如何为ggplot的geom_tile图层添加含元数据的列？

Great question! To add metadata columns to your geom_tile plot, you’ll need to integrate your metadata into your dataset, reshape the data for ggplot compatibility, and then customize the plot to distinguish metadata from your genetic markers. Here’s a complete, reproducible example:

Step 1: Combine Your Genetic Data with Metadata #

First, create your metadata variables (e.g., sample type, antibiotic resistance status) and merge them with your existing genetic data into a single data frame. I’ll use realistic metadata examples below:

# Your original genetic data
set.seed(123) # For consistent random samples
id <- 1:80
gyrA <- sample(c(1,0), 80, replace = TRUE)
parC <- sample(c(1,0), 80, replace = TRUE)
marR <- sample(c(1,0), 80, replace = TRUE)
qnrS <- sample(c(1,0), 80, replace = TRUE)
marA <- sample(c(1,0), 80, replace = TRUE)
ydhE <- sample(c(1,0), 80, replace = TRUE)
qnrA <- sample(c(1,0), 80, replace = TRUE)
qnrB <- sample(c(1,0), 80, replace = TRUE)
qnrD <- sample(c(1,0), 80, replace = TRUE)
mcbE <- sample(c(1,0), 80, replace = TRUE)

# Add example metadata
sample_type <- sample(c("Clinical", "Environmental"), 80, replace = TRUE)
abx_resistance <- sample(c("Resistant", "Susceptible"), 80, replace = TRUE)

# Combine into one data frame
df <- data.frame(id, sample_type, abx_resistance, gyrA, parC, marR, qnrS, marA, ydhE, qnrA, qnrB, qnrD, mcbE)

Step 2: Reshape Data to Long Format #

geom_tile works best with long-format data (each row represents one sample-variable pair). We’ll use pivot_longer from the tidyr package to reshape the data, and add a column to categorize variables as either "Metadata" or "Gene":

library(tidyr)
library(dplyr)

df_long <- df %>%
  pivot_longer(cols = -id, # Keep sample ID as a separate column
               names_to = "Variable",
               values_to = "Value") %>%
  # Tag variables as metadata or genetic markers
  mutate(Var_Category = ifelse(Variable %in% c("sample_type", "abx_resistance"), 
                               "Metadata", "Gene"))

Step 3: Create the Tile Plot with Metadata Columns #

Now we can build the plot. We’ll place samples on the y-axis, all variables (metadata + genes) on the x-axis, and style metadata differently to make it stand out:

library(ggplot2)

ggplot(df_long, aes(x = Variable, y = factor(id), fill = Value)) +
  # Add tile layer with white borders for clarity
  geom_tile(color = "white") +
  # Add a thick vertical line to separate metadata from genes
  geom_vline(xintercept = which(unique(df_long$Variable) == "abx_resistance") + 0.5, 
             color = "black", size = 1.2) +
  # Customize fill colors: distinct colors for metadata categories and gene presence/absence
  scale_fill_manual(values = c(
    "0" = "#ffffff", "1" = "#0072B2",
    "Clinical" = "#E69F00", "Environmental" = "#56B4E9",
    "Resistant" = "#D55E00", "Susceptible" = "#009E73"
  )) +
  # Clean up labels and theme
  labs(x = "", y = "Sample ID", fill = "Status") +
  theme_minimal() +
  theme(
    axis.text.x = element_text(angle = 45, hjust = 1, size = 10),
    axis.text.y = element_text(size = 8),
    panel.grid = element_blank(),
    legend.position = "bottom"
  )

Key Notes: #

• Customization: Adjust the metadata variables, fill colors, and theme to match your actual data and preferences.
• Categorical Metadata: If your metadata is numerical, you can use a continuous fill scale instead of manual values.
• Variable Order: Use factor(Variable, levels = c(...)) to control the order of columns in the plot (e.g., place metadata first).

内容的提问来源于stack exchange，提问作者Haakonkas