geneSCOPE: gene spatial co-occurrence of pairwise expression.

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Spatial transcriptomics enables context-dependent gene expression analysis.
• Existing methods lack consistent scale accounting and rely on user-defined graphs, limiting comparability.
• There is a need for scalable and interpretable spatial analysis frameworks.

Purpose of the Study:

• To introduce geneSCOPE (gene Spatial Co-Occurrence of Pairwise Expression), a novel framework for gene-centric spatial analysis.
• To integrate spatial statistics and network analysis for capturing measurement scale and spatial information.
• To enhance cross-study comparability and interpretability in spatial transcriptomics.

Main Methods:

• Molecules binned on a grid with scale determined by Morisita's Iδ-width curves.
• Pairwise spatial associations quantified using Lee's L statistic.
• Spatial gene network assembly and module identification via consensus clustering.
• Cell-cell interaction identification based on spatial association and co-expression.

Main Results:

• geneSCOPE applied to human colorectal cancer and lymph node data.
• Identified spatial gene modules corresponding to microanatomical compartments.
• Highlighted intercellular interactions at tumor-stroma interfaces, including LGR5 and C3 patterns.
• Demonstrated superior concordance with known gene interactions compared to other methods.

Conclusions:

• geneSCOPE offers a scalable, interpretable, and cross-study comparable framework for spatial transcriptomics.
• The method effectively captures spatial gene expression patterns and cell-cell interactions.
• geneSCOPE advances the analysis of tissue microenvironments and disease contexts.