weight edge widths in emapplot

[cstubben]

Is it possible to weight the edge widths in emapplot so the term similarity is clearly visible? For example, the edges in the default plot below all look the same, but I would like the line widths to range from 0.2 to 1 (times the cex_line). So if the term similarity is 0.25 , it should be 1/4 the width of an edge connecting nodes with a term similarity=1

data(geneList)
de <- names(geneList)[abs(geneList) > 2]
edo <- enrichKEGG(de)
edo <- pairwise_termsim(edo)
emapplot(edo, cex_line=3)

[cstubben]

[huerqiang]

The width of edges in emapplot are not the same, they are proportional to the similarity of terms.

[cstubben]

I think the widths should range from the min_edge or 0.2 to 1, but it's very difficult to tell which pairs of terms in the plot have 100% similarity compared to those with 80% or 40% or even 20% similarity. I thought 20% similarity should be 1/5 the size of lines with 100% similarity? Also, is there a way to make the term similarity more obvious, maybe a color scale for edges (light gray to black) or display the term similarity in the middle of the lines. I will also check ggraph and see if I can scale these lines any better.

[cstubben]

This answer on StackOverflow might help and maybe all that's needed is the option to adjust the scale_edge_width?

library(ggraph)
library(igraph)
data=data.frame(w1=c(rep('apple', 4), 'pear', 'pear', 'peach'), 
           w2 = c('orange', 'pear','peach', 'banana', 'peach', 'banana', 'orange'),
           weight= c(.2,.4,.5,.8, 1, .75, .25))
data %>% 
  graph_from_data_frame() %>% 
  ggraph(layout = "fr") +
  geom_edge_link(alpha = .25, 
                 aes(width = weight)) +
  geom_node_point(color = "blue", size = 2) + 
  geom_node_text(aes(label = name),  repel = TRUE)+
  scale_edge_width(range = c(1,10)) +
  theme_graph()

[cstubben]

[cstubben]

This looks much better but the legend does not make sense (4 to 6?).
emapplot(edo, cex_line=3) + scale_edge_width(range = c(1,10))