ggplot facet_wrap multiple panels

Question

I work in Genetics and very recently started practicing R, ggplot in particular, to make images for publication.

I run several analyses and combined in a single table (Attempt_mod_TopTen) 3 dataset containing:

• y axis (GO_term) = names that have to be in common for the entire graph
• x1 axis (N_Genes_AB) = numbers of genes associated with y, and their p_value
• X2 axis (N_Genes_A) = numbers of genes associated with y, and their p_value
• X3 axis (N_Genes_B) = numbers of genes associated with y, and their p_value

I am able to make a figure with single panel using only one set of data (either X1, X2, or X3)

    library(ggplot2)
    theme_set(
      theme_bw() 
    )
    ggplot(data = Attempt_mod_TopTen)+geom_point(aes(x=N_Genes_AB, y=GO_Term, size=P_value_AB)) +
      scale_size(range = c(4,.2))   
           

Bubble plot displaying the relation between the GO_terms, the number of genes and their P value:

I saw that there are posts on how to make bubble plot with multiple panels (i.e. facet_wrap), but I guess I am doing something wrong with my data because I can't find how to incorporate data for X2 and X3. I would like to display what I am able to make in a single panel in 3 panels next to each other that share the same y axis.

Answer 1

As Allan Cameron suggested in the comments, the key here will be changing the way your data are structured via tidyr::pivot_longer(). This can be accomplished like so:

library(tidyverse)

dat <- Attempt_mod_TopTen %>%
    pivot_longer(
        !GO_Term,
        names_to = c(".value", "grp"),
        names_pattern = "([A-Z]_[A-Za-z]+)_([A-Z]+)"
    )

This is a straightforward application of the "Multiple observations per row" subsection of the "Longer" section in the "Pivoting" vignette of the tidyr package, which you can access via the R command vignette("pivot", package = "tidyr"); I would strongly suggest reading through it for an in-depth understanding of what was done there. However, we can get an idea by simply looking at the result. You can see we've turned every row into three rows, one for group "A", one for group "B", and one for group "AB". Then we don't need six columns for our observations of "N_Genes" and "P_value", but just two. Now we can easily use facet_wrap(), with the newly created grp column dictating the facets:

ggplot(data = dat) +
    facet_wrap(~grp) +
    geom_point(aes(x = N_Genes, y = GO_Term, size = P_value)) +
    scale_size(range = c(4,.2)) +
    theme(
        axis.text = element_blank(),
        axis.title = element_blank(),
        axis.ticks = element_blank()
    )