N-Orbit: towards a universal model and metric for comparing tissue microenvironments.

Area of Science:

• Computational biology
• Spatial omics
• Tissue architecture

Background:

• Spatial omics technologies offer deep insights into tissue microenvironments.
• A unified framework for modeling and comparing tissue architecture is needed.

Purpose of the Study:

• Introduce N-Orbit, a mathematical model for tissue microenvironment analysis.
• Develop a vector-based approach to encode cell-type composition and spatial relationships.

Main Methods:

• N-Orbit encodes cellular neighborhoods as vectors for distance calculations.
• Benchmarked N-Orbit using spatial omics datasets with ground-truth neighborhoods and clinical outcomes.

Main Results:

• N-Orbit outperforms cell-type-enrichment metrics in discriminating neighborhood types.
• The model accurately predicts clinical variables and identifies homologous structures across species.
• N-Orbit enhances interpretability by linking neighborhoods to spatial motifs.

Conclusions:

• N-Orbit provides a robust framework for modeling and comparing tissue architecture.
• The model significantly advances the analysis of spatial omics data.
• N-Orbit has potential for understanding tissue remodeling in development, disease, and evolution.