Development Of Predictive Regression Model For Study Of Therapeutic Properties Of Chemical Compounds

Regression analysis is an important tool in the computer based drug design practitioner’s toolbox for a number of reasons. First, this method saves lot of time and by using this model, lot of compound can be screened for their therapeutic properties. Secondly, varieties of regression analysis methods are available depending on the nature of the problem being studied. In current project, a regression model “Regression Analysis Software Package (RASTP)” is developed as a measure for Study of Therapeutic Properties of Chemical Compounds. This model will compare the unknown molecule with the set of known molecules with respect to their structural properties and select or reject the given molecule or set of molecule on the basis of correlation and Regression Coefficient parameters. In this study, the known set of molecules, which are anticancer compounds with known biological activity, will be retrieved from the databank. These set of molecules will act as training set during the model building and will also used for validation purpose of the model. From the data obtained from model, it is clear that the important regression analysis parameters for the predictor such as coefficient estimate, standard error; mean R squared error; adjusted R square, etc. are in good accord with the respective parameters for the known set. Hence, “Regression Analysis Software Package (RASTP)” model is a good tool for the calculation of biological activity which in turn is used to predict the therapeutic importance of the chemical compound.