R-CCS Cafe 第274回（2024年12月2日）

Title:
Efficient All-Atom Structure Sampling via Parallelized Coarse-Grained Elastic Network Model and CG-to-AA Reconstruction
Abstract:
Efficient simulation of large-scale structural changes in target proteins is crucial for practical drug development. However, unbiased all-atom molecular dynamics (AA-MD) simulations often fail to capture long-timescale (ms-s) dynamics of large biomolecules due to computational limitations. Coarse-grained molecular dynamics (CG-MD), which simplifies each residue to a single degree of freedom, offers a computationally efficient alternative but lacks essential structural details, such as side chains, critical for drug design. To address these challenges, we developed a framework combining parallelized CG elastic network model (CG-ENM) sampling with efficient CG-to-AA reconstruction. This method enhances sampling coverage and restores atomic-level details required for detailed analysis. In this presentation, we will describe the methodology, performance improvements, and potential applications.