摘要
基于分子电性距离矢量(Mt)表征21个芳胺喹唑啉衍生物的分子结构,并与其对人胃癌细胞的抗癌活性(pI)关联。通过最佳变量子集回归方法建立上述化合物抗胃癌活性的三参数(M15、M18、M82)定量构效关系模型。其交叉验证系数(Rcv2)、非交叉验证系数(R2)依次为0.386和0.737。通过R、Rcv2、VIF等检验,模型具有较好的相关性、稳健性和预测能力。结果显示,-CH2-、-CH<、-O-、-S-和-NH-等分子结构单元直接影响这些化合物的抗胃癌活性。将M15、M18、M82作为人工神经网络的输入层结点,采用3∶3∶1的网络结构,利用BP算法获得了令人满意的pI模型,其R2和标准偏差SD分别为0.992和0.080,表明pI与三参数呈现优异的非线性关系。
The structures of 21 arylamine quinazoline derivatives were characterized on the basis of a molecular electronegativity distance vector(Mt),which were related to anti-gastric cancer activity(pI)to human gastric cancer cells.The three-parameter(M15,M18,M82)QSAR model was established by using Leaps-and-Bounds regression analysis for pI of above compounds along with the Mt.The coefficients of the cross-validation(Rcv2)and non cross-validation(R2)for the pI model were 0.386 and 0.737,respectively.The QSAR model had better correlations,robustness,and prediction capability by using R,Rcv2,and VIF tests.The results showed that the molecular structural units-CH2-,-CH<,-O-,-S-and-NH-are the main factors that directly affect the loss of traction activity of these compounds.When M15,M18,M82 were used as the input neurons of artificial neural network and the 3∶3∶1 network architecture was employed,a satisfactory BP-pI model were constructed with the back-propagation algorithm.Its correlation coefficient(R2)and the standard error(SD)were 0.992 and 0.080,respectively,indicating that pI had an excellent nonlinear correlation with three structural parameters.
作者
杨杰元
杨雪颖
杨沛艳
冯惠
冯长君
Yang Jieyuan;Yang Xueying;Yang Peiyan;Feng Hui;Feng Changjun(School of Material&Chemical Engineering,Xuzhou University of Technology,Xuzhou,221018)
出处
《化学通报》
CAS
CSCD
北大核心
2021年第8期853-856,846,共5页
Chemistry
基金
国家自然科学基金项目(21075138)
结构化学国家重点实验室开放基金(2016003)
江苏省大学生创新创业训练项目(xcx2020143)
徐州工程学院大学生创新创业基金项目(201913)资助。
关键词
芳胺喹唑啉衍生物
抗胃癌活性
电性距离矢量
人工神经网络
定量构效关系
Arylamine quinazoline derivative
Anti-gastric cancer activity
Electronegativity distance vector
Artificial neural network(ANN)
Quantitative structure-activity relationship(QSAR)
