Converting a 1D scatter plot to a histogram or density plot using the ggplot2 package in R (CC307)

October 17, 2024 • PD Schloss • 1 min read • •

Pat critiques a recent figure that he recreated and creates a new plot as a histogram. Along the way, he experiments with overlaying a density plot. He customizes the appearance using the annotate, labs, scale_x_continuous, scale_y_continuous, and scale_fill_manual functions. The plot is based on simulated data for Figure 1A from the paper, “Exploring novel microbial metabolites and drugs for inhibiting Clostridioides difficile” by Ahmed Abouelkhair and Mohamed Seleem (https://journals.asm.org/doi/10.1128/msphere.00273-24). The newsletter describing how I would go about generating the figure can be found at https://shop.riffomonas.org/posts/building-data-visualization-intuition.

Code

library(tidyverse)
set.seed(19760620)

sim_data <- tibble(
  compound = 1:527,
  percent_inhibition = sample(
    c(rnorm(n = 464, mean = 0, sd = 30),
      runif(63, 91, 100))
    )
  ) %>%
  mutate(strong = percent_inhibition >= 90)

sim_data %>%
  ggplot(aes(x = percent_inhibition, fill = strong)) +
#  geom_histogram(binwidth = 7) +
  geom_histogram(breaks = seq(-100, 100, 5),
                 show.legend = FALSE) +
  annotate(geom = "text",
                x = c(0, 45),
                y = c(5, 35),
                label = c("464 compounds\nwere below 90%",
                          "63 compounds\nwere above 90%"),
                fontface = "bold",
           color = c("black", "red"),
           hjust = c(0.5, 0)) +
  labs(x = "Inhibition of bacterial growth (%)",
       y = "Number of compounds") +
  scale_y_continuous(expand = c(0, 0)) +
  scale_x_continuous(limits = c(-100, 100)) +
  scale_fill_manual(breaks = c(FALSE, TRUE),
                   values = c("gray", "red")) +
  theme_classic()

ggsave("inhibition_histogram.png", width = 6, height = 4)

# sim_data %>%
#   ggplot(aes(x = percent_inhibition, fill = strong)) +
#   geom_density()
# 
# sim_data %>%
#   ggplot(aes(x = percent_inhibition, fill = strong, color = strong)) +
#   geom_histogram(binwidth = 7, aes(y = after_stat(density))) +
#   geom_density(alpha = 0.2)