Estrada Ernesto, Delgado Eduardo J, Alderete Joel B, Jaña Gonzalo A
Molecular Informatics, X-rays Unit, RIAIDT, Edificio CACTUS, University of Santiago de Compostela, 15782 Santiago de Compostela, Spain.
J Comput Chem. 2004 Nov 15;25(14):1787-96. doi: 10.1002/jcc.20099.
Quantum-connectivity indices are topographic descriptors combining quantum-chemical and topological information. They are used to describe the water solubility of a noncongeneric data set of organic compounds. A QSPR model is obtained with two quantum-connectivity indices that accounts for more than 90% of the variance in the water solubility of these chemicals. This model is compared to other five QSPR models using constitutional, electrostatic, geometric, quantum-chemical, and topological descriptors calculated by CODESSA. None of these models accounts for more than 85% of the variance in water solubility of the compounds in this data set. The QSPR model obtained with quantum-connectivity indices is also better than that generated from the general pool of 508 CODESSA indices. Models with up to five variables were explored and compared with the model obtained here. It is shown that quantum-connectivity indices contain more structural information than other classes of descriptors at least for describing the water solubility of these 53 chemicals. Structural interpretation of the QSPR model developed as well as the role of the quantum-connectivity indices included in it are also analyzed.
量子连接性指数是结合量子化学和拓扑信息的拓扑描述符。它们用于描述一组非同类有机化合物的水溶性。通过两个量子连接性指数获得了一个定量构效关系（QSPR）模型，该模型解释了这些化学物质水溶性中超过90%的方差。将该模型与使用由CODESSA计算的组成、静电、几何、量子化学和拓扑描述符得到的其他五个QSPR模型进行了比较。这些模型中没有一个能解释该数据集中化合物水溶性方差的85%以上。用量子连接性指数获得的QSPR模型也优于从508个CODESSA指数的通用库中生成的模型。探索了多达五个变量的模型，并与这里获得的模型进行了比较。结果表明，至少对于描述这53种化学物质的水溶性，量子连接性指数比其他类型的描述符包含更多的结构信息。还分析了所开发的QSPR模型的结构解释以及其中包含的量子连接性指数的作用。
