Application of novel atom-type AI topological indices in the structure-property correlations

摘要

Novel atom-type AI topological indices are generated as new parameters in quantitative structure-property/activity relationships (QSPR/QSAR) models to encode the structural environment of each atom-type in a molecule. These AI indices, along with previously proposed Xu index, are extended to complex compounds with heteroatoms by using the novel vertex degree v~m, which is derived on the basis of the valence connectivity δ~v of Kier-Hall. The efficiency of the approach is demonstrated through three high quality QSPR models of the molar volumes (MV), molar refractions (MR), and molecular total surface areas (TSA) for three data sets of compounds consisting of alkanes and alcohols. The results indicate that combination of the atomic-based AI indices and Xu index can produce a significant improvement in the statistical quality of the models obtained for the three properties. The significant improvement indicates the high potential of these indices for application to various physical properties and structural types, especially complex compounds with special functional groups. For the final multiple linear regression models, the correlation coefficients r are 0.9965, 0.9993, and 0.9990, and the standard errors s are 2.603, 0.3223, and 0.393 for MV, MR, and TSA, respectively. In addition, the results indicate that three properties are dominated by molecular size but other atomic groups are also important although their contributions are much smaller than that of the molecular size. The cross-validation using the more general leave-n-out method demonstrates the final models to be highly statistically reliable.