Plotting a heatmap out of common value matrix in R

Question

I have a tab-separated file like the following:

        C1P C2P C3P C4P C5P
sam1    3 cp    3 cp    3 cp    3 cp    3 cp
sam2    S3c 4 cp    3 cp    3 cp    S3c
sam3    3 cp    3 cp    3 cp    3 cp    3 cp
sam4    3 cp    3 cp    LOH LOH 3 cp
sam5    3 cp    3 cp    3 cp    3 cp    3 cp
sam6    4 cp    4 cp    UPD UPD UPT

Now I want to assign a color to each value [condition]...and let the color appear wherever there is the same value, thereby making a heatmap that represents occurences of a specific condition [values in the column] in each sample [row headers].

Right now I am assigning numeric values to each condition, then generating the heatmap through pheatmap. But I was looking for more robust way of doing it.

Appreciate any help.

Answer 1

This should get you fairly close:

library(tidyverse)
df %>%
    rownames_to_column("row") %>%
    gather(col, Value, -row) %>%
    mutate(
        row = factor(row, levels = rev(unique(row))),
        Value = factor(Value)) %>%
    ggplot(aes(col, row, fill = Value)) +
    geom_tile() +
    scale_fill_manual(values = c(
        `3 cp` = "yellow",
        `4 cp` = "red",
        LOH = "blue",
        S3c = "lightgreen",
        UPD = "darkgreen",
        UPT = "black")) +
    labs(x = "", y = "") +
    theme_minimal()

Explanation:

1. Reshape data from wide to long.
2. Use geom_tile to draw a heatmap, where the fill colour is given by the factor levels of your values.
3. The rest is aesthetic "fluff" to increase the similarity to the image you link to.

---

Sample data

df <- read.table(text =
    "        C1P C2P C3P C4P C5P
sam1    '3 cp'    '3 cp'    '3 cp'    '3 cp'    '3 cp'
sam2    'S3c' '4 cp'    '3 cp'    '3 cp'    'S3c'
sam3    '3 cp'    '3 cp'    '3 cp'    '3 cp'    '3 cp'
sam4    '3 cp'    '3 cp'    LOH LOH '3 cp'
sam5    '3 cp'    '3 cp'    '3 cp'    '3 cp'    '3 cp'
sam6    '4 cp'    '4 cp'    UPD UPD UPT", header = T)