Abstract Multimodal spatial omics enables systematic characterization of tissue organization by jointly profiling transcriptomic, proteomic, metabolomic, and other spatially resolved modalities within their spatial context. A central challenge in realizing this potential is achieving robust spatial alignment across modalities and sections. Although numerous alignment methods have been developed, most are designed for single-modality sections or specific modality combinations, with few enabling modality-agnostic alignment. Cross-modal alignment remains difficult due to the absence of shared molecular features, partial spatial overlap, and nonrigid tissue deformations. To address these challenges, we introduce SPOmiAlign, a modality-agnostic framework for spatial multimodal alignment, enabled by a feature-matching foundation model that serves as a general computational primitive for spatial multi-omics alignment. The framework enables accurate cross-modal spatial alignment without manual intervention or modality-specific tuning. Across diverse multimodal benchmarks, SPOmiAlign consistently achieves higher alignment accuracy than existing methods. We further demonstrate its utility through automated registration to a common coordinate framework, enabling standardized anatomical annotation. Finally, applying SPOmiAlign to integrate spatial transcriptomic, proteomic, and metabolomic data in mouse brain facilitates the identification of spatial domains that were difficult to resolve with less accurate registration, highlighting its utility for multi-omic integration and biological interpretation.
Wang et al. (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: