Li Xiuyong, Luan Feng, Si Hongzong, Hu Zhide, Liu Mancang
Department of Chemistry, Lanzhou University, Lanzhou, Gansu 730000, China.
Toxicol Lett. 2007 Dec 10;175(1-3):136-44. doi: 10.1016/j.toxlet.2007.10.005. Epub 2007 Oct 18.
Quantitative structure-retention relationship (QSRR) studies were performed for predicting the retention times (RTs) of 110 kinds of pesticides or toxicants. Chemical descriptors were calculated from the molecular structure of the compounds alone. The QSRR models were built using the heuristic method (HM) and support vector machine (SVM), respectively. The obtained linear model of HM had a square of a correlation coefficient: R(2)=0.913, F=116.70 with a root mean square error (RMS) error of 0.0387 for the training set, while R(2)=0.907, F=195.49, and RMS=0.0408 for the test set. The non-linear model by SVM gave better results: for the training set R(2)=0.966, F=2420.5, RMS=0.0231 and for the test set R(2)=0.944, F=339.7, RMS=0.0313. The prediction results are in good agreement with the experimental values. And the proposed model could identify and provide some insight into what structural features are related to retention time of these compounds.
开展了定量结构-保留关系（QSRR）研究，以预测110种农药或有毒物质的保留时间（RTs）。化学描述符仅根据化合物的分子结构计算得出。分别使用启发式方法（HM）和支持向量机（SVM）构建了QSRR模型。所获得的HM线性模型，训练集的相关系数平方为R(2)=0.913，F=116.70，均方根误差（RMS）为0.0387，而测试集的R(2)=0.907，F=195.49，RMS=0.0408。SVM的非线性模型给出了更好的结果：训练集的R(2)=0.966，F=2420.5，RMS=0.0231，测试集的R(2)=0.944，F=339.7，RMS=0.0313。预测结果与实验值吻合良好。并且所提出的模型能够识别并揭示与这些化合物保留时间相关的结构特征。
