MolCluster: Integrating Graph Neural Network with Community Detection for Coarse-Grained Mapping

2025/09/25 by Zhixuan Zhong, Zhong, Zhixuan, Linbo Ma +3
Computer Science · Materials Science · #Advanced Graph Neural Networks #Chemical Physics (physics.chem-ph) #Computational Physics (physics.comp-ph) #FOS: Physical sciences #Machine Learning in Materials Science #Neural Networks and Applications

paper · pdf · doi:10.48550/arxiv.2509.20893

openalex publication_date 2025/09/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Coarse-grained (CG) modeling simplifies molecular systems by mapping groups of atoms into representative units. However, traditional CG approaches rely on fixed mapping rules, which limit their ability to handle diverse chemical systems and require extensive manual intervention. Thus, supervised learning-based CG methods have been proposed, enabling more automated and adaptable mapping. Nevertheless, these methods suffer from limited labeled datasets and the inability to control mapping resolution, which is essential for multiscale modeling. To overcome these limitations, we propose MolCluster, an unsupervised model that integrates a graph neural network and a community detection algorithm to extract CG representations. Additionally, a predefined group pair loss ensures the preservation of target groups, and a bisection strategy enables precise, customizable resolution across different molecular systems. In the case of the downstream task, evaluations on the MARTINI2 dataset demonstrate that MolCluster, benefiting from its label-free pretraining strategy, outperforms both traditional clustering and supervised models. Overall, these results highlight the potential of MolCluster as a core model for customizable and chemically consistent CG mapping.

Citations

• Physical embedding machine learning force fields for organic systems
• A Neural-Network-Based Mapping and Optimization Framework for High-Precision Coarse-Grained Simulation
• Efficient Machine Learning Force Field for Large-Scale Molecular Simulations of Organic Systems
• Statistically Optimal Force Aggregation for Coarse-Graining Molecular Dynamics
• Hierarchical Nearest Neighbor Graph Embedding for Efficient Dimensionality Reduction
• Molformer: Motif-based Transformer on 3D Heterogeneous Molecular Graphs
• GeomGCL: Geometric Graph Contrastive Learning for Molecular Property Prediction
• Martini 3: a general purpose force field for coarse-grained molecular dynamics
• Efficient Parameter-free Clustering Using First Neighbor Relations
• Machine Learning of coarse-grained Molecular Dynamics Force Fields
• Attention Is All You Need
• Los carnavales de La Bañeza: Aspectos antropológicos
• Perspective: Coarse-grained models for biomolecular systems
• A Fast and High Quality Multilevel Scheme for Partitioning Irregular Graphs

Related

• Graph-Coarsening for Machine Learning Coarse-grained Molecular Dynamics
• Mapping Still Matters: Coarse-Graining with Machine Learning Potentials
• Coarse graining molecular dynamics with graph neural networks
• MCGM: Multi-stage Clustered Global Modeling for Long-range Interactions in Molecules
• HEroBM: a deep equivariant graph neural network for universal backmapping from coarse-grained to all-atom representations