The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Example with a TCGA dataset

8 September 2014

This is an R Markdown document, which demonstrates the use of gapmap and dendsort packages to generate a gapped cluster heatmap visualization.

Let’s start by loading the data file from the package, and creating two color palettes.

library(gapmap)
data("sample_tcga")
library(RColorBrewer)
## Warning: package 'RColorBrewer' was built under R version 4.1.2
RdBu = rev(brewer.pal(11, name="RdBu"))
RdYlBu = rev(brewer.pal(11, name="RdYlBu"))

Now you have the data matrix loaded, let’s calculate correlation-based distance and perform hierarchical clustering. In this example, we use the correlation-based dissimilarity (Pearson Correlation) and the complete linkage for hierarchical clustering.

#transpose
dataTable <- t(sample_tcga)
#calculate the correlation based distance
row_dist <- as.dist(1-cor(t(dataTable), method = "pearson"))
col_dist <- as.dist(1-cor(dataTable, method = "pearson"))
#hierarchical clustering
col_hc <- hclust(col_dist, method = "complete")
row_hc <- hclust(row_dist, method = "complete")
col_d <- as.dendrogram(col_hc)
row_d <- as.dendrogram(row_hc)

Now you are ready to plot the data. First, we will plot a cluster heatmap without any gaps.

gapmap(m = as.matrix(dataTable), d_row = rev(row_d), d_col = col_d, ratio = 0, verbose=FALSE, col=RdBu,
       label_size=2, v_ratio= c(0.1,0.8,0.1), h_ratio=c(0.1,0.8,0.1))

This gapmap package was designed to encode the similarity between two nodes by adjusting the position of each leaves. In the traditional representation, all the information about the similarity two adjacent nodes is in the height of the branch in a dendrogram. By positioning leaves based on the similarity, we introduce gaps in both dendrograms and heat map visualization. In the figure below, we exponentially map a distance (dissimilarity) of two nodes to a scale of gap size.

gapmap(m = as.matrix(dataTable), d_row = rev(row_d), d_col = col_d, mode = "quantitative", mapping="exponential", col=RdBu,
       ratio = 0.3, verbose=FALSE, scale = 0.5, label_size=2, v_ratio= c(0.1,0.8,0.1), h_ratio=c(0.1,0.8,0.1))

Since the background is white, we can use another color scale where the value 0 is encoded in yellow.

gapmap(m = as.matrix(dataTable), d_row = rev(row_d), d_col = col_d, mode = "quantitative", mapping="exponential", col=RdYlBu,
       ratio = 0.3, verbose=FALSE, scale = 0.5, label_size=2, v_ratio= c(0.1,0.8,0.1), h_ratio=c(0.1,0.8,0.1))

dendsort

This package works well with the dendsort package to reorder the structure of dendrograms. For further information for the dendsort, please see the paper.

library(dendsort)
gapmap(m = as.matrix(dataTable), d_row = rev(dendsort(row_d, type = "average")), d_col = dendsort(col_d, type = "average"),  
       mode = "quantitative", mapping="exponential", ratio = 0.3, verbose=FALSE, scale = 0.5, v_ratio= c(0.1,0.8,0.1), 
       h_ratio=c(0.1,0.8,0.1), label_size=2, show_legend=TRUE, col=RdBu)

dendrograms

You can also plot gapped dendrogram. First you need to create a gapdata class object by calling gap_data(). To bring the text labels closer to the dendrogram, we set a negative value to axis.tick.margin. This value should be adjusted depending on your plot size. If anyone has a better solution to adjust the position of the axis labels, please let me know.

row_data <- gap_data(d= row_d, mode = "quantitative", mapping="exponential", ratio=0.3, scale= 0.5)
dend <- gap_dendrogram(data = row_data, leaf_labels = TRUE, rotate_label = TRUE)
dend + theme(axis.ticks.length= grid::unit(0,"lines") )+ theme(axis.ticks.margin = grid::unit(-0.8, "lines"))
## Warning: The `axis.ticks.margin` argument of `theme()` is deprecated as of ggplot2
## 2.0.0.
## ℹ Please set `margin` property of `axis.text` instead
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

Here, we can also apply dendsort.

row_data <- gap_data(d= dendsort(row_d, type = "average"), mode = "quantitative", mapping="exponential", ratio=0.3, scale= 0.5)
dend <- gap_dendrogram(data = row_data, leaf_labels = TRUE, rotate_label = TRUE)
dend + theme(axis.ticks.length= grid::unit(0,"lines") )+ theme(axis.ticks.margin = grid::unit(-0.8, "lines"))

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.