitable agreement with the experimental results. the obtained results suggested that the pls method could be more helpful to predict the biological activity of iminochromene derivatives. this study is be useful to predict the activity of other compounds in the same derivatives