Wang G, Bai N
Sino-Canadian High-tech Center of Resources and Environment, Chinese Academy of Siences, Beijing, China.
Chemosphere. 1998 Mar;36(7):1475-83. doi: 10.1016/s0045-6535(97)10046-7.
Structure-activity relationships between acute toxicities of 95 alcohols to rat and mouse (oral LD50) and four special sub-structure factors, hydroxyl number, carbon atom number were examined by means of expert system method. The results showed that the expert system based QSAR model was excellent for classification for miscellaneous alcohols (only 9 of them were wrong classified). It was also used to predict the toxicity of other 25 alcohols, and the false prediction rate was only 12%.
采用专家系统方法研究了95种醇类对大鼠和小鼠的急性毒性(经口半数致死量)与四个特殊子结构因素(羟基数、碳原子数)之间的构效关系。结果表明,基于专家系统的定量构效关系模型对杂醇类的分类效果极佳(仅有9种分类错误)。该模型还用于预测其他25种醇类的毒性,误预测率仅为12%。