ggupset

While UpSetR is a popular package for intersection visualization, ggupset is designed to integrate seamlessly with the ggplot2 ecosystem.

Key Differences:

• UpSetR: Provides a standalone plot with an additional view showing the overall size of each set. It requires data to be pivoted into a wide format.
• ggupset: Works directly with ggplot2 objects and can be combined with any ggplot that uses a categorical x-axis. This allows for non-standard plots, such as combining an UpSet-style axis with violin plots or boxplots.

UpSet plots in ggupset require data to be in a tidy format where categories are stored in a list-column. If you start with a wide matrix (e.g., genes vs. pathways), you must:

1. Convert the matrix to a tidy tibble using as_tibble() and gather() (or pivot_longer()).
2. Filter for TRUE memberships.
3. Group by the primary identifier (e.g., Gene) and use summarize(Pathways = list(Pathway)) to create the required list-column.

# 1. Tidy the matrix
tidy_pathway_member <- gene_pathway_membership %>%
  as_tibble(rownames = "Pathway") %>%
  gather(Gene, Member, -Pathway) %>%
  filter(Member) %>%
  select(- Member)

# 2. Create the list-column
tidy_pathway_member %>%
  group_by(Gene) %>%
  summarize(Pathways = list(Pathway)) %>%
  ggplot(aes(x = Pathways)) +
    geom_bar() +
    scale_x_upset()

You can install the released version of ggupset from CRAN or get the latest development version directly from GitHub using devtools.

# Download package from CRAN
install.packages("ggupset")

# Or get the latest version directly from GitHub
devtools::install_github("const-ae/ggupset")

When saving plots that include a combination matrix, do not rely on ggsave()'s default behavior of saving last_plot(), as it may only capture the combination matrix component. Instead, explicitly assign your complete ggplot object to a variable and pass that variable to the plot argument in ggsave().

pl <- tidy_movies %>% 
  ggplot(aes(x=Genres)) + 
  geom_bar() + 
  scale_x_upset(n_intersections = 20)

ggsave("/tmp/movie_genre_barchart.png", plot = pl)

To create an UpSet plot, use ggplot2 with a list-column in your data. Map the list-column to the x aesthetic and apply scale_x_upset() to transform the x-axis into a combination matrix. This allows you to visualize intersections of categories (e.g., genres) instead of a standard categorical axis.

library(ggplot2)
library(tidyverse)
library(ggupset)

# Assuming tidy_movies has a list-column 'Genres'
tidy_movies %>%
  distinct(title, year, length, .keep_all=TRUE) %>%
  ggplot(aes(x=Genres)) +
    geom_bar() +
    scale_x_upset(n_intersections = 20)

To customize the appearance of the combination matrix (the grid of dots/lines below the bar plot), use theme_combmatrix(). This function provides control over various elements of the matrix panel.

tidy_movies %>%
  distinct(title, year, length, .keep_all=TRUE) %>%
  ggplot(aes(x=Genres)) +
    geom_bar() +
    scale_x_upset(order_by = "degree") +
    theme_combmatrix(
      combmatrix.panel.point.color.fill = "green",
      combmatrix.panel.line.size = 0,
      combmatrix.label.make_space = FALSE
    )

The scale_x_upset function allows you to add an UpSet-style combination matrix axis to a ggplot object. It is particularly useful when your x-axis contains list-columns or combined categorical labels.

Common parameters:

• order_by: Determines the ordering of intersections (e.g., "degree" or "freq").
• n_sets: Limits the number of sets shown.
• n_intersections: Limits the number of intersections shown.
• sets: A character vector specifying the sets to include.
• position: The position of the axis (e.g., "top").
• name: The name of the axis.

ggplot(aes(x=Genres)) +
  geom_bar() +
  scale_x_upset(order_by = "degree", n_sets = 5)

If you prefer not to use the full UpSet matrix, you can use alternative scaling methods:

• scale_x_mergelist(sep = "-"): Automatically collapses list elements into a single delimited string (e.g., "Action-Animation") to be used as categorical axis labels.
• axis_combmatrix(sep = "-"): Replaces axis labels with a combination matrix, which can be used in conjunction with scale_x_mergelist() to improve readability.

# Using merged list labels with a combination matrix
tidy_movies %>%
  distinct(title, year, length, .keep_all=TRUE) %>%
  ggplot(aes(x=Genres)) +
    geom_bar() +
    scale_x_mergelist(sep = "-") +
    axis_combmatrix(sep = "-")

The scale_x_upset() function allows you to automatically order the categories and genres. Use the order_by argument set to "degree" to order them based on the number of intersections.

tidy_movies %>%
  distinct(title, year, length, .keep_all=TRUE) %>%
  ggplot(aes(x=Genres)) +
    geom_bar() +
    scale_x_upset(order_by = "degree")