Saved in:
Bibliographic Details
Main Authors: Zhong, Zhixuan, Ma, Linbo, Jiang, Jian
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.20893
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915513322438656
author Zhong, Zhixuan
Ma, Linbo
Jiang, Jian
author_facet Zhong, Zhixuan
Ma, Linbo
Jiang, Jian
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2509_20893
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MolCluster: Integrating Graph Neural Network with Community Detection for Coarse-Grained Mapping
Zhong, Zhixuan
Ma, Linbo
Jiang, Jian
Computational Physics
Chemical Physics
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.
title MolCluster: Integrating Graph Neural Network with Community Detection for Coarse-Grained Mapping
topic Computational Physics
Chemical Physics
url https://arxiv.org/abs/2509.20893