基于启发式方法和支持向量机定量研究药物的血脑屏障通透性

QSAR study of blood-brain barrier permeation using the heuristic method and support vector machine

构建55个药物分子与其血脑屏障通透性(log B)之间的定量结构—性质关系模型,探讨影响药物血脑屏障通透性的结构因素.应用CODESSA软件计算55个化合物的组成、拓扑、几何、静电和量子化学等结构参数,通过启发式方法筛选得到最佳的结构参数,并建立线性回归模型;用所选的4个结构参数作为支持向量机的输入,建立非线性的支持向量机回归模型.预测结果表明:支持向量机回归模型性能(R^2=0.89,MSE=0.06)要优于启发式回归模型的性能(R^2=0.82,MSE=0.11).描述符HASA2,NO,FPSA3和E(CH)都是影响log B的主要结构因素.支持向量机模型简单快速,在药物设计中可以用来预测候选药物的log B值。

To build the quantitative structure-activity relationships （QSAR） between the molecular structures and the brain-blood barrier （BBB） permeability of 55 compounds and to further discuss the structural factors that influenced the BBB permeability of compounds, the topological, constitutional, geometrical, electrostatic and quantum-chemical descriptors of 55 compounds were calculated by CODESSA, and these descriptors were pre-selected using the heuristic method. As a result, the four-descriptor linear model was developed to describe the relati.onship between the molecular structures and BBB permeability. Meanwhile, the non-linear regression model was built based on support vector machine （SVM） using the same four descriptors. The predicted results indicated that the performance of the non-linear SVM model （R2=0.89, MSE=0.06） was better than that of the linear model （R2=0.82, MSE=O.11）. Descriptor HASA2, NO, FPSA3 and E（CH） were the important structural factors that influenced the log B of compounds. The SVM model was a simple and efficient tool to predict the log B values of the candidate molecules in drug design.