QSAR studies on some $C14$-urea tetrandrine compounds as potent anti-cancer agents against Leukemia cell line (K562)

Abstract

This research applied Quantitative Structure Activity Relationship (QSAR) technique indeveloping a Multiple-Linear Regression (MLR) model using Genetic Functional Approximation (GFA)method in selecting optimum molecular descriptors from the structures of 24 C14-urea tetrandrinecompounds. Firstly, the compounds were optimized at the Density Functional Theory (DFT) level usingBecke’s three-parameter Lee-Yang-Parr hybrid functional (B3LYP) with the 6-31G* basis set in theSpartan 14 Version 1.1.4 software. The descriptors of the compounds were computed using Padelsoftware,and data set was divided into training and test set. A model was built from the training setwith internal validation parameter R2train as 0.9104. The external validation of the model was done usingthe test set compounds with validation parameter R2test as 0.6443 that passed the criteria foracceptability of a QSAR model globally. The coefficient of determination (𝑐𝑅2𝑝) parameter was calculatedas 0.8192 which is greater than 0.5, this affirms that the generated model is robust. Furthermore,AST4p, GATS8v, and MLFER are descriptors in the model with the positive mean effect of 0.0899, 0.9098and 0.0002 respectively. This study depicts a route in designing and synthesizing new C14-ureatetrandrine compounds with better inhibitory potentials.