SpNeigh: spatial neighborhood and differential expression analysis for high-resolution spatial transcriptomics

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Spatial transcriptomics technologies like Xenium, MERFISH, and Visium HD offer high-resolution gene expression profiling within tissue architecture.
• Existing computational methods often overlook local tissue context, such as boundaries, neighborhoods, and gradients, limiting in-depth spatial analysis.

Purpose of the Study:

• To introduce SpNeigh, an R package designed for spatial neighborhood analysis and spatially aware differential expression modeling.
• To provide tools for boundary detection, neighborhood extraction, and gradient-based statistical testing in spatial transcriptomics data.

Main Methods:

• Development of the SpNeigh R package implementing boundary detection, spatial neighborhood extraction, and distance-based weighting.
• Integration of spline-based regression for smooth spatial modeling and a spatial enrichment index for gene identification.
• Application of SpNeigh to diverse spatial transcriptomics datasets from mouse brain, human breast cancer, and human liver.

Main Results:

• SpNeigh effectively models local tissue context, enabling spatially aware differential expression analysis.
• The package identified intermediate cell populations at tissue interfaces, differences in immune microenvironments, and spatially zoned gene expression patterns.
• Demonstrated utility across multiple spatial transcriptomics platforms and tissue types.

Conclusions:

• SpNeigh provides a flexible and interpretable framework for dissecting spatial gene expression dynamics.
• The package enhances the analysis of complex tissues by explicitly modeling spatial relationships and context.
• Facilitates deeper understanding of tissue microenvironments and cellular heterogeneity.