Application Of QSAR Studies In Nervous System Drug And Anti-HIV Drug

Quantitative structure activity relationships (QSAR) methods have been widely used in chemistry, environmental chemistry and drug design. QSAR is the most promising tools to provide some useful meaning for the prediction of the activities properties of compounds that have not been synthesized.In our dissertation, we researched the application of QSAR studies in nervous system drug and anti-HIV drug. In addition, we mainly discussed the support vector machine (SVM) and 3D-QSAR methods.In Chapter 1, the dissertation included a brief description of the history, principle, methods and applications of QSAR. Much emphasis was put on the introduction of SVM,3D-QSAR methods and the realization process of QSAR.In Chapter 2, we concluded the application of 2D-QSAR in the nervous system drug. A brief description was given below:(1) A QSAR study was performed on fatty acid amide hydrolase (FAAH) inhibitors. Heuristic method (HM) and support vector machine (SVM) were used to build QSAR models. The root mean square error (RMSE) for the training set given by HM and SVM were 0.555 and 0.404, respectively, which shows the performance of SVM model is better than that of the HM model. This paper provides a new and effective method for predicting the activities of FAAH inhibitors.(2) SVM was used to develop a non-linear QASR model for the prediction of the activities of the adenosine A2A receptor antagonists. The results obtained by SVM were compared with those obtained by HM. The root mean squared errors (RMSE) for the training set given by HM and SVM are 0.291 and 0.223, respectively, which shows the performance of SVM model is better than that of the HM model.Chapter 3 described the application of 3D-QSAR in anti-HIV drug. A brief introduction was given as follows:(1)A series of human immunodeficiency virus type 1 (HIV-1) attachment inhibitors were subjected to 3D-QSAR studies. The CoMFA model yielded satisfactory statistical data with the cross-validated q2 value of 0.589. The cross-validated q2 value of CoMSIA model was 0.621. From the cross-validated results, it can be seen that the CoMSIA model has a better predictive ability than CoMFA model. Based on the above results, the CoMFA and CoMSIA analyses can be used in the design of more potent HIV-1 attachment inhibitors.(2) A series of human immunodeficiency virus type 1 (HIV-1) integrase inhibitors were subjected to three-dimensional quantitative structure-activity relationship (3D-QSAR) studies. Statistically reliable models were obtained with good predictive power. The CoMFA model includes steric and electrostatic fields for the training set with the cross-validated q2 value of 0.67 and the noncross-validated R2 value of 0.98. The cross-validated q2 value of CoMSIA Model is 0.76 and the noncross-vaildated R2 value is 0.99. From the cross-validated results, it can be seen that the CoMSIA model has a better predictive ability than CoMFA model. Based on the above results, the CoMFA and CoMSIA analyses can be used in the design of more potent HIV-1 integrase inhibitors.