Path Pendeccentric Connectivity Indices: Detour Matrix Based Molecular Descriptors for QSAR/QSPR Studies, Part 2

Abstract

A pivotal role of tyrosine threonine kinase (TTK) has been established in tumor initiation, survival of genomically unstable and aneuploid cancer cells. In present study, path pendeccentric connectivity indices reported in part 1 of the manuscript were successfully applied for developing models for predicting TTK inhibitory activity of acetamide/carboxamide analogs. Diverse 2D and 3D molecular descriptors (MDs) were successfully utilized for developing models using artificial neural networks (ANN) and moving average analysis (MAA). The overall accuracy of prediction achieved for ANN and MAA based models was up to 96% for the training set and up to 92% during cross validation. The statistical utility of the said models was also evaluated through Matthews correlation coefficient, non error rate, sensitivity and intercorrelation analysis. Low IC50 values obtained for active ranges of the proposed MAA based models indicate the tremendous potential of said models for furnishing lead molecules for developing potent TTK inhibiting acetamide/carboxamide analogs.