MONTAGE: A Computation Framework to Identify Spatially Resolved Functional Enrichment Gradients in the Tissue...

Area of Science:

• Computational biology
• Cancer research
• Spatial transcriptomics

Background:

• Single-cell spatial-omics advances tumor microenvironment understanding.
• Current methods offer limited insight into functional and clinical relevance of cellular organization.

Purpose of the Study:

• Introduce MONTAGE, a computational framework to reconstruct, functionally analyze, and identify clinically relevant cellular spatial communities (SCs).
• Enable deeper understanding of tumor heterogeneity and its clinical implications.

Main Methods:

• MONTAGE integrates single-cell transcriptomics data to generate a gene signature matrix of SCs.
• Reconstructs tissue maps reflecting sequential SC compositional changes along spatial gradients.
• Applies MONTAGE signature matrix to deconvolve bulk and spot-based spatial transcriptomic data.
• Integrates clinical metadata to assess SC clinical relevance.

Main Results:

• MONTAGE reveals distinct SCs enriched with specific cell types (e.g., malignant cells, granulocytes, macrophages, fibroblasts).
• Identified shifts in SC composition along epithelial-mesenchymal transition gradients.
• Demonstrated clinical prognostic significance of identified SCs in head and neck cancer.

Conclusions:

• MONTAGE provides a novel framework for analyzing cellular spatial organization and its functional and clinical relevance.
• This approach enhances the utility of spatial transcriptomics for cancer research.
• MONTAGE facilitates the discovery of novel biomarkers and therapeutic targets based on cellular communities.