Identification of Leukemia Enzyme Inhibitors by Molecular Modeling and Machine Learning Approaches

Background: Acute Myeloid Leukemia (AML) results in the uncontrolled proliferation of aberrant white blood cells, leading to severe health consequences and increased mortality rates. It predominantly affects males worldwide. Targeting key enzymes involved in leukemia cell proliferation through enzyme inhibition is a promising therapeutic strategy. Objectives: This study employs computational drug discovery approaches to identify potential AML inhibitors targeting four key enzymes: Tyrosine Kinase (TK), Protein Kinase C Beta (PKC), Myeloid Cell Leukemia 1 (MCL-1), and Histone Acetyltransferase (HAT). Methods: A collection of inhibitors was retrieved from the ChEMBL database, followed by molecular descriptor computations using PaDEL. Machine learning models were evaluated using LazyPredict, with RMSE and R-squared as assessment metrics, and results were visualized using Seaborn and Matplotlib. QSAR modeling was conducted using the Random Forest Regressor, achieving over 90% accuracy in predicting pIC50 values. Lead compound binding affinities were assessed through molecular docking using PyRX, while drug-likeness and pharmacokinetic properties were analyzed using ADMETlab2.0. Additionally, molecular dynamics simulations were performed via the NMSim server to evaluate the stability of lead compounds. Results: The QSAR model demonstrated strong predictive capability, with a high correlation between predicted and experimental pIC50 values. Molecular docking and ADMET analysis identified a promising novel pharmaceutical molecule with potential as an AML enzyme inhibitor. Conclusion: Computational approaches successfully identified lead compounds with strong inhibitory potential against key enzymes involved in AML. These findings offer promising avenues for future therapeutic development in AML treatment. conclusion: Computational approaches successfully identified a novel lead compound with strong inhibitory potential against key enzymes involved in AML. These findings offer promising avenues for future therapeutic development in AML treatment.