External test set validation provided correlation coefficient values of 0.708 and 0.706 for multiple linear regressions and partial least squares analysis respectively. A comparable partial least squares model with correlation coefficient (0.723) and neural model with correlation coefficient (0.731) indicated good internal predictability of the model. For partial least squares, a statistical significance value of 0.988 and a fraction of variance explained 0.723 were observed. The most significant model with n=179, regression coefficient (0.851), correlation coefficient (0.725), cross-validated correlation coefficient (0.709), standard error (0.349), Fischer statistic value (114.706) was developed using multiple linear regression analysis. 179 compounds generated a final quantitative structure-activity relationship model with the leave-out one row method of cross-validation to evaluate the predictive ability of the model. The study was performed using 268 compounds (data set) by division into training and test set by the random selection method. Attempts have been made to derive and comprehend a correlation between biological activity (dependent variable) and descriptors (independent variables). Two dimensional-quantitative structure-activity relationships have been performed on a series of tetrahydropyrido-pyrazole nucleus using TSAR 3.3. The anti-arthritic activity of cathepsin S enzyme has been acknowledged and regarded as an emerging target for the development of novel therapeutic agents for the treatment of various autoimmune disorders and other inflammatory diseases. DOI: 10.36468/pharmaceutical-sciences.1010 Abstract
0 kommentar(er)
