有机污染物生物富集因子的预测模型  

Prediction model of bioaccumulation of organic pollutants

作　　者：贾艳平[1] 李燕[2] 周一兵[1] 王永华[1] 

机构地区：[1]大连水产学院辽宁省高校海洋生物资源可持续利用重点实验室,辽宁大连116023 [2]大连理工大学化工学院,辽宁大连116023

出　　处：《大连水产学院学报》2008年第4期288-295,共8页Journal of Dalian Fisheries University

基　　金：国家“863”计划资助项目(2006AA10Z410);国家自然科学基金资助项目(30571419);辽宁省自然科学基金资助项目(2040605)

摘　　要：为了定量预测鱼体内生物富集因子(BCF),确定其最佳预测模型,将214个有机物分子作为数据集,其中随机选取54个分子作为检验集,其它160个分子为训练集,构建了定量构效关系模型(QSAR)。每个分子计算了193个分子参数,采用多元线性回归分析中逐步筛选法获得最佳模型(应用于训练集与检验集的相关系数分别为R2=0.81、R2=0.66)。此外,lgBCF与lgP(油水分配系数)的关联结果也显示出较强的相关性(R2=0.62)。A QSAR （Quantitative Structure -Activity Relationship） model was put forward to quantitatively predict bioaccumulation of organic compounds in fish on the basis of a dataset of 214 diversified compounds in which 54 compounds served as test set and the rest as training set. About 193 molecular indices were calculated for each compound. From the analysis of the models, the predictability of stepwise MLR was satisfactory （R^2 = 0. 81, R^2 = 0.66） for the present work. In addition, the relationship between log BCF and lg P was also reasonable （R^2 = 0.62） , indicating that all these studies was helpful for prediction of bioaccumulation of organic compounds in fish.

关 键 词：生物富集因子(BCF) 分子参数 多元线性回归 QSAR 

分 类 号：R913[医药卫生—药学]