论文部分内容阅读
用F统计结合偏最小二乘算法作为蚁群算法选择变量的目标函数,对非竞争性AMPA受体拮抗剂2,3-苯并二氮杂-4(硫)酮和1,2-酞嗪及异喹啉衍生物的定量构效关系进行了研究。所得的结果与多元线性回归、全变量偏最小二乘构效关系模型进行了比较,结果显示改进的蚁群优化算法可以得到更好的预测结果。指出了影响2,3-苯并二氮-4(硫)酮和1,2-酞嗪及异喹啉衍生物抑制活性的主要描述符,可以为合成高活性药物结构提供理论指导。
F-statistic combined with partial least squares algorithm was used as the objective function of the ant colony algorithm to select variables. The non-competitive AMPA receptor antagonists 2,3-benzodiazepine-4 (thio) ketone and 1,2-phthalide The quantitative structure-activity relationship of oxazine and isoquinoline derivatives was studied. The obtained results are compared with multivariate linear regression and full-variable partial least squares construction-effect-relation model. The results show that the improved ant colony optimization algorithm can get better prediction results. The main descriptors affecting the inhibitory activity of 2,3-benzodiazepine-4 (thio) ketone, 1,2-phthalazine and isoquinoline derivatives are pointed out, which can provide theoretical guidance for the synthesis of highly active drug structures.