QSAR of Topoisomerase I Inhibitors using Cluster based Descriptor Procedure for Camptothecin Analogs

Abstract

This article aimed to derive Quantitative structure-activity relationship (QSAR) model for 413 camptothecin (CPT) analogs which are potent Topoisomerase I (Topo I) inhibitors by devising a methodology to integrate the molecules, choose descriptors for entire dataset, have less outliers and increase the cross validated r² (r²_cv) and conventional r² (q²) values with minimum Standard Error of Prediction. A hierarchical clustering based QSAR equation is modeled for CPT analogs which would help to understand the CPT and the Topo I characteristics as well as assist in developing novel analogs with better inhibitory activity. The BASE model r²_cv value was found to be 0.78 and q² value 0.82. The 3 cluster based equations gave q² value of 0.82, 0.86 and 0.86, respectively with less number of outliers and high predictive r² (r²_pred) value. The models were compared with COMFA QSAR models and they gave better prediction results. In this novel work we have optimized the 3D QSAR methodology by executing a model which identifies all the SAR characteristics of CPT and demonstrate the predictive capacity of the model for the entire CPT family unlike in earlier study wherein QSAR models are based on only specific group substitutions.