Geometry-Aware Ligand–Receptor Analysis for Spatial Transcriptomics

A geometry-aware framework for prioritizing biologically meaningful ligand–receptor interactions in spatial transcriptomics.

Overview

Cell–cell communication analyses frequently identify large numbers of statistically significant ligand–receptor interactions, making it difficult to distinguish biologically meaningful signaling programs from interactions driven by tissue composition or spatial organization alone.

This study introduces a geometry-aware framework that separates interface association from spatial localization and evaluates ligand–receptor interactions using distance-weighted boundary scoring, geometry-preserving null models, and localization-based prioritization metrics.

Applied across multiple tumor microenvironments, the framework reveals conserved and tissue-specific communication programs while providing quantitative measures of spatial constraint.

Framework

Overview of the evaluation framework used to assess interpretability in concept bottleneck models across multiple biological regimes. The framework quantifies concept selectivity, concept-space geometry, and gene-level mechanistic support, demonstrating that structured representations do not necessarily imply mechanistic validity.

Scientific Contributions

• Introduces a geometry-aware framework for ligand–receptor analysis in spatial transcriptomics.
• Separates interface enrichment from spatial localization.
• Implements geometry-preserving permutation-based null models.
• Quantifies ligand–receptor retention around cellular interfaces.
• Prioritizes communication programs using spatial evidence rather than expression alone.
• Reveals a continuum of spatial constraint rather than discrete communication regimes.
• Identifies conserved and tissue-specific signaling programs across tumor types.

Study Design

The framework was applied to four spatial transcriptomic cancer datasets:

• Breast cancer
• Colorectal cancer (CRC)
• Melanoma
• Pancreatic ductal adenocarcinoma (PDAC)

Each dataset was analyzed using an identical computational workflow to enable direct cross-tissue comparisons.

Repository Resources

Manuscript

• Paper.pdf Geometry-aware ligand-receptor analysis distinguishes interface association from spatial localization and reveals a continuum of tumor communication

Publication Figures

• Figure1_Geometry_Aware_Framework.pdf
• Figure2A_Top_Pathways_By_Tissue.pdf
• Figure2B_Recurrent_Pathway_Composition.pdf
• Figure3A_Observed_vs_Null.pdf
• Figure3B_Null_Gap_Distribution.pdf
• Figure3C_Effect_Size_vs_Significance.pdf
• Figure4A_Boundary_vs_Geometry_Aware_Score.pdf
• Figure4B_Retention_Distribution.pdf
• Figure4C_Sample_Level_Retention.pdf
• Figure5A_Retention_vs_Diffuse_Fraction.pdf
• Figure5B_Retention_vs_Null_Gap.pdf
• Figure5C_Retention_By_Tissue.pdf
• Figure6A_Conserved_Pathway_Retention.pdf
• Figure6B_Conserved_Pathway_Composition.pdf
• Figure6C_Tissue_Specific_Pathway_Retention.pdf
• Figure6D_Global_Regime_Composition.pdf

Analysis Modules

Notebook	Scientific Focus
breast_cancer_geometry_aware_lr_analysis.ipynb	Geometry-aware ligand–receptor analysis in breast cancer spatial transcriptomics.
colorectal_cancer_geometry_aware_lr_analysis.ipynb	Geometry-aware ligand–receptor analysis in colorectal cancer.
melanoma_geometry_aware_lr_analysis.ipynb	Characterization of geometry-aware cell–cell communication in melanoma.
pdac_geometry_aware_lr_analysis.ipynb	Analysis of spatially constrained signaling programs in pancreatic ductal adenocarcinoma.
manuscript_figure_generation.ipynb	Reproduction of publication figures and cross-tissue summaries.

Key Findings

• Interface enrichment and spatial localization are distinct properties.
• Strong interface association alone does not establish spatial specificity.
• Geometry-aware refinement substantially changes ligand–receptor prioritization.
• Conserved signaling pathways are deployed under different spatial constraints.
• Tumor communication is better described as a continuum of spatial constraint.

Reproducibility

All analyses are provided as documented Jupyter notebooks together with publication figures and manuscript resources. The repository is intended as a reproducible research resource for spatial transcriptomics, tumor microenvironment analysis, and cell–cell communication studies.

Citation

Yepes S. (2026). Geometry-aware ligand–receptor analysis reveals tumor communication patterns.

DOI: https://doi.org/10.5281/zenodo.19476574