How to Create a Beautiful Interactive Heatmap in R

Prerequisites

Install the required R package:

install.packages("heatmaply")

Load the package:

library("heatmaply")

Data preparation

Normalize the data to make the variables values comparable.

df <- normalize(mtcars)

Basic heatmap

heatmaply(df)

Heatmaply has also the option to produce a static heatmap using the ggheatmap R function:

ggheatmap(df)

Replicating the dendrogram ordering of gplots::heatmap.2().

• Create a heatmap using the gplots R package:

gplots::heatmap.2(
  as.matrix(df),
  trace = "none",
  col = viridis(100),
  key = FALSE
)

• Create a similar version using the heatmaply R package:

heatmaply(
  as.matrix(df),
  seriate = "mean", 
  row_dend_left = TRUE,
  plot_method = "plotly"
)

Split rows and columns dendrograms into k groups

The k-means algorithm is used.

heatmaply(
  df,
  k_col = 2, 
  k_row = 2
)

Change color palettes

The default color palette is viridis. Other excellent color palettes are available in the packages cetcolor and RColorBrewer.

Use the viridis colors with the option “magma”:

heatmaply(
  df,
  colors = viridis(n = 256,  option = "magma"),
  k_col = 2, 
  k_row = 2
)

Use the RColorBrewer palette :

library(RColorBrewer)
heatmaply(
  df,
  colors = colorRampPalette(brewer.pal(3, "RdBu"))(256),
  k_col = 2, 
  k_row = 2
)

Specify customized gradient colors using the function scale_fill_gradient2() [ggplot2 package].

gradient_col <- ggplot2::scale_fill_gradient2(
   low = "blue", high = "red", 
    midpoint = 0.5, limits = c(0, 1)
  )
heatmaply(
  df,
  scale_fill_gradient_fun = gradient_col
)

Customize dendrograms using dendextend

A user can supply their own dendrograms for the rows/columns of the heatmaply using the Rowv and the Colv parameters:

library(dendextend)
# Create dendrogram for rows
mycols <- c("#2E9FDF", "#00AFBB", "#E7B800", "#FC4E07")
row_dend <-  df %>%
  dist() %>%
  hclust() %>%
  as.dendrogram() %>%
  set("branches_lwd", 1) %>% 
  set("branches_k_color", mycols[1:3], k = 3) 

# Create dendrogram for columns
col_dend <-  df %>%
  t() %>%
  dist() %>%
  hclust() %>%
  as.dendrogram() %>%
  set("branches_lwd", 1) %>% 
  set("branches_k_color", mycols[1:2], k = 2)

# Visualize the heatmap
heatmaply(
  df,
  Rowv = row_dend,
  Colv = col_dend
)

Add annotation based on additional factors

The following R code adds annotation on column and row sides:

heatmaply(
  df[, -c(8, 9)],
  col_side_colors = c(rep(0, 5), rep(1, 4)),
  row_side_colors = df[, 8:9]
)

Add text annotations

By default, the colour of the text on each cell is chosen to ensure legibility, with black text shown over light cells and white text shown over dark cells.

heatmaply(df, cellnote = mtcars)

Add custom hover text

mat <- df
mat[] <- paste("This cell is", rownames(mat))
mat[] <- lapply(colnames(mat), function(colname) {
    paste0(mat[, colname], ", ", colname)
})

heatmaply(
  df,
  custom_hovertext = mat
)

Saving your heatmaply into a file

Create an interactive html file:

dir.create("folder")
heatmaply(mtcars, file = "folder/heatmaply_plot.html")
browseURL("folder/heatmaply_plot.html")

Saving a static file (png/jpeg/pdf). Before the first time using this code you may need to first run: webshot::install_phantomjs() or to install plotly’s orca program.

dir.create("folder")
heatmaply(mtcars, file = "folder/heatmaply_plot.png")
browseURL("folder/heatmaply_plot.png")

Save the file, without plotting it in the console:

tmp <- heatmaply(mtcars, file = "folder/heatmaply_plot.png")
rm(tmp)

References

• Introduction to heatmaply