论文部分内容阅读
杂环类化合物在卫生保健和药物分子设计领域发挥关键作用,在西药中占有重要地位.本工作针对扩展距离矩阵提出了一组范数指数,基于扩展距离矩阵的范数指数构建了一个新的构效关系模型,并对杂环类化合物二苯并呋喃的芳烃受体亲和性(pEC50)以及芳香和杂环芳香胺的诱导有机体变异力(lnR)进行了计算预测.结果表明,基于扩展距离矩阵范数指数建立的构效关系模型可以很好地预测pEC50和lnR.其中,pEC50预测结果的平均绝对误差(AAD)为0.403,相关性系数r2=0.876,lnR预测结果的AAD为0.702,r2=0.779.与其他预测方法的对比结果表明,本工作不仅能够利用一个完全相同的数学表达模型同时对pEC50和lnR进行预测,而且预测结果在准确性和稳定性上都有显著改善.
Heterocyclic compounds play a key role in the field of health care and molecular design of drugs, and occupy an important position in western medicine.In this work, a set of norm indices for extended distance matrices are proposed, and based on the norm index of extended distance matrices, a new Structure-activity relationship model, and the prediction of the aromatic receptor affinity (pEC50) of the heterocyclic compound dibenzofuran and the induced organism mutation force (lnR) of the aromatic and heterocyclic aromatic amines were calculated and predicted.The results show that based on the expansion The structure-activity relationship model established by the matrix norm index can predict pEC50 and lnR well, among them, the average absolute error (AAD) of pEC50 prediction is 0.403, the correlation coefficient r2 is 0.876, the AAD of lnR prediction is 0.702, r2 = 0.779. Comparing with other prediction methods, this work not only makes it possible to predict pEC50 and lnR simultaneously by using an identical mathematical expression model, but also predicts the accuracy and stability of the prediction results.