METHODS article
Front. Genet.
Sec. Computational Genomics
This article is part of the Research TopicMethods for Imaging and Omics Data Science: Advances, Applications, and Spatiotemporal InnovationsView all 10 articles
DWGCN: Distance-Weighted Graph Convolutional Network for Robust Spatial Domain Identification in Spatial Transcriptomics
Provisionally accepted
- 1Department of Statistical Science, School of Mathematics, Sun Yat-sen University, Guangzhou, China
- 2Center for Biomedical Data Science, Duke-NUS, Singapore, Singapore
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Background: Graph Convolutional Networks (GCNs) are widely applied for spatial domain 4 identification in spatial transcriptomics (ST), where node representations are learned by 5 aggregating information from neighboring spots. However, most ST workflows construct spatial 6 graphs by assigning equal weights to neighbors and self-loops, and then applying degree-based 7 normalization. This procedure often yields near-uniform adjacency matrices, suppressing natural 8 distance heterogeneity, diminishing spatial resolution, aggravating GCN over-smoothing, and 9 obscuring fine-grained tissue boundaries. Methods: We introduce DWGCN, a Distance-Weighted 10 Graph Convolutional Network that replaces uniform neighbor assignment with inverse-distance 11 weighting (IDW) and spot-wise normalization. DWGCN enhances locality-sensitive aggregation by 12 assigning larger weights to proximal neighbors, while preserving self-loop dominance to maintain 13 intrinsic spot information and reduce hub-driven dilution. Results: Across four real and four 14 simulated ST datasets, integrating DWGCN with representative GCN-based frameworks (SEDR, 15 GraphST, SpaNCMG, SpaGIC) generally improved clustering accuracy, particularly in tissues 16 with complex spatial architectures. Conclusion: These results demonstrate that DWGCN offers 17 a broadly applicable approach for restoring distance-aware structure in spatial graphs, thereby 18 improving the delineation of spatial domain identification.
Keywords: clustering, graph convolutional networks, representation learning, Spatial domain identification, Spatial transcriptomics
Received: 02 Jan 2026; Accepted: 27 Jan 2026.
Copyright: © 2026 Peng, Li, Wu, Fan and Guo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Xiaobo Guo
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.