Integrating Deep Learning and Molecular Dynamics to Identify GPR17 Ligands for Glioblastoma Therapy

Background: Guanine Protein-coupled Receptor 17 (GPR17) plays pivotal roles in various physiological processes and diseases. However, the discovery of ligands binding to GPR17 remains an active area of research. Methods: In this study, we utilized our recently published GPCR-specific deep learning approach, molecular docking, and molecular dynamics simulations. Specifically, the DeepGPCR model, employing graph convolutional networks, was used to screen the extensive ZINC database for potential ligands. Results: This computational pipeline identified three highly promising lead compounds, ZINC000044404209, ZINC000229938097, and ZINC000005158963. Molecular dynamics simulations confirmed the stability of the protein-ligand complexes while binding free energy calculations highlighted the crucial molecular forces stabilizing these interactions. Notably, ZINC000229938097 exhibited particularly favorable binding energy values among the compounds assessed. Conclusion: Our study underscores the efficacy of computational methodologies in identifying potential drug candidates targeting GPR17. Understanding the molecular mechanisms underlying GPR17 activation holds significant promise for developing tailored therapies for Glioblastoma Multiforme. conclusion: The current study highlights the utility of computational tools in drug development and provides information on GPR17 as a target for GBM treatment and the directions for further experimental testing and utilization.