WGCNA基因网络igraph可视化：模块着色与节点标签/大小设置问询

Alright, let's walk through how to solve these two visualization challenges for your WGCNA-derived network. I'll assume you're working in R since that's the standard environment for WGCNA and igraph together.

1. Color Nodes by Source/Target Module #

First, we need to make sure your igraph nodes have the right attributes to map module colors. There are two common scenarios here:

Scenario A: Source nodes belong to Module1 (single color) and Target nodes belong to Module2 (single color) #

If all source nodes share Module1's color and all target nodes share Module2's color, you can directly assign colors based on node type:

# Define your module colors (use the actual colors from your WGCNA results)
module1_color <- "#E41A1C"  # Example: WGCNA's "red" module
module2_color <- "#377EB8"  # Example: WGCNA's "blue" module

# Assign color based on node type (make sure your graph has a "node_type" attribute first)
V(g)$color <- ifelse(V(g)$node_type == "source", module1_color, module2_color)

Scenario B: Source/Target nodes have sub-modules with unique colors #

If your Module1/Module2 are made up of smaller sub-modules (each with their own color from WGCNA), map the pre-existing WGCNA module colors to your igraph nodes:

# Assume you have a WGCNA module color vector named `moduleColors`, where names are gene IDs
# Match node names (gene IDs) to their corresponding module colors
V(g)$color <- moduleColors[match(V(g)$name, names(moduleColors))]

Note: If your graph doesn't have a node_type attribute yet, you can create it by extracting source and target nodes from the edge list:

# Extract unique source and target nodes from the graph
source_nodes <- unique(head_of(g))
target_nodes <- unique(tail_of(g))

# Add the node_type attribute to your igraph object
V(g)$node_type <- ifelse(V(g)$name %in% source_nodes, "source", "target")

2. Set Node Size by Degree & Show Labels Only for Source Nodes #

Next, we'll adjust node size based on degree and control label visibility:

Step 1: Set node size based on degree #

Degree is a measure of how many connections a node has. You can use raw degree values or scale them (like log scaling) to avoid extreme size differences:

# Option 1: Use raw degree values (tweak the multiplier to adjust size)
V(g)$size <- degree(g) * 1.5

# Option 2: Log scale for large degree differences (prevents tiny/huge nodes)
V(g)$size <- log(degree(g) + 1) * 4  # +1 avoids log(0) issues

Step 2: Show labels only for source nodes #

We'll set target node labels to empty strings so they don't display:

# Assign labels: source nodes show their name, target nodes show nothing
V(g)$label <- ifelse(V(g)$node_type == "source", V(g)$name, "")

Full Plotting Code #

Put it all together and plot your network with clean styling:

plot(g,
     layout = layout_with_fr(),  # Fruchterman-Reingold layout (great for network structure)
     vertex.label.cex = 0.7,  # Adjust label size
     vertex.frame.color = "white",  # Remove node borders (or set to a color you prefer)
     edge.color = "gray80",  # Fade edges to keep focus on nodes
     vertex.label.color = "black"  # Ensure labels are readable
)

内容的提问来源于stack exchange，提问作者PesKchan