估算醇的溶解度和辛醇水分配系数的价连接性指数模型

Valence connectivity index method for predicting solubility and octanol-water partition coefficients of alcohol

在分子图的邻接矩阵基础上提出了一个新的价连接性指数 mL ,用多元线性回归技术分别建立了 6 0个醇的价连接性指数 mL与它们的辛醇 水分配系数lgKow及水溶解度lgSw 的定量结构—性质相关性模型 .该模型具有良好的稳定性和预测能力 .辛醇 水分配系数的线性方程的相关系数为 0 .995 7,溶解度的线性相关方程的相关系数为 0 .9882 .由于指数充分考虑了空间结构和电性结构 ,因此对含杂原子化合物的区分能力得到加强 ,计算也较为简单 ,预测的精度也优于文献报道结果 .对辛醇 水分配系数预测值的平均误差仅为 0 .10 ,对水溶解度预测值的平均误差也只有 0 .16 ,均小于文献方法的预测误差 ,而且本指数能较好地反映化合物的结构特征 .

Based on the adjacency matrix,a novel valence connectivity index mL is derived. Quantitative structure property relationship on solubility and octanol-water partition coefficients of 60 alcohols were proposed respectively by multiple linear regression method with good stability and predictivity. The correlation coefficient of the linear regression equation for octanol-water partition coefficients is 0.995?7. The correlation coefficient of the linear regression equation for solubility is 0.988?2. Because of the sufficient discussion of the space structure and electricity structure,the capability of the index is enhanced for compounds including non-carbon atom. The calculating is simpler, and the precision is superior to those reported by references. The average error of the prediction for octanol-water partition coefficients is 0.10, and that for solubility is 0.16, which are all smaller than those reported by references. Furthermore, the results demonstrate that the structure property of compounds can be described by valence connectivity indices.