首页> 中文期刊>海峡药学 >大环内酯化合物细胞毒性的定量构效关系研究

大环内酯化合物细胞毒性的定量构效关系研究

     

摘要

OBJECTIVE The article focuses on Macrolide as the research object,build MLR,ANN and RBF model,investigate their structure-activity relationships between Macrolide and cytotoxicity,and confirm that what kind of model is more exact to predict the relationship between the two.METHODS This paper mainly adopted MMFF 94 method,calculated electron structure parameter of 30 Macrolide chemical compounds,constructed comparisons of MLR model,ANN model,and RBF model between its structure and Cytotoxicity effect.RESULTS Screened out the five significant variables which affected cytotoxicity,the correlation coefficient of MLR model equals to 0.870,stand-ard deviation equals to 2332.724.The correlation coefficient of RBF model equals to 0.991,standard deviation e-quals to 50.873.The correlation coefficient of ANN model equals to 0.986 , standard deviation equals to 16.737.CONCLUSION Come to conclusion shows that built ANN model possesses preferable stability and pre-dictable ability.States that molar refractivity and covalent links and cytotoxicity present positively correlated.Partition coefficient,molecular topology parameter and excluded volume of connery solvent and cytotoxicity present negatively correlated.%目的:以大环内酯化合物为研究对象,建立MLR、ANN和RBF模型,研究大环内酯和细胞毒性的构效关系,并确立哪种模型对二者之间的预测更准确。方法采用分子力学MMFF94方法,计算了30个大环内酯化合物的电子结构参数,建立其结构和细胞毒性之间的MLR模型、ANN模型、RBF模型对比。结果筛选出影响细胞毒性的显著的5个变量,建立的 MLR 模型的相关系数 R =0.870,标准误差是 SE =2332.724;RBF模型的相关系数是0.991,标准误差是50.873。 ANN模型的相关系数是0.986,标准误差是16.737。结论采用神经网络建立的模型具有更好的稳定性和预测能力,摩尔折射率和共价链接与细胞毒性呈现正相关,分配系数、分子拓扑指数和康纳利溶剂排除体积与细胞毒性呈现负相关。

著录项

相似文献

  • 中文文献
  • 外文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号