Detecting global and local hierarchical structures in cell-cell communication using CrossChat.

Area of Science:

• Computational Biology
• Systems Biology
• Genomics

Background:

• Cell-cell communication (CCC) is crucial across biological scales, forming complex hierarchical structures.
• Existing methods for inferring CCC from single-cell RNA sequencing (scRNA-seq) and Spatial Transcriptomics overlook these inherent hierarchies.
• Understanding hierarchical CCC is vital for deciphering complex tissue functions.

Purpose of the Study:

• To develop CrossChat, a novel computational framework for inferring and analyzing hierarchical cell-cell communication structures.
• To provide a comprehensive approach for understanding both global and local hierarchical relationships within CCC.
• To demonstrate the utility of CrossChat across diverse biological datasets.

Main Methods:

• CrossChat employs two complementary approaches: multi-resolution clustering for global hierarchies and tree detection for local hierarchies.
• The framework analyzes signaling molecules, specifically ligands and receptors, to define hierarchical structures.
• Applied to nonspatial scRNA-seq and spatial transcriptomics datasets from human and mouse samples.

Main Results:

• CrossChat successfully infers and visualizes global and local hierarchical CCC structures.
• The framework identified distinct signaling properties within sender and receiver cell groups.
• Demonstrated CrossChat's capability in analyzing CCC hierarchies in COVID-19 patient data, mouse embryonic skin, and wounded mouse skin.

Conclusions:

• CrossChat offers a powerful new tool for dissecting the complexities of cell-cell communication hierarchies.
• The framework enhances our understanding of how hierarchical CCC governs tissue functions.
• CrossChat's application across multiple datasets validates its versatility and effectiveness in biological research.