Interpretable spatial multi-omics data integration and dimensionality reduction with SpaMV
8 affiliations
- 1Department of Computer Science, Hong Kong Baptist University, Hong Kong SAR, China
- 2College of Biomedical Engineering, Sichuan University, Chengdu, China
- 3Department of Computer Science, Vanderbilt University, Nashville, TN, USA
- 4Department of Statistics, Chinese University of Hong Kong, Hong Kong SAR, China
- 5Department of Computer Science, City University of Hong Kong, Hong Kong SAR, China
- 6Department of Biomedical Engineering, Vanderbilt University, Nashville, TN, USA. [email protected]
- 7Department of Computer Science, Vanderbilt University, Nashville, TN, USA. [email protected]
- 8Department of Computer Science, Hong Kong Baptist University, Hong Kong SAR, China. [email protected]
Abstract
Spatial multi-omics technologies have revolutionized our understanding of biological systems by providing spatially resolved molecular profiles from multiple perspectives. Existing spatial multi-omics integration methods often assume that data from different modalities share a common underlying distribution, aiming to project them into a single unified latent space. This assumption, however, can obscure the unique insights offered by each modality, thereby limiting the full potential of multi-omics analyses. To address this limitation, we present the Spatial Multi-View (SpaMV) representation learning algorithm, which captures both the shared information across modalities and the distinct, modality-specific information, enabling a more comprehensive and interpretable representation of spatial multi-omics data. Through extensive evaluation on both simulated and real-world datasets, SpaMV demonstrates superior spatial domain clustering performance and provides users with more interpretable dimension reduction for downstream analysis. Moreover, SpaMV effectively annotates cell types within clusters of a mouse thymus dataset, highlighting its effectiveness in interpretable dimensionality reduction.
Related
- smiMAD: adaptive multi-view graph learning for paired spatial omics integration
- SubPoE: a subgraph-induced Product-of-Experts framework for spatial multi-omics integration
- Sampling Budget as a Confound in Dimensionality-Reduction Analyses of Molecular Simulation
- Interpretable and scalable spatial gene set activity analysis with GESSO uncovers functional tissue architecture
- A deep learning-based multiscale integration of spatial omics with tumor morphology
BibTeX
@article{liu2026interpretable,
title = {{Interpretable spatial multi-omics data integration and dimensionality reduction with SpaMV}},
author = {Yang Liu and Kexin Ma and Haoran Xu and Ke Xu and Yunfei Hu and Zhenhan Lin and Jiangli Lin and Bo Han and Shuai Cheng Li and Zhixiang Lin and Xin Zhou and Lu Zhang},
journal = {Nature Communications},
year = {2026},
month = jul,
doi = {10.1038/s41467-026-74718-1},
url = {https://doi.org/10.1038/s41467-026-74718-1},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Interpretable spatial multi-omics data integration and dimensionality reduction with SpaMV
AU - Yang Liu
AU - Kexin Ma
AU - Haoran Xu
AU - Ke Xu
AU - Yunfei Hu
AU - Zhenhan Lin
AU - Jiangli Lin
AU - Bo Han
AU - Shuai Cheng Li
AU - Zhixiang Lin
AU - Xin Zhou
AU - Lu Zhang
PY - 2026
DA - 2026/07/01
JO - Nature Communications
DO - 10.1038/s41467-026-74718-1
UR - https://doi.org/10.1038/s41467-026-74718-1
AB - Spatial multi-omics technologies have revolutionized our understanding of biological systems by providing spatially resolved molecular profiles from multiple perspectives. Existing spatial multi-omics integration methods often assume that data from different modalities share a common underlying distribution, aiming to project them into a single unified latent space. This assumption, however, can obscure the unique insights offered by each modality, thereby limiting the full potential of multi-omics analyses. To address this limitation, we present the Spatial Multi-View (SpaMV) representation learning algorithm, which captures both the shared information across modalities and the distinct, modality-specific information, enabling a more comprehensive and interpretable representation of spatial multi-omics data. Through extensive evaluation on both simulated and real-world datasets, SpaMV demonstrates superior spatial domain clustering performance and provides users with more interpretable dimension reduction for downstream analysis. Moreover, SpaMV effectively annotates cell types within clusters of a mouse thymus dataset, highlighting its effectiveness in interpretable dimensionality reduction.
SN - 2041-1723
ER -