基于偏最小二乘法的二萜类生物碱化合物毒性定量构效关系研究  

The investigation on quantitative structure-toxicity relationship for norditerpenoid alkaloids using partial least square method

作　　者：李晓[1] 毕开顺[1] 李清[1] 隋振宇[1] 常丽娟[1] 李佐静[2] 

机构地区：[1]沈阳药科大学药学院,辽宁沈阳110016 [2]沈阳药科大学医疗器械学院,辽宁沈阳110016

出　　处：《计算机与应用化学》2016年第2期228-230,共3页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(41310164);辽宁省高校优秀人才支持计划(51120915);辽宁省自然科学基金资助项目(51120920)

摘　　要：二萜类生物碱是一种从植物中提取的结构复杂的物质,具有很大药用价值。它不仅含有有效成分,还含有毒性成分。经研究它的毒性与其化学结构有关。为了研究它们之间的关系,本文使用定量构效关系方法研究其化合物的量化参数与毒性之间的关系,先计算30种该类化合物的三维结构理化参数,以这些结构参数作为自变量,以毒性数据值作为因变量,并利用偏最小二乘判别分析方法建立模型。结果表明该模型具有良好的预测能力。Norditerpenoid alkaloids whose structure is complex are extracted from plants and have medicinal value. These compounds not only include active ingredients, but also include toxic ingredients. Its toxicity is closely related with the chemical structure. We used quantitative structure-toxicity relationship （QSTR） method to study the relationship between toxicity and structure of the series of compounds in this paper. Firstly, we calculated the three-dimensional structure parameters of thirty kinds of compounds by using HyperChem software, and established the model with these parameters as dependent variable, the independent variable is the toxicity data from rats by intravenous injection which is obtained from chemidplus database. The partial least squares method of SIMCA-P software was applied to establish the model of quantitative structure-toxicity relationship between toxicity and structure of these compounds. And we use this model to predict toxicity, indicating that the model has good ability to predict toxicity.

关 键 词：二萜类生物碱 偏最小二乘法 定量构效关系 

分 类 号：TQ015.9[化学工程] TP391.9[自动化与计算机技术—计算机应用技术]