CellChat

CellChat infers and quantitatively analyzes cell–cell communication networks from single-cell RNA sequencing (scRNA-seq) data to characterize signaling inputs and outputs among cells.

Key Features:

• R implementation: Implemented as an R toolkit for analysis of scRNA-seq datasets.
• Ligand–receptor–cofactor database: Contains an extensive database of interactions among ligands, receptors, and their cofactors with explicit representation of known heteromeric molecular complexes.
• Mass action modeling: Uses mass action models to quantitatively infer intercellular communication network strengths.
• Network analysis and pattern recognition: Applies network analysis and pattern recognition techniques to identify signaling inputs, outputs, and communication patterns.
• Manifold learning and quantitative contrasts: Employs manifold learning and quantitative contrasts to classify signaling pathways as conserved or context-specific across datasets.
• Per-cell signaling inference: Infers major signaling inputs and outputs for individual cells.
• Visualization of communication networks: Provides functions to visualize inferred cell–cell communication networks and signaling patterns.

Scientific Applications:

• Cell–cell communication mapping: Reconstruction and quantitative analysis of intercellular signaling networks from scRNA-seq data.
• Pathway classification across contexts: Classification of conserved versus context-specific signaling pathways across multiple datasets using manifold learning and quantitative contrasts.
• Discovery of signaling patterns: Extraction of known and novel signaling patterns via network analysis and pattern recognition.
• Developmental and injury biology: Application to mouse skin scRNA-seq datasets, including embryonic development and adult wound healing studies.
• Comparative tissue atlasing: Support for building comparative cell–cell communication atlases across diverse tissues.

Methodology:

Leverages a ligand–receptor–cofactor interaction database with heteromeric complex representation, applies mass action models to infer quantitative interaction strengths, and uses network analysis, pattern recognition, manifold learning, and quantitative contrasts to analyze and classify signaling pathways.