Topological Models for Prediction of Adductability of Branched Aliphatic Compounds in Urea

Abstract

The relationship of adductability of branched chain aliphatic compounds in urea with topological descriptors has been investigated. Wiener’s index – a distance-based topological descriptor, molecular connectivity index, an adjacency-based topological descriptor and eccentric connectivity index – an adjacency-cum-distance based topological descriptor were employed for the present study. A data set comprising of 133 branched aliphatic compounds was segregated into training and test sets. The values of all the three topological indices for all the compounds constituting the training and test sets were computed using an in-house computer program. Resulting data of the training set was analyzed and suitable models were developed after identification of the adductible ranges. Subsequently, each compound in the training set was either classified as adductible or non-adductible using these models, which was then compared with the reported adductability in urea. An accuracy of prediction of ≥86% was observed using these models in the training set. These models were then cross-validated using the test set. An accuracy of prediction of ≥80% was observed during cross-validation of these models in an independent test set.