SpatialData: an open and universal data framework for spatial omics

Area of Science:

• Spatial biology
• Genomics
• Multi-omics data analysis

Background:

• Spatially resolved omics technologies offer unprecedented insights into tissue organization.
• Existing uni- and multimodal spatial omics datasets present challenges due to large volumes, data heterogeneity, and lack of flexible, spatially aware structures.

Purpose of the Study:

• To introduce SpatialData, a novel framework addressing the challenges in handling spatial omics data.
• To provide a unified, extensible, and multiplatform solution for managing and analyzing spatial omics datasets.

Main Methods:

• Development of SpatialData, a framework featuring a unified file format.
• Implementation of lazy data representation for larger-than-memory datasets.
• Inclusion of data transformations and alignment to common coordinate systems.

Main Results:

• SpatialData facilitates seamless spatial annotations and cross-modal aggregation.
• The framework supports integrative analysis of multimodal spatial omics studies.
• Demonstrated utility through vignettes, including a breast cancer study integrating Xenium and Visium data.

Conclusions:

• SpatialData offers a robust solution for managing and analyzing complex spatial omics data.
• The framework enhances the accessibility and utility of spatial omics technologies for biological research.