Potential anti-colon cancer agents: Molecular modelling, docking, pharmacokinetics studies and molecular dynamic simulations

Abstract

Objective

The objective of this investigation is to create a trustworthy Quantitative Structure-Activity Relationship (QSAR) model that generates little to no side effects and is low-cost for treating colon cancer using experimental data obtained from the literature.

Methods

ChemDraw software was used for creating molecular structures, which were then optimized using Spartan 14 software to generate quantum chemical descriptors. Data pre-treatment and data division were performed using specific software packages. Additionally, analysis and validation tasks were carried out using software tools such as Discovery Studio Visualizer, PyRx for docking, SwissADME for pharmacokinetics studies, and Desmond for molecular dynamic (MD) simulation.

Results

The developed QSAR model demonstrates good predictive quality with a Mean Absolute Error (MAE) of 1.3313 and high internal validation metrics (R² ​= ​0.9407, adjusted R² ​= ​0.9329). External validation on a test set yields satisfactory results (R² ​= ​0.9012, adjusted R² ​= ​0.8436, CCC ​= ​0.9229). Docking analysis identifies compounds 111 and 112 as having the lowest binding affinity of −10.4 kJ/mol, characterized by specific molecular properties. Additionally, MD simulation provides insights into the dynamic behavior and interaction types of the protein-ligand complex, contributing to a deeper understanding of their stability and fluctuations.

Conclusion

The model validation parameters confirm the reliability and robustness of the model. The pharmacokinetics study validates the drug-likeness of the drug candidate through various parameters. The MD simulation sheds light on the dynamic behavior and interaction types of the protein-ligand complex, enhancing our understanding of their stability and fluctuations.