Sopa: a technology-invariant pipeline for analyses of image-based spatial omics

Area of Science:

• Computational biology
• Bioinformatics
• Spatial omics

Background:

• Spatial omics technologies provide single-cell resolution insights into tissue architecture.
• Data complexity and diverse technologies (e.g., Xenium, MERSCOPE, CosMX, MACSima, PhenoCycler) create analytical challenges.
• Existing tools often lack generality across different spatial omics platforms.

Purpose of the Study:

• To introduce Sopa, a technology-invariant and memory-efficient pipeline for image-based spatial omics data.
• To provide a unified visualizer for diverse spatial omics datasets.
• To streamline the analysis of cellular organization and dynamics in biological systems.

Main Methods:

• Development of Sopa, a pipeline built on the SpatialData framework.
• Optimization of key tasks including segmentation, transcript/channel aggregation, and geometric/spatial analysis.
• Integration of a unified visualizer and generation of user-friendly web reports.

Main Results:

• Sopa offers a unified approach for analyzing various spatial omics data types.
• The pipeline is memory-efficient and optimizes complex analytical tasks.
• Outputs include comprehensive data files, web reports, and visualizer files for enhanced analysis.

Conclusions:

• Sopa significantly advances spatial data analysis by unifying diverse technologies.
• The tool facilitates a deeper understanding of cellular interactions and tissue organization.
• Sopa represents a crucial step towards standardized and accessible spatial omics research.