A robust statistical approach for finding informative spatially associated pathways.

Area of Science:

• Genomics
• Computational Biology
• Systems Biology

Background:

• Spatial transcriptomics maps gene expression to tissue locations, aiding cellular function and communication studies.
• Current methods often focus on individual genes, potentially missing complex pathway interactions and spatial heterogeneity.
• Understanding spatial gene expression patterns is crucial for tissue architecture and disease pathology.

Purpose of the Study:

• To develop a novel framework for directly identifying functional pathways associated with spatial variability in transcriptomic data.
• To overcome limitations of gene selection and parameter selection in existing spatial transcriptomics analysis methods.
• To explore the heterogeneity of biological functions across spatial domains within tissues.

Main Methods:

• Adaptation of the Brownian distance covariance test to analyze spatial transcriptomic data.
• A statistical testing approach that avoids gene and parameter selection, accommodating nonlinear dependencies.
• Application of the framework to human and mouse spatial transcriptomic datasets.

Main Results:

• Identification of significant biological pathways associated with spatial gene expression variation.
• Discovery of distinct pathway patterns in inner- versus edge-cancer regions of tissues.
• Demonstration of the framework's ability to capture complex, nonlinear spatial dependencies.

Conclusions:

• The novel framework provides a new perspective for analyzing spatial transcriptomics data by focusing on pathways.
• It enhances understanding of how cells coordinate activities across spatial domains through biological pathways.
• The method contributes to a deeper comprehension of tissue architecture and disease mechanisms.