|
Quantitative structure activity relationship (QSAR), which is to investigate the quantitative relationship between the molecular structural parameters and biological activities or other relative activities, has got a wide and rapid development in forecasting toxicity of organic pollutants and drug discovery etc. Molecular structural parameterization, as an important technique, is the key step in the QSAR studies. Firstly, topological indices can be derived directly from molecular structure without experiments, so they can be applied in QSAR as a simple, direct and effective molecular structure parameterization method. Secondly, quantum chemistry, an important method in studying molecular structure and reaction theory, has been widely applied in QSAR, thus greatly increasing the accuracy of QSAR theory. Thirdly, compared to 2D descriptors, 3D descriptors in QSAR is a more accurate technique in structure identification because 3D descriptors will indicate non-bonding interactions of ligand-receptor. In the dissertation, we developed a series of novel 2D and 3D descriptors based on the basic molecular structure character, considering common intramolecular and intermolecular non-bonding interactions, like electrostatic interaction, steric interaction, and hydrophobic interaction. Molecular structure parameterization methods and modeling methods have been investigated and applied in QSAR as simple, direct and effective molecular structure parameterization methods. At the same time, the quantitative relationships of selected organic pollutants and several representative drug structures have been built. The results will provide some useful basic information for analyzing biological activities of molecular, function, reaction mechanism, drug design, and deficiency of medicine exploitation.A series of new molecular structure parameterization methods, which had been new developed in our group were used to describe the chemical structures of the selected organic pollutants and representative drug. Here some quantitative structure activity relationship (QSAR) models were built by multiple linear regression (MLR) and partial least square regression (PLS). The estimation stability and generalization ability of the models were strictly analyzed by the correlation coefficient (Rcum), leave-one-out (LOO) cross- validation (RCV), predicted values versus experimental ones of external samples (Qext). The results showed that the new descriptors could preferably express information of molecular structure, which had both favorable estimation stability and good prediction capabilities.In Chapter 1 of this dissertation, not noly a brief description of QSAR development process and research status was given, and also expatiate on the study objective and significance of this dissertation.In Chapter 2 of this thesis, both the structural parameters and research methods in QSAR study were outlined. Including the theory definitions, calculation methods, and characteristics of each descriptor of the molecular electronegativity-distance vector (MEDV), molecular electronegativity–interaction vector (MEIV), generalized correlative index (GCI), 3D holographic vector of atomic interaction field (3D-HoVAIF), et al. In addition, the methods and principle of multiple linear regression (MLR) and partial least squares (PLS) were summarized in this chapter.In Chapter 3 of this thesis, research process and results of the chlorinated phenols and chlorinated benzenes, polychlorinated dibenzofurans, substituted aromatic hydmcarbons, et al, which are organic pollutants and 5-pheny-l -phenylamino-1H-imidazole, which is representative anti-HIV drug were summarized in this chapter.In Chapter 4 of this thesis, these of study contents, innovation and insufficiency overall this dissertation were summarized. Thus, a conclusion about the further development and ideas based on the summarizations is given.
|