文献类型：期刊文献

中文题名：氯代环醚系列化合物结构与急性毒性关系研究

英文题名：Studies on QSAR for cycloethers and polychlorinated cycloethers

作者：白乃彬[1];杜玉[2];谢飞[2];王兴平[2]

第一作者：白乃彬

机构：[1]中国科学院生态环境研究中心;[2]北京联合大学化学工程学院

第一机构：中国科学院生态环境研究中心,北京100085

年份：2001

卷号：21

期号：4

起止页码：330-334

中文期刊名：中国环境科学

外文期刊名：China Environmental Science

收录：CSTPCD;;Scopus;北大核心:【北大核心2000】；CSCD:【CSCD2011_2012】；CSSCI:【CSSCI2000_2002】；

基金：国家自然科学基金资助项目(29877032)

语种：中文

中文关键词：神经网络;氯化环醚;急性毒性;构效关系;化合物结构

外文关键词：neural networks；polychlorinated cycloethers；acute toxicity；quantitative structure-activity relationship

摘要：由化学物质毒性效应登录(RTECS)1998年版光盘系统检索获得52种环醚和34种氯代环醚急性毒性数据.应用主成份分析法从11种分子描述符中选择出对急性毒性有明显影响的5种:氯原子数,取代苯环上氢的氯原子数,取代烷基氢的氯原子数,烷基数和氧原子数.通过样本学习集训练并优化神经网络结构、建模分类、预报,对于86个样本,低中和高剧两类毒性,正确分类率达到91%;而对于低、中、高、剧4类毒性,正确分类率达到80%.结果表明,氯代环醚的特殊结构具有很强的致毒作用.
Acute toxicity data of 52 cycloethers and 34 polychlorinated cycloethers were collected from a optical disk system of Registry of Toxic Effects of Chemical Substances (RTECS) of 1998 version. By using principal component analysis, 5 kinds having marked influence on acute toxicity were chosen from 11 kinds of molecular structure descriptors; the number of chloro atom, the number of substituted chloro atom in the benzene-cycle, the number of substituted chloro atom in the alkyl, the numbers of alkyl atom and oxygen atom. After learning of above sample sets and neural network structure optimized, the established models were classified, it was predicted: for 86 samples, the correct classifying rate of low-middle and high-severe classes reached 91%; and reached 80% for 4 classes of low, middle, high and severe. Results showed that the special structure of polychlorinated cycloethers possess very high toxicities.