基于区间活性数据的定量结构-活性关系  

Quantitative structure-activity relationship based on the interval censored data

作　　者：赵劲松[1] 

机构地区：[1]农业部长江中下游耕地保育重点实验室,华中农业大学资源与环境学院,武汉430070

出　　处：《环境化学》2013年第7期1188-1193,共6页Environmental Chemistry

基　　金：国家自然科学基金(40601085);中央高校基本科研业务费专项资金(2012ZYTS031)资助

摘　　要：利用贝叶斯统计方法构建了基于区间活性数据的取代苯胺和苯酚类化合物对大型溞(Daphniamagna)24 h急性毒性的定量结构-活性关系模型,并与基于平均数和中位数的点估计活性数据的定量结构-活性关系模型进行了比较.结果表明,前者可以充分利用化合物的活性数据信息,模型具有更好的拟合效果与预测能力以及较宽的应用范围.基于区间活性数据的定量结构-活性关系模型可为生态风险评价等提供更加可靠的预测数据.In this study, by using Bayesian statistical analysis, quantitative structure-activity relationship （QSAR） models of substituted anilines and phenols for Daphnia magna were developed based on the interval censored data of 24 h acute toxicities. They were compared with QSAR models built from the point estimation, i. e. , the mean and median of the interval censored dataset. The results showed that the former QSAR models could effectively use the biological activity information of specific compounds, and therefore had a better goodness-of- fit and predictive ability, and a wider range of applications. QSAR model based on the interval censored data could provide a more reliable predicted biological activity for the ecological risk assessment.

关 键 词：区间数据 贝叶斯统计 定量结构-活性关系 取代苯胺和苯酚类化合物 

分 类 号：X13[环境科学与工程—环境科学]