Predicting anti-HIV activity: computational approach using a novel topological descriptor.

The discriminating power of a novel topological descriptor termed as eccentric adjacency index in the estimation of anti-HIV activity, for a data set of 107 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT) derivatives was investigated in the present study. The value of eccentric adjacency index of each derivative was computed and active range was identified using moving average analysis. Subsequently, each derivative was assigned a biological activity which was then compared with the reported anti-HIV activity. The accuracy of prediction was found to be more than ninety percent in the active range using eccentric adjacency index. The proposed index offers a vast potential for structure-activity/property studies.