R igraph vertex coloring

Question

I'm trying to color the vertices according to the dummy variable [color_thresh]. Data looks like the following:

I have tried several suggested methods, but the closest (and simplest) I can do produces the following:

Here is my code for reproducing data and the graph above:

library(igraph)
library(data.tree)

Level_0 = c("TAC_310B_124","TAC_310B_124","TAC_310B_124","TAC_310B_124","TAC_310B_124","TAC_310B_124","TAC_310B_124","TAC_310B_124","TAC_310B_124")
Level_1 = c("","SC_776E_86","SS_A66D_9","SC_776E_86","SC_776E_86","SC_776E_86","SC_776E_86","SC_776E_86","SC_776E_86")
Level_2 = c("","","","TGW_1DD0_42","TGW_1DD0_42","TGW_1DD0_42","TGW_1DD0_42","TGW_1DD0_42","TGW_1DD0_42")
Level_3 = c("","","","","CORB_F2E4_17","CORB_F2E4_22","CORB_AE16_10921","CORB_AE16_10921","CORB_AE16_10921")
Level_4 = c("","","","","","","","CORB_D81E_45","CORB_D81E_45")
Level_5 = c("","","","","","","","","SC_B949_3156")
color_thresh = c(1,0,0,0,0,0,1,0,1)

dataset <- data.frame(Level_0,Level_1,Level_2,Level_3,Level_4,Level_5,color_thresh)

dataset$pathString <- paste(dataset$Level_0, 
                            dataset$Level_1, 
                            dataset$Level_2, 
                            dataset$Level_3, 
                            dataset$Level_4, 
                            dataset$Level_5, 
                            sep = "/")

g <- as.Node(dataset)
g <- as.igraph(g, directed=TRUE)

V(g)$color <- ifelse(dataset$color_thresh==1, "red", "green")

plot.igraph(g, vertex.label.color="black", vertex.label.dist=0, vertex.label.family="Helvetica", vertex.frame.color="white", vertex.size=25, layout=layout_as_tree)

The colors change when i sort the data frame by [color_thresh], but still doesn't give me the right colors. Any comments helping me to understand how the conditional coloring works is much appreciated! Thank you so much in advance.

Thanks for your comment, patL, but can you be more specific? How can i sort the color according to the nodes? — Daniel Safai
If you have generated a vector for colors called colors it could be done with V(g)$color[colors] and it will sort your colors according to the colors vector. — patL
The first approach is better as the color_thresh variable is dynamic so the graph has to update the colors with the most recent value. Thanks again! I would never have come up with such elegant solution as the first you gave. — Daniel Safai

missuse missuse · Accepted Answer · 2017-11-02T09:05:32

Here is an approach:

g <- as.Node(dataset)
g <- as.igraph(g, directed=TRUE)

Check how the nodes are sorted

g_data <-  get.data.frame(g,  what="vertices")$name

see how the nodes are sorted in your data:

data_vert <- apply(dataset[1:6], 1, function(x) tail(x[x!=""], 1))

Sort data according to nodes and apply color

 V(g)$color <- ifelse(dataset$color_thresh[match(g_data , data_vert)] == 1, "red", "green")


plot.igraph(g, vertex.label.color="black", vertex.label.dist=0, vertex.label.family="Helvetica", vertex.frame.color="white", vertex.size=25, layout=layout_as_tree)

Is that the correct color coding?

another approach would be explicitly stating which nodes should have which color:

g <- as.Node(dataset)
g <- as.igraph(g, directed = TRUE)

V(g)$color <- "green" 
V(g)["TAC_310B_124"]$color <- "red"
V(g)["SC_B949_3156" ]$color <- "red"
V(g)["CORB_AE16_10921"]$color <- "red"

same output image

R igraph vertex coloring

1 Answers