中文摘要：

为了研究60种取代芳烃生物降解最大去除率（QTOD）的定量构效关系,利用最佳变量子集回归建立50种有机污染物（训练集）QTOD与新型分子连接性指数（mGtv）的五元QSBR（定量结构-生物降解相关性）模型,其传统判定系数（R2）、逐一剔除法（LOO）的交互验证系数（Q2）分别为0.841和0.792.据此模型可知取代芳烃生物降解控制步骤为酶催化反应,其中包含＂诱导契合机理＂.利用反向传播（BP）算法获得了一个QSBR模型,其R2和估计标准误差分别为0.993和2.151,表明QTOD与这5个结构参数具有良好的非线性关系.

英文摘要：

Quantitative structure-activity relationship （QSAR） between the maximum specific removal rates （QTOD） of biodegradation for 60 substituted aromatic molecules was studied.The five quantitative structure-biode-gradability relationship （QSBR） model between QTOD of biodegradation and a novel molecular connectivity index（mGvt） for 50 organic pollutants in the training set was developed by leaps-and-bounds regression（LBR）,the traditional correlation coefficient（R2）,the cross-validation correlation coefficient（Q2） of leave-one-out（LOO） were 0.841,0.792,respectively.From the model,the control step of biodegradation for substituted aromatic compounds is enzyme reaction,which includes ＂induced fit mechanism＂.A satisfying BP-QTOD model could be constructed with the back-propagation algorithm of artificial neural network,with the correlation coefficient and the standard error being 0.993 and 2.151,respectively,showing that the relationship between QTOD and the five structural parameters has a good nonlinear correlation.