RS-REMD Protein-GAG Interaction Dataset in CHARMM36m

The dataset includes input files, simulation parameters, and analysis scripts used in Repulsive Scaling Replica Exchange Molecular Dynamics (RS-REMD) simulations to study protein–glycosaminoglycan (GAG) interactions. In this study, the RS-REMD method was applied for molecular docking of GAGs and carbohydrates to selected protein targets. Molecular Mechanics Generalized Born Surface Area (MM-GBSA) served as the scoring function, and a Fully Connected Neural Network (FCNN) model was subsequently trained using MM-GBSA energies and structural properties to predict the Root Mean Square Atom Type Deviation (RMSatd), a metric that quantifies structural similarity. The dataset includes structures selected based on MM-GBSA and FCNN-predicted RMSatd values, supporting binding site identification and energy evaluation. The provided Python scripts facilitate force field modification, RMSatd calculations, and machine learning model training and application.