Interpretable niche-based cell‒cell communication inference using multi-view graph neural networks

Area of Science:

• Computational Biology
• Genomics
• Systems Biology

Background:

• Cell-cell communication (CCC) is vital for biological systems.
• Existing CCC inference methods often overlook niche heterogeneity within cell types.
• Understanding niche-specific interactions is crucial for biological and disease contexts.

Purpose of the Study:

• To develop a novel tool, STCase, for single-cell/spot level CCC inference using spatial transcriptomics.
• To identify distinct cellular niches and their specific CCC events.
• To overcome limitations of current methods that aggregate cells by type.

Main Methods:

• Spatial Transcriptomics-based cell-cell Communication And Subtype Exploration (STCase) tool.
• Utilizes an interpretable multi-view graph neural network with CCC-aware attention.
• Analyzes CCC events at the single-cell/spot resolution.

Main Results:

• STCase accurately captures known immune-related CCC events in human bronchial glands.
• Identified three distinct niches in oral squamous cell carcinoma, previously obscured by other methods.
• Discovered niche-specific CCC events potentially impacting oral cancer prognosis.

Conclusions:

• STCase provides a powerful method for high-resolution CCC analysis.
• The tool effectively identifies niche heterogeneity and its role in biological processes.
• Findings highlight the importance of niche-specific CCC in disease progression and suggest new avenues for therapeutic strategies.