application of pls,osc-pls and ga-pls for a qsar study of inhibitory activity of p38map kinase derivatives and design new compounds for kinas using image processing

The purpose of the multivariate image analysis quantitative structure–activity relationship (mia-qsar) study is to obtain information from a set of descriptors that descriptors are in fact pixels of images of two-dimensional structures of molecules. the genetic algorithm application for selecting pixels and image processing was investigated in the qsar study of protein p38 map kinase compounds. using the information obtained, there is a quantitative relationship between the structure and the ic50. protein p38 map kinase is used in the treatment of malignant tumors that act as inhibitors. in this study, the generation of a model for predicting ic50 of these compounds was considered. to this end, the molecules were first plotted in the chemsketch software and fixed in the same coordinates, and finally the images were processed in the matlab program. quantitative models were obtained with pls, osc-pls and pls-ga methods, and ic50 prediction was performed. the results of a series of statistical parameters such as rmsep, rsep showed the highest predictive power for the pls-ga model.