stGCL: a versatile cross-modality fusion method based on multi-modal graph contrastive learning for spatial transcriptomics

Journal article

Advances in spatial transcriptomics have enabled high-resolution mapping of tissue architecture at the molecular level, yet integrating its multi-modal data remains challenging. Here, we present stGCL, a framework for accurate and robust integration of gene expression, spatial coordinates, and histological features. stGCL employs a histology-based Vision Transformer to extract morphological features and a multi-modal graph autoencoder with contrastive learning for cross-modal fusion. In addition, we introduce a spatial coordinate correction and registration strategy to support multi-slice integration. We demonstrate that stGCL reliably identifies spatial domains, integrates vertical and horizontal tissue slices, and highlight its generalizability across platforms and resolutions.

Authors: Yu, N; Zhang, D; Zhang, W; Liu, Z; Qiao, X

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: BioMed Central
• Journal: Genome Biology
• Volume: 27
• Issue: 1
• Article number: 51
• Publication date: 2026-01-28
• Acceptance date: 2025-12-01
• DOI: 10.1186/s13059-025-03896-w
• EISSN: 1474-760X
• ISSN: 1474-7596
• Language: English
• Keywords: Autoencoder; Graph; Spatial analysis; Pattern recognition (psychology); Generalizability theory; Spatial reference system; Transcriptome; Proteogenomics