CellChat Skill

Comprehensive assistance with CellChat development and cell-cell communication analysis, generated from official documentation.

When to Use This Skill

This skill should be triggered when:

• •Working with CellChat - Analyzing cell-cell communication in single-cell RNA-seq data
• •Data preprocessing - Converting Seurat, SingleCellExperiment, or AnnData objects to CellChat format
• •Communication inference - Computing cell-cell communication probabilities and networks
• •Comparative analysis - Comparing communication patterns across multiple conditions or datasets
• •Visualization - Creating network plots, hierarchy plots, and chord diagrams
• •Advanced analysis - Performing centrality analysis, pattern identification, and spatial analysis
• •Tool integration - Interfacing CellChat with other single-cell analysis toolkits
• •Debugging CellChat - Troubleshooting data input issues and analysis problems

Quick Reference

Essential Setup Patterns

Pattern 1: Basic CellChat Setup

library(CellChat)
library(patchwork)
options(stringsAsFactors = FALSE)

Pattern 2: Create CellChat from Data Matrix

# Load normalized data matrix and metadata
data.input = data_humanSkin$data  # normalized data matrix
meta = data_humanSkin$meta         # dataframe with cell metadata

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

Pattern 3: Create from Seurat Object

# Extract normalized data from Seurat
data.input <- seurat_object[["RNA"]]@data  # or seurat_object[["RNA"]]$data for v5+
labels <- Idents(seurat_object)
meta <- data.frame(group = labels, row.names = names(labels))

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

Pattern 4: Create from SingleCellExperiment

# Extract data from SCE
data.input <- SingleCellExperiment::logcounts(object)
meta <- as.data.frame(SingleCellExperiment::colData(object))
meta$labels <- meta[["sce.clusters"]]

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

Pattern 5: Create from AnnData (Python)

library(anndata)
ad <- read_h5ad("scanpy_object.h5ad")
counts <- t(as.matrix(ad$X))

# Normalize if needed
library.size <- Matrix::colSums(counts)
data.input <- as(log1p(Matrix::t(Matrix::t(counts)/library.size) * 10000), "dgCMatrix")
meta <- ad$obs
meta$labels <- meta[["ad_clusters"]]

Core Analysis Workflow

Pattern 6: Set Up Database and Preprocess

# Load CellChat database
CellChatDB <- CellChatDB.human
cellchat@DB <- CellChatDB

# Subset expression data to signaling genes
cellchat <- subsetData(cellchat)

# Normalize data if using raw counts
cellchat <- normalizeData(cellchat)

Pattern 7: Infer Cell-Cell Communication

# Compute communication probability
cellchat <- computeCommunProb(cellchat)

# Compute pathway-level communication
cellchat <- computeCommunProbPathway(cellchat)

# Filter significant communications
cellchat <- filterCommunication(cellchat, min.cells = 10)

Pattern 8: Aggregate Network for Visualization

# Calculate aggregated network
cellchat <- aggregateNet(cellchat)

# Extract interaction matrix
net <- cellchat@net$count  # number of interactions
weight <- cellchat@net$weight  # interaction weights

Comparative Analysis Patterns

Pattern 9: Compare Multiple Datasets

# Load multiple CellChat objects
object.list <- list(E13 = cellchat.E13, E14 = cellchat.E14)

# Merge objects
cellchat <- mergeCellChat(object.list, add.names = names(object.list), cell.prefix = TRUE)

# Compare interaction strength
netVisual_diffInteraction(object.list, slot.name = "netP")

Pattern 10: Lift Cell Groups for Comparison

# Define unified cell labels
group.new = levels(cellchat.E14@idents)
cellchat.E13 <- liftCellChat(cellchat.E13, group.new)

# Merge lifted objects
object.list <- list(E13 = cellchat.E13, E14 = cellchat.E14)
cellchat <- mergeCellChat(object.list, add.names = names(object.list))

Visualization Patterns

Pattern 11: Circle Plot for Signaling Pathways

# Circle plot of specific pathway
pathways.show <- c("WNT")
weight.max <- getMaxWeight(object.list, slot.name = "netP", attribute = pathways.show)

par(mfrow = c(1,2), xpd=TRUE)
for (i in 1:length(object.list)) {
  netVisual_aggregate(object.list[[i]], signaling = pathways.show,
                     layout = "circle", edge.weight.max = weight.max[1],
                     edge.width.max = 10, signaling.name = paste(pathways.show, names(object.list)[i]))
}

Pattern 12: Hierarchy Plot

# Hierarchy plot with specific receiver cells
vertex.receiver = seq(1,10)  # Left portion for dermal cells
par(mfrow = c(1,2), xpd=TRUE)
for (i in 1:length(object.list)) {
  netVisual_aggregate(object.list[[i]], signaling = pathways.show,
                     vertex.receiver = vertex.receiver, edge.weight.max = weight.max[1])
}

Pattern 13: Chord Diagram

# Group cells for chord diagram
group.merged <- c(rep("Dermal", 10), rep("Epidermal", 3))
names(group.merged) <- levels(object.list[[1]]@idents)

# Create chord diagram
for (i in 1:length(object.list)) {
  netVisual_chord_cell(object.list[[i]], signaling = pathways.show,
                      group = group.merged, title.name = paste0(pathways.show, " signaling - ", names(object.list)[i]))
}

Advanced Analysis Patterns

Pattern 14: Identify Communication Patterns

# Select optimal number of patterns
selectK(cellchat, slot.name = "netP", pattern = "outgoing", k.range = seq(2, 10))

# Identify patterns
cellchat <- identifyCommunicationPatterns(cellchat, slot.name = "netP",
                                       pattern = c("outgoing", "incoming"), k = 4)

Pattern 15: Compute Centrality Measures

# Compute centrality for signaling networks
cellchat <- computeCentrality(cellchat, slot.name = "netP")

# Access centrality measures
centrality_outgoing <- cellchat@netP$centrality_outgoing
centrality_incoming <- cellchat@netP$centrality_incoming

Pattern 16: Data Smoothing for Shallow Sequencing

# Smooth data to reduce dropout effects
cellchat <- smoothData(cellchat, method = "netSmooth", alpha = 0.5)

# Use smoothed data for communication inference
cellchat <- computeCommunProb(cellchat, raw.use = FALSE)

Key Concepts

CellChat Object Structure

• •@data: Raw and processed expression data
• •@meta: Cell metadata and group labels
• •@DB: Ligand-receptor interaction database
• •@LR: Significant ligand-receptor pairs
• •@net: Communication probability arrays
• •@netP: Pathway-level communication networks
• •@idents: Cell group identities

Communication Probability

• •Computed using gene expression of ligands and receptors
• •Incorporates population size and expression thresholds
• •Can be computed at ligand-receptor or pathway level
• •Supports spatial distance constraints

Database Types

• •CellChatDB.human: Human ligand-receptor interactions
• •CellChatDB.mouse: Mouse interactions
• •CellChatDB.zebrafish: Zebrafish interactions
• •Custom databases can be integrated

Reference Files

This skill includes comprehensive documentation organized by topic:

Core Documentation

• •getting_started.md - Introduction and basic workflow setup
• •core_classes.md - CellChat object structure and class definitions
• •data_preprocessing.md - Data normalization, smoothing, and formatting

Analysis Methods

• •communication_inference.md - Computing communication probabilities and networks
• •advanced_methods.md - Network aggregation, merging, and lifting methods
• •centrality_analysis.md - Network centrality measures and importance analysis
• •network_analysis.md - Network topology and structure analysis

Comparative Analysis

• •comparison_analysis.md - Multiple dataset comparison with different cell compositions
• •enrichment_analysis.md - Functional enrichment and pathway analysis

Specialized Features

• •spatial_functions.md - Spatial transcriptomics analysis methods
• •spatial_tutorials.md - Spatial analysis workflows and examples
• •database_management.md - Custom database integration and management
• •tool_integration.md - Integration with Seurat, Scanpy, and other tools

Visualization

• •visualization_tools.md - Plotting functions and network visualizations
• •plotting_utilities.md - Helper functions for custom plots

Technical Details

• •ligand_receptor.md - Ligand-receptor interaction details
• •mathematical_functions.md - Mathematical foundations and algorithms
• •cpp_source_code.md - C++ source code for performance-critical functions
• •r_source_code.md - R function implementations

Additional Resources

• •expression_analysis.md - Gene expression analysis utilities
• •dimensionality_reduction.md - Data reduction methods
• •interaction_analysis.md - Detailed interaction analysis methods
• •other.md - Additional utilities and miscellaneous functions

Working with This Skill

For Beginners

1. •Start with getting_started.md for basic CellChat setup
2. •Learn data preparation from data_preprocessing.md and tool_integration.md
3. •Follow basic workflow using patterns 1-8 from Quick Reference
4. •Explore visualizations with visualization_tools.md

For Intermediate Users

1. •Master comparative analysis using comparison_analysis.md
2. •Learn advanced methods from advanced_methods.md and network_analysis.md
3. •Apply pattern identification using patterns 14-15
4. •Integrate with other tools using tool_integration.md

For Advanced Users

1. •Develop custom databases using database_management.md
2. •Implement spatial analysis with spatial_functions.md
3. •Optimize performance using cpp_source_code.md insights
4. •Extend functionality with mathematical_functions.md

Navigation Tips

• •Use specific searches - Search for function names or analysis types
• •Follow workflow order - Data preprocessing → Communication inference → Analysis → Visualization
• •Check examples - Each reference file contains code examples with language annotations
• •Cross-reference - Related topics are linked across documentation files

Common Use Cases

Single Dataset Analysis

1. •Load and preprocess data (Patterns 1-6)
2. •Infer communications (Pattern 7)
3. •Visualize networks (Patterns 11-13)
4. •Identify patterns (Pattern 14)

Multi-Dataset Comparison

1. •Create CellChat objects for each dataset
2. •Lift cell groups if needed (Pattern 10)
3. •Merge objects (Pattern 9)
4. •Compare communications (Pattern 9, 11-13)

Spatial Analysis

1. •Load spatial data with coordinates
2. •Use distance constraints in computeCommunProb
3. •Visualize spatial communication patterns
4. •Analyze neighborhood-specific interactions

Tool Integration

• •Seurat: Use Pattern 3 for direct conversion
• •Scanpy/AnnData: Use Pattern 4 or 5 for Python integration
• •SingleCellExperiment: Use Pattern 4 for Bioconductor workflows

Resources

Documentation Structure

• •references/: Complete categorized documentation with examples
• •scripts/: Add your automation scripts here
• •assets/: Store templates and example projects

Code Examples

All code examples in this skill are:

• •Extracted from official documentation
• •Language-annotated for proper syntax highlighting
• •Tested and verified in real-world scenarios
• •Contextually organized by difficulty and use case

Updating

To refresh this skill with updated documentation:

1. •Re-run the scraper with the same configuration
2. •The skill will be rebuilt with the latest information
3. •Local enhancements preserve custom improvements

Notes

• •This skill was generated from official CellChat documentation
• •Reference files maintain original structure and examples
• •Code examples include proper language detection
• •Quick reference patterns are extracted from common usage examples
• •All examples are practical and ready-to-use