Cellchat Pkg Local Complete

CellChat-Pkg-Local-Complete Skill

Comprehensive assistance with CellChat package for cell-cell communication analysis, generated from official documentation and source code.

When to Use This Skill

This skill should be triggered when:

Data Preparation & Object Creation:

• Converting single-cell data from Seurat, SingleCellExperiment, or AnnData objects to CellChat
• Preparing normalized expression matrices and cell metadata for CellChat analysis
• Setting up CellChat objects from different data formats (count matrices, spatial data)
• Handling data input issues from Python/Scanpy or other single-cell tools

Cell-Cell Communication Analysis:

• Inferring ligand-receptor interactions and signaling pathways
• Computing communication probabilities and network analysis
• Analyzing spatially resolved cell-cell communication
• Performing comparative analysis across conditions or time points

Network Analysis & Visualization:

• Visualizing communication networks using hierarchy, circle, or chord diagrams
• Computing network centrality measures to identify key signaling players
• Creating spatial plots of communication on tissue sections
• Extracting and filtering specific communications of interest

Advanced Applications:

• Integrating custom ligand-receptor databases
• Analyzing contact-dependent signaling
• Projecting expression data onto protein-protein interaction networks
• Batch correction and multi-sample analysis

Troubleshooting & Optimization:

• Debugging issues with object creation or data conversion
• Optimizing parameters for communication inference
• Handling spatial transcriptomics data from different platforms
• Understanding and customizing analysis workflows

Quick Reference

Core Examples

Example 1: Create CellChat object from Seurat

library(CellChat)
cellchat <- createCellChat(object = seurat.obj, group.by = "ident", assay = "RNA")

Example 2: Create from expression matrix

cellchat <- createCellChat(object = data.input, meta = meta, group.by = "labels")

Example 3: Set up database and preprocess

cellchat@DB <- CellChatDB.human
cellchat <- subsetData(cellchat)
cellchat <- identifyOverExpressedGenes(cellchat)
cellchat <- identifyOverExpressedInteractions(cellchat)

Example 4: Infer communication network

cellchat <- computeCommunProb(cellchat)
cellchat <- computeCommunProbPathway(cellchat)
cellchat <- aggregateNet(cellchat)

Example 5: Extract communications

# All communications
df.net <- subsetCommunication(cellchat)

# Specific pathways
df.net <- subsetCommunication(cellchat, signaling = c("WNT", "TGFb"))

# Specific cell groups
df.net <- subsetCommunication(cellchat, sources.use = c(1,2), targets.use = c(4,5))

Example 6: Visualize signaling network

netVisual(cellchat, signaling = "TGFb", layout = "hierarchy")

Example 7: Network centrality analysis

cellchat <- netAnalysis_computeCentrality(cellchat, slot.name = "netP")
netAnalysis_signalingRole_network(cellchat, signaling = "TGFb")

Example 8: Spatial data setup

cellchat <- createCellChat(object = data.input, meta = meta, group.by = "labels",
                          coordinates = coordinates, spatial.factors = spatial.factors)

Example 9: Data conversion from AnnData

library(anndata)
ad <- read_h5ad("scanpy_object.h5ad")
counts <- t(as.matrix(ad$X))
library.size <- Matrix::colSums(counts)
data.input <- as(log1p(Matrix::t(Matrix::t(counts)/library.size) * 10000), "dgCMatrix")
meta <- ad$obs

Example 10: Rank signaling pathways

rankNet(cellchat, mode = "single", measure = "weight")

Key Concepts

Core CellChat Objects:

• cellchat@data: Normalized expression data
• cellchat@meta: Cell metadata and group information
• cellchat@net: Ligand-receptor level communication network
• cellchat@netP: Pathway level communication network
• cellchat@idents: Cell group identities

Data Requirements:

• Expression Matrix: Genes in rows, cells in columns, normalized data required
• Cell Metadata: Dataframe with cell labels and sample information
• Spatial Data: Coordinates and distance factors for spatial analysis

Communication Inference:

• Probability Calculation: Based on mass action law integrating expression and prior knowledge
• Permutation Testing: Statistical significance assessment
• Pathway Aggregation: Summarizing multiple L-R pairs per signaling pathway

Network Analysis Types:

• Functional Networks: Based on inferred communication probabilities
• Structural Networks: Based on ligand-receptor interaction structure
• Centrality Measures: Identifying key senders, receivers, mediators, influencers

Reference Files

This skill includes comprehensive documentation in references/:

api_functions.md (134 pages)

Complete API documentation for all CellChat functions:

• subsetCommunication: Extract specific cell-cell communications
• computeAveExpr: Calculate average expression per cell group
• netVisual_embedding: 2D visualization of signaling manifolds
• rankNet: Rank signaling networks by information flow
• netAnalysis_computeCentrality: Network centrality analysis
• computeExpr_antagonist: Model antagonist effects

cpp_source.md (2 pages)

C++ source code for performance-critical operations:

• ComputeSNN: Shared nearest neighbor computation
• Rcpp bindings for high-performance calculations

r_source.md (10 pages)

R source code for core functionality:

• visualization: Network visualization functions and themes
• Color palettes and plotting utilities
• Core analysis algorithms

tutorials.md (27 pages)

Comprehensive tutorials covering:

• Spatial transcriptomics analysis: Complete workflow for spatial data
• Data preparation: From various single-cell formats
• Network inference: Step-by-step communication analysis
• Visualization: Multiple plotting options and customization
• Advanced applications: Custom databases and comparative analysis

Working with This Skill

For Beginners

1. Start with Data Preparation: Use the tutorials to understand data input requirements
2. Basic Workflow: Follow the spatial transcriptomics tutorial for a complete example
3. Simple Visualizations: Begin with circle plots and hierarchy diagrams
4. Reference Functions: Use api_functions.md for detailed parameter explanations

For Intermediate Users

1. Custom Analyses: Explore different type parameters in computeAveExpr
2. Advanced Visualizations: Try chord diagrams and spatial plots
3. Network Analysis: Use centrality measures to identify key regulators
4. Multi-sample Analysis: Learn comparative analysis across conditions

For Advanced Users

1. Custom Databases: Update CellChatDB with domain-specific interactions
2. Spatial Applications: Adapt parameters for different spatial technologies
3. Performance Optimization: Use C++ functions for large datasets
4. Integration Workflows: Combine with other single-cell analysis tools

Navigation Tips

• Function Lookup: Search api_functions.md for specific function names
• Parameter Details: Each function entry includes complete parameter descriptions
• Code Examples: All tutorials include reproducible R code
• Troubleshooting: Check the tutorials section for common issues and solutions

Resources

references/

Organized documentation extracted from official sources:

• Detailed explanations of all parameters and return values
• Code examples with proper language annotations
• Links to original documentation for deeper exploration
• Table of contents for quick navigation

scripts/

Add helper scripts here for common automation tasks:

• Data conversion utilities
• Custom visualization functions
• Batch processing workflows

assets/

Store templates and examples:

• Example datasets in proper format
• Custom ligand-receptor databases
• Configuration files for different analysis types

Common Workflows

Basic CellChat Analysis

# 1. Create object
cellchat <- createCellChat(object = data.input, meta = meta, group.by = "labels")

# 2. Set database and preprocess
cellchat@DB <- CellChatDB.human
cellchat <- subsetData(cellchat)
cellchat <- normalizeData(cellchat)

# 3. Identify over-expressed genes/interactions
cellchat <- identifyOverExpressedGenes(cellchat)
cellchat <- identifyOverExpressedInteractions(cellchat)

# 4. Infer communication
cellchat <- computeCommunProb(cellchat)
cellchat <- computeCommunProbPathway(cellchat)
cellchat <- aggregateNet(cellchat)

# 5. Visualize
netVisual(cellchat, signaling = "TGFb")

Spatial Transcriptomics Analysis

# 1. Create with spatial information
cellchat <- createCellChat(object = data.input, meta = meta, group.by = "labels",
                          coordinates = coordinates, spatial.factors = spatial.factors)

# 2. Spatial communication inference
cellchat <- computeCommunProb(cellchat, contact.dependent = FALSE,
                             interaction.range = 250, scale.distance = 1)

# 3. Spatial visualization
netVisual_spatial(cellchat, signaling = "WNT")

Comparative Analysis

# 1. Create separate objects for each condition
cellchat1 <- createCellChat(object = data.input1, meta = meta1, group.by = "labels")
cellchat2 <- createCellChat(object = data.input2, meta = meta2, group.by = "labels")

# 2. Process each object
# ... (preprocessing steps for each)

# 3. Merge for comparison
cellchat <- mergeCellChat(cellchat1, cellchat2, add.names = c("LS", "NL"))

# 4. Compare signaling
rankNet(cellchat, mode = "comparison", comparison = c(1, 2))

Notes

• Coverage: This skill covers 173 files including tutorials, API docs, and source code
• Languages: Primarily R with some C++ components
• Data Types: Supports scRNA-seq, spatial transcriptomics, and bulk data
• Integration: Works with Seurat, SingleCellExperiment, AnnData objects
• Documentation: Preserves original structure and examples from official sources

Updating

To refresh this skill with updated documentation:

1. Restart the local CellChat documentation server
2. Re-run the documentation scraper with the same configuration
3. The skill will be rebuilt with the latest information and examples