摘要
基于Kier和Hall的分子连接性指数及邻接矩阵,定义新型分子连接性指数mKvt,并计算了30种取代苯酚和取代苯甲酸的分子连接性指数。经最佳变量子集回归建立了25种有机污染物生化需氧量(BOD)与2Kpv、5Kpv的定量结构-生物降解相关性(QSBR)模型,该模型判定系数R2、逐一剔除法(LOO)的交互验证系数Q2及Kubinyi函数(FIT)分别为0.818、0.776和3.410,该模型具有高度的稳定性及良好预测能力。据此模型可知,影响有机污染物BOD的主要因素是取代基的电子效应以及分子的柔韧性、折叠程度等空间因素。将这2个结构参数作为人工神经网络的输入层结点,采用2∶5∶1的网络结构,利用BP算法获得了一个令人满意的QSBR模型,其R2和标准偏差s分别为0.967和3.688,表明BOD与2Kpv、5Kvp具有良好的非线性关系。由此可见,新建的连接性指数对有机物生化需氧量的表征是合理有效的,可望在物质构效关系研究中获得广泛应用。
On the basis of the revision of Kier and Hall＇s molecular connectivity index and conjugation matrix, a novel molecular connectivity index ^mKt^v is defined and calculated for 30 substituted phenol and benzoic acid molecules in this paper. The quantitative structure-biodegradability relationship （QSBR） model between biochemical oxygen demand （BOD） and ^2Kp^e,^5Kp^v for 25 organic pollutants among above molecules is developed by Leaps-and-Bounds regression （LBR）, the traditional correlation coefficient R^2, the cross-validation correlation coefficient Q^2 of leave-one-out （LOO） and Kubinyi function （FIT） are 0. 818, 0. 776 and 3. 410, respectively. The result demonstrates that the model is highly reliable and has good predictive ability from the point of statistics. The model shows that the dominant influencing factors of BOD are the electronic effects of substituents and the space factors of molecule： the flexibility and the puckered degree of molecules. The two structural parameters are used as the input neurons of the artificial neural network, and a 2： 5：1 network architecture is employed. A satisfying QSBR model can be constructed with the back-propagation algo- rithm, with the correlation coefficient R^2 and the standard errors being 0. 967 and 3. 688, respectively, showing that therelationship between BOD and the two structural parameters has a good nonlinear correlation. The results show that the new molecular connectivity index have good rationality and efficiency for the biochemical oxygen demand of organic compounds. It can be expected that the ^mKt^v will be used widely in quantitative structure-property/activity relationship research.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2009年第5期700-706,共7页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金(20776149)
环境模拟与污染控制国家重点联合实验室基金(KJ2007001)
江苏省高校自然科学基金(08KJD610003)
徐州工程学院培育课题(XKY2008313)
关键词
分子连接性指数
取代苯酚
取代苯甲酸
生化需氧量
人工神经网络
定量结构-生物降解相关性
molecular connectivity index
substituted phenol
substituted benzoic acid
biochemical oxygen demand
artificial neural network
quantitative structure-biodegradability relationship
