How To Make Heatmaps in R with ComplexHeatmap?

Heatmap is a Visualization technique in the form of a matrix. It is mainly used for analyzing the numeric feature in a dataset and to visualize the trend over the data. This technique can highlight the range of values using color intensity i.e it indicates higher values with brighter colors and the color gradually fades out for smaller values.

In this article, we will see how to plot heatmaps using the Heatmap( ) function of ComplexHeatmap Package in R programming language.

Installation

To create heatmaps using ComplexHeatmap it has to be first installed and loaded into the working space.

Syntax:

# installing the package  
install.packages("ComplexHeatmap")
library(ComplexHeatmap)

Heatmap 

To create a heatmap using this module a simple heatmap() function is used with appropriate parameters.

Syntax:

Heatmap(m, col = c(), cluster_columns = FALSE, cluster_rows = FALSE, rect_gp = gpar() … )

Parameters:

• m :  matrix or vector
• cluster_columns : specify clusters for columns
• cluster_rows : specify clusters for rows
• rect_gp : to add gap spaces to heatmap

Example: Simple Heatmap 

R

library(ComplexHeatmap)

# code to create a simple heatmap
set.seed(2020)

# generate random numbers
d1 <- runif(500, 0, 1)
m<-matrix(d1,nrow = 20,ncol=10)

# randomize the data points
m<-m[,sample(ncol(m))]

Heatmap(m)

Output:

Customizing colors of a Heatmap

To change the color of the heatmap col parameter is used with the values to change the color with.

Example: Changing color of a heatmap 

R

library(ComplexHeatmap)

# code to create a simple heatmap
set.seed(2020)

# generate random numbers
d1 <- runif(500, 0, 1)
m<-matrix(d1,nrow = 20,ncol=10)

# randomize the data points
m<-m[,sample(ncol(m))]

# heatmap with custom colors
Heatmap(m,col = c('#b3ffff','#006600','#ffffb3'))

Output:

Heatmap with Column Clusters

We assign cluster_rows as FALSE to remove row clusters.

Example: Removing row clusters

R

library(ComplexHeatmap)

# code to create a simple heatmap
set.seed(2020)

# generate random numbers
d1 <- runif(500, 0, 1)
m<-matrix(d1,nrow = 20,ncol=10)

# randomize the data points
m<-m[,sample(ncol(m))]

# heatmap with col clusters
Heatmap(m,cluster_rows = FALSE,col = c('#1a53ff','#ffcccc'))

Output:

Heatmap with Row Clusters

We assign cluster_columns as FALSE to remove column clusters.

Example: Heatmap with row clusters

R

library(ComplexHeatmap)

# code to create a simple heatmap
set.seed(2020)

# generate random numbers
d1 <- runif(500, 0, 1)
m<-matrix(d1,nrow = 20,ncol=10)

# randomize the data points
m<-m[,sample(ncol(m))]

# heatmap with row clusters
Heatmap(m,cluster_columns = FALSE,
        rect_gp = gpar(col = "white", lwd = 2),
        col = rev(rainbow(10)))

Output:

Heatmap with no clusters

Let’s visualize a heatmap with no clusters. For this cluster_rows and cluster_columns are set to FALSE.

Example: Heatmap with no clusters

R

library(ComplexHeatmap)

# code to create a simple heatmap
set.seed(2020)

# generate random numbers
d1 <- runif(500, 0, 1)
m<-matrix(d1,nrow = 20,ncol=10)

# randomize the data points
m<-m[,sample(ncol(m))]

# heatmap with no clusters
Heatmap(m,cluster_rows = FALSE,cluster_columns = FALSE)

Output:

Combine two Heatmap together

Let’s add 2 heatmaps and visualize them together. For this plots are created independently and then joined together.

Example: Combining two heatmap together

R

library(ComplexHeatmap)

# code to create a simple heatmap
set.seed(2020)

# generate random numbers
d1 <- runif(500, 0, 1)
m<-matrix(d1,nrow = 20,ncol=10)

# randomize the data points
m<-m[,sample(ncol(m))]

# heatmap 1
h1<-Heatmap(m,col = c('#b3ffff','#006600','#ffffb3'))

# heatmap 2
h2<-Heatmap(m,cluster_rows = FALSE,col = c('#1a53ff','#ffcccc'))

# add to combine 2 heatmaps
h1+h2

Output: