如何使用image()函数绘制黑白双色PWM热力图？

Hey there! I assume you're working with R's image() function since that's the standard tool for this kind of plot. Let's walk through exactly how to create a black-and-white PWM (Position Weight Matrix) heatmap step by step, with code you can copy and adapt.

First, make sure your PWM is structured as a numerical matrix—typically rows represent nucleotides (A/C/G/T) and columns represent positions in the motif. Here's an example PWM to test with:

# Example 4x5 PWM matrix (rows = A/C/G/T, columns = positions 1-5)
pwm <- matrix(c(
  0.8, 0.1, 0.05, 0.05, 0.7,
  0.1, 0.8, 0.05, 0.05, 0.1,
  0.05, 0.05, 0.8, 0.1, 0.1,
  0.05, 0.05, 0.1, 0.8, 0.1
), nrow = 4, byrow = TRUE)
rownames(pwm) <- c("A", "C", "G", "T")
colnames(pwm) <- paste0("Pos", 1:5)

We need a palette that maps your PWM values to only two colors. Use colorRampPalette to generate a 2-color scale—swap the order if you want black for low values instead of high:

# White = low PWM values, Black = high PWM values
bw_palette <- colorRampPalette(c("white", "black"))(2)

# Reverse if you want the opposite: Black = low, White = high
# bw_palette <- colorRampPalette(c("black", "white"))(2)

By default, image() transposes the matrix, so we'll transpose our PWM first to get positions on the x-axis and nucleotides on the y-axis (the standard way to display PWMs). We'll also fix the y-axis labels to show our nucleotide names:

# Plot the transposed PWM
image(t(pwm), col = bw_palette,
      xlab = "Motif Position", 
      ylab = "Nucleotide",
      main = "Black-and-White PWM Heatmap",
      # Disable default y-axis to add custom labels
      yaxt = "n")

# Add custom y-axis labels for nucleotides
axis(2, at = seq(0, 1, length.out = nrow(pwm)), labels = rownames(pwm))

If you want to strictly split values into black/white based on a threshold (e.g., all values ≥0.5 are black, others white), binarize your PWM first:

# Set your threshold value
threshold <- 0.5

# Binarize the PWM: 1 = above threshold, 0 = below
pwm_binarized <- ifelse(pwm >= threshold, 1, 0)

# Plot the binarized heatmap
image(t(pwm_binarized), col = c("white", "black"),
      xlab = "Motif Position", 
      ylab = "Nucleotide",
      main = "Thresholded Black-and-White PWM Heatmap",
      yaxt = "n")
axis(2, at = seq(0, 1, length.out = nrow(pwm)), labels = rownames(pwm))

• If your PWM values aren't scaled between 0 and 1, normalize them first with pwm_normalized <- (pwm - min(pwm)) / (max(pwm) - min(pwm)) to ensure the color mapping works correctly.
• Adjust the main, xlab, and ylab arguments to match your dataset's context.

内容的提问来源于stack exchange，提问作者JJJ