GroScore: Automated Scoring of Protein–Protein Interactions Using Bidirectional Steered Molecular Dynamics

Abstract

Abstract GroScore is an open-source Python toolkit that automates the scoring of protein–protein structures using bidirectional steered molecular dynamics (SMD) simulations with explicit solvent. Building on a previously published unidirectional pulling approach, the current version introduces bidirectional pulling (unbinding and rebinding), interface cutout for reduced computational cost, support for three force fields (AMBER19SB, CHARMM36m, GROMOS 54A8), and automated handling of small molecules and noncanonical amino acids (using OpenFF), structural ions, and crystal waters. GroScore integrates with GROMACS and SLURM for high-performance computing and is validated on protein–protein affinity benchmarks, showing significant correlation between computed scores and experimental binding affinities across all tested force fields. The software is freely available at https://github.com/jwperthold/GroScore under the MIT License.