《抗结核病药物的3D-QSAR定量结构性质关系研究》-毕业设计（论文）.docVIP

PAGE \* ROMAN IV PAGE \* ROMAN III 抗结核病药物的3D-QSAR定量结构性质关系研究 摘 要 采用三维全息原子场作用矢量(3D-HoVAIF)和基于虚拟蛋白质受体原子探针的分子表面随机采样分析(RaSMS)对25个查耳酮类抗结核病药物进行定量构效关系(QSAR)研究。运用多元线性回归(Multiple linear regression, MLR)及偏最小二乘回归(partial least square regression, PLS)建模，同时采用内部及外部双重验证的办法对所得模型稳定性能进行了深入分析和检验。采用3D-HoVAIF 的MLR及PLS建模的复相关系数(Rcum2)、留一法(leave-one-out, LOO)交互校验(cross-validation, CV)复相关系数(QLOO2)和外部样本校验复相关系数(Qext2)分别为0.984、0.920、0.808和0.998、0.975、0.903；采用RaSMS分别为 0.975、0.914、0.764和0.926、0.905、0.829。结果表明，3D-HoVAIF和RaSMS都有较好表征抗结核药物分子结构信息，因而能建立具有良好稳定性和预测能力的QSAR模型。 关键词：体外抗菌活性，抗结核病药物，三维全息原子场作用矢量，比较分子/虚拟受体相互作用分析法，定量构效关系 Anti-TB drugs by 3D-QSAR Quantitative Structure-Property Relationship Study ABSTRACT A newly developed three-dimensional holographic vector of atomic interaction field (3D-HoVAIF) and Protein-based virtual surface receptor molecule atom probe analysis of random sampling (RaSMS) were used to describe the chemical structures of 25 Chalcone as antituberculosis agents. Here 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 these models were strictly analyzed by both internal and external validations. When use 3D-HoVAIF, the correlation coefficient (R2) of established MLR and PLS， leave-one-out (LOO) cross-validation (CV), predicted values versus experimental ones of external samples were 0.998、0.975、0.903 and 0.926、0.785、0.785, respectively. When use (RaSMS), the results were 0951、0.914、0.829 and 0.928、0.905、0.764. The results indicated that results of PLS here had favorable estimation stability and good prediction capabilities. Satisfactory results showed that 3D-HoVAIF and RaSMS could preferably express information related to biological activity of anti-tuberculosis drug. KEY WORDS: in vitro antibacterial activity, anti-tuberculosis drug, three-dimensional holographic vector of atomic interaction field, quantitative (3D-H