Generate a node/edge matrix from long data or single columns of variables (or other shared analyses)-CodePudding

I am trying to find the number of shared and unique features between several clusters (nodes) and visualize them. The data I have is in 2 columns. I think I need to get it into target/source format but I cannot figure it out from this current format.

Here is some example data:

df <- data.frame(cluster = c(rep(1, 5), rep(2, 5), rep(3, 5)),
                 feature = c(letters[1:3], letters[7:8], letters[1:3], letters[9:10], letters[2:3], letters[9:11]))

> df
   cluster feature
1        1       a
2        1       b
3        1       c
4        1       g
5        1       h
6        2       a
7        2       b
8        2       c
9        2       i
10       2       j
11       3       b
12       3       c
13       3       i
14       3       j
15       3       k

I want to show that cluster 1 shares a with 1 other cluster, cluster 1 shares b with 2 other clusters, cluster 2 shares i with cluster 3 etc.

I have tried so many combinations of tidyr, plyr, dplyr code but I can't figure it out. For example this basic code gives me the number of shared partners between clusters, but not which partner it is shared with.

df2 <- df %>%
  group_by(feature) %>%
  mutate(n_gene = n())


> df2
# A tibble: 15 × 3
# Groups:   feature [8]
   cluster feature n_gene
     <dbl> <chr>    <int>
 1       1 a            2
 2       1 b            3
 3       1 c            3
 4       1 g            1
 5       1 h            1
 6       2 a            2
 7       2 b            3
 8       2 c            3
 9       2 i            2
10       2 j            2
11       3 b            3
12       3 c            3
13       3 i            2
14       3 j            2
15       3 k            1

My goal is to have either a network like something below (credit to

I tried plotting your graph using iagraph as well. What it does is, create a node between a cluster and a feature that is wrong but still, I'll include it to show you what's wrong.

library(igraph)
df.g <- graph.data.frame(d = df, directed = TRUE)
df.g
plot(df.g, vertex.label = V(df.g)$name)

It looks like this:

CodePudding user response：

Maybe something like this?

df %>%
  graph_from_data_frame() %>%
  set_vertex_attr(
    name = "color",
    value = c("green", "red")[1   (names(V(.)) %in% df$feature)]
  ) %>%
  plot()