Jiao Long, Gao Mingming, Wang Xiaofei, Li Hua
College of Chemistry and Chemical Engineering, Xi'an Shiyou University, Xi'an 710065, People's Republic of China ; College of Chemistry and Materials Science, Northwest University, Xi'an 710069, People's Republic of China.
No.203 Research lnstitute of Nuclear industry, Xianyang 712000, People's Republic of China.
Chem Cent J. 2014 Jun 10;8:36. doi: 10.1186/1752-153X-8-36. eCollection 2014.
The quantitative structure property relationship (QSPR) for octanol/air partition coefficient (K OA) of polybrominated diphenyl ethers (PBDEs) was investigated. Molecular distance-edge vector (MDEV) index was used as the structural descriptor of PBDEs. The quantitative relationship between the MDEV index and the lgK OA of PBDEs was modeled by multivariate linear regression (MLR) and artificial neural network (ANN) respectively. Leave one out cross validation and external validation was carried out to assess the predictive ability of the developed models. The investigated 22 PBDEs were randomly split into two groups: Group I, which comprises 16 PBDEs, and Group II, which comprises 6 PBDEs.
The MLR model and the ANN model for predicting the K OA of PBDEs were established. For the MLR model, the prediction root mean square relative error (RMSRE) of leave one out cross validation and external validation is 2.82 and 2.95, respectively. For the L-ANN model, the prediction RMSRE of leave one out cross validation and external validation is 2.55 and 2.69, respectively.
The developed MLR and ANN model are practicable and easy-to-use for predicting the K OA of PBDEs. The MDEV index of PBDEs is shown to be quantitatively related to the K OA of PBDEs. MLR and ANN are both practicable for modeling the quantitative relationship between the MDEV index and the K OA of PBDEs. The prediction accuracy of the ANN model is slightly higher than that of the MLR model. The obtained ANN model shoud be a more promising model for studying the octanol/air partition behavior of PBDEs.
研究了多溴二苯醚（PBDEs）正辛醇/空气分配系数（K OA）的定量结构-性质关系（QSPR）。采用分子距离-边向量（MDEV）指数作为PBDEs的结构描述符。分别通过多元线性回归（MLR）和人工神经网络（ANN）建立了MDEV指数与PBDEs的lgK OA之间的定量关系。采用留一法交叉验证和外部验证来评估所建立模型的预测能力。将所研究的22种PBDEs随机分为两组：第一组包含16种PBDEs，第二组包含6种PBDEs。
建立了预测PBDEs的K OA的MLR模型和ANN模型。对于MLR模型，留一法交叉验证和外部验证的预测均方根相对误差（RMSRE）分别为2.82和2.95。对于L-ANN模型，留一法交叉验证和外部验证的预测RMSRE分别为2.55和2.69。
所建立的MLR和ANN模型对于预测PBDEs的K OA是可行且易于使用的。结果表明PBDEs的MDEV指数与PBDEs的K OA存在定量关系。MLR和ANN均可用于建立MDEV指数与PBDEs的K OA之间的定量关系模型。ANN模型的预测精度略高于MLR模型。所获得的ANN模型应该是研究PBDEs正辛醇/空气分配行为的更有前景的模型。
