Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 9 min ago
Peer-reviewedCC BY 4.0Bioinformatics

smiMAD: adaptive multi-view graph learning for paired spatial omics integration

  • Yuhao Tao1,2
  • Fei Wang1,2
  1. 1College of Computer Science and Artificial Intelligence, Fudan University , Shanghai 200433,
  2. 2Shanghai Key Laboratory of Intelligent Information Processing, Fudan University , Shanghai 200433,
Open at sourcePDFDOICite

Abstract

Motivation: Paired spatial omics technologies measure transcriptomic profiles together with epigenomic or proteomic profiles while preserving tissue context. Integration remains challenging because modalities differ in feature distributions, sparsity, and spatial dependencies. Existing methods may not fully model complementary graph views or adapt spatial constraints, limiting preservation of spatial organization and modality-specific information.

Results: We present smiMAD for paired spatial omics integration. smiMAD models spatial-neighborhood and molecular-feature graphs through graph-view-specific and common-information branches. An adaptive gate integrates graph-view and modality information, while an alignment–decorrelation objective promotes common-information embedding consistency and reduces branch redundancy. A dynamic spatial-consistency strategy progressively relaxes spatial constraints during training. Across 17 simulated and 15 real RNA–ATAC (assay for transposase-accessible chromatin) and RNA–protein datasets, smiMAD achieved competitive clustering accuracy, spatial coherence, and multimodal integration performance. Molecular marker patterns provided support for the inferred spatial regions.

Availability and implementation: The source code is available at https://github.com/fdu-wangfeilab/smiMAD, with the version used in this study archived at https://zenodo.org/records/22931624. Data are available at https://zenodo.org/records/17207079.

Related

  1. A deep learning-based multiscale integration of spatial omics with tumor morphology
    Benoît Schmauch, Loïc Herpin, Antoine Olivier +14
    Nature CommunicationsPeer-reviewedCC BY-NC-NDFull text26 citedMolecular biology & gene regulationHAL
  2. scGSI: Graph-guided self-supervised integration of paired single-cell multi-omics
    Xiang Chen, Zihan Yang, Xiaoyu Liu +2
    PLOS Computational BiologyPeer-reviewedCC BY 4.0Full textMolecular biology & gene regulationEcologyPLOS
  3. Interpretable spatial multi-omics data integration and dimensionality reduction with SpaMV
    Yang Liu, Kexin Ma, Haoran Xu +9
    Nature CommunicationsPeer-reviewedCC BY-NC-ND8 citedMolecular biology & gene regulationOpenAlex
  4. Morphodynamic domains enable integration of live morphometrics and spatial transcriptomics
    Adrien Leroy, Eric van Leen, Maria Balakireva +11
    BioRxivPreprintCC BY-NC-NDFull textMolecular biology & gene regulationHAL
  5. SubPoE: a subgraph-induced Product-of-Experts framework for spatial multi-omics integration
    Yuanfang Li, Wentao Gao, Junhao Ren +6
    Frontiers in GeneticsPeer-reviewedCC BY 4.0Full textMolecular biology & gene regulationPMC

Cite

BibTeXRIS
BibTeX
@article{tao2026smimad,
  title = {{smiMAD: adaptive multi-view graph learning for paired spatial omics integration}},
  author = {Yuhao Tao and Fei Wang},
  journal = {Bioinformatics},
  year = {2026},
  month = oct,
  doi = {10.1093/bioinformatics/btag758},
  url = {https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag758/8910347},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - smiMAD: adaptive multi-view graph learning for paired spatial omics integration
AU  - Yuhao Tao
AU  - Fei Wang
PY  - 2026
DA  - 2026/10/12
JO  - Bioinformatics
DO  - 10.1093/bioinformatics/btag758
UR  - https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag758/8910347
AB  - Motivation: Paired spatial omics technologies measure transcriptomic profiles together with epigenomic or proteomic profiles while preserving tissue context. Integration remains challenging because modalities differ in feature distributions, sparsity, and spatial dependencies. Existing methods may not fully model complementary graph views or adapt spatial constraints, limiting preservation of spatial organization and modality-specific information.  Results: We present smiMAD for paired spatial omics integration. smiMAD models spatial-neighborhood and molecular-feature graphs through graph-view-specific and common-information branches. An adaptive gate integrates graph-view and modality information, while an alignment–decorrelation objective promotes common-information embedding consistency and reduces branch redundancy. A dynamic spatial-consistency strategy progressively relaxes spatial constraints during training. Across 17 simulated and 15 real RNA–ATAC (assay for transposase-accessible chromatin) and RNA–protein datasets, smiMAD achieved competitive clustering accuracy, spatial coherence, and multimodal integration performance. Molecular marker patterns provided support for the inferred spatial regions.  Availability and implementation: The source code is available at https://github.com/fdu-wangfeilab/smiMAD, with the version used in this study archived at https://zenodo.org/records/22931624. Data are available at https://zenodo.org/records/17207079.
SN  - 1367-4803
SN  - 1367-4811
ER  -