论文部分内容阅读
肝细胞生长因子(HGF)/酪氨酸蛋白激酶(c-Met)介导的细胞信号是导致肿瘤细胞产生和转移的主要途径之一。c-Met抑制剂能够阻断HGF/c-Met信号,抑制人类肿瘤的转移发生。本文选用77个结构多样的吡唑啉酮类衍生物作为c-Met抑制剂分子的数据集,随机选取其中16个分子作为检验集,其余作为训练集,采用多元线性回归(MLR)和主成分回归分析(PCA)法对每个分子的648个参数进行回归分析,分别建立定量构效关系的最优预测模型。结果表明,多元线性回归中的逐步筛选法是最佳的建模方法,其所建模型统计结果良好(R2=0.81,SEE=0.37),应用于检验集的结果也较理想(R2=0.83,SEP=0.42),模型可靠性和预测能力较强,能直观反映影响活性的主要因素。此模型的确立有助于指导新型高效c-Met抑制剂药物的筛选和开发。
Hepatocyte growth factor (HGF) / tyrosine protein kinase (c-Met) -mediated cell signaling is one of the major pathways leading to tumor cell production and metastasis. c-Met inhibitors block HGF / c-Met signaling and inhibit the metastasis of human tumors. In this paper, we selected 77 pyrazolone derivatives with various structures as data sets of c-Met inhibitor molecules, randomly selected 16 of them as the test set and the rest as the training set. Multiple linear regression (MLR) and principal components Regression analysis (PCA) method was used to carry out regression analysis on 648 parameters of each molecule to establish the optimal prediction model of quantitative structure-activity relationship. The results showed that the stepwise screening method in multiple linear regression method was the best modeling method, and its model was statistically good (R2 = 0.81, SEE = 0.37), and the result of the test set was also ideal (R2 = 0.83, SEP = 0.42). The model has strong reliability and predictive ability and can directly reflect the main factors affecting the activity. The establishment of this model will help guide the selection and development of new and highly effective c-Met inhibitor drugs.