The present study deals with successful utilization of both correlation and classification techniques for the development
of diverse models for the prediction of dual mTOR and PI3Kα inhibitory activities using a dataset comprising
of 39 analogues of 4-morpholinopyrrolopyrimidines. Decision tree, random forest, moving average analysis (MAA) and
multiple linear regression (MLR) were used to develop models for mTOR and PI3Kα inhibitory activities. The proposed
models were also found to be sensitive towards the cell inhibition against PC3. The models were assessed for statistical
significance in terms of overall accuracy of prediction, specificity, sensitivity, Mathew’s correlation coefficient (MCC)
and intercorrelation analysis. The high predictability of the proposed models of diverse nature offers enormous potential
for providing lead molecules for the development of potent medicinal agents for dual mTOR and PI3Kα inhibition.
Keywords: mTOR inhibitors and PI3Kα inhibitors, 4-morpholinopyrrolopyrimidines, Topological descriptors, Decision Tree,
Random forest, Moving average analysis.
Rights & PermissionsPrintExport