| Title: | Modular Leaf Ordering Methods for Dendrogram Nodes |
|---|---|
| Description: | An implementation of functions to optimize ordering of nodes in a dendrogram, without affecting the meaning of the dendrogram. A dendrogram can be sorted based on the average distance of subtrees, or based on the smallest distance value. These sorting methods improve readability and interpretability of tree structure, especially for tasks such as comparison of different distance measures or linkage types and identification of tight clusters and outliers. As a result, it also introduces more meaningful reordering for a coupled heatmap visualization. This method is described in "dendsort: modular leaf ordering methods for dendrogram representations in R", F1000Research 2014, 3: 177 <doi:10.12688/f1000research.4784.1>. |
| Authors: | Ryo Sakai [aut], Evan Biederstedt [cre, aut] |
| Maintainer: | Evan Biederstedt <[email protected]> |
| License: | GPL-2 | GPL-3 |
| Version: | 0.3.4 |
| Built: | 2024-11-07 02:56:16 UTC |
| Source: | https://github.com/evanbiederstedt/dendsort |
Modular Leaf Ordering Methods for Dendrogram Nodes
This package includes functions to optimize ordering of nodes in a dendrogram, without affecting the meaning of the dendrogram. A dendrogram can be sorted based on the average distance of subtrees, or based on the smallest distance value. These sorting methods improve readability and interpretability of tree structure, especially for tasks such as comparison of different distance measures or linkage types and identification of tight clusters and outliers. As a result, it also introduces more meaningful reordering for a coupled heatmap visualization.
Ryo Sakai [email protected]
dendsort sorts a dendrogram object which is
typically a result of hierarchical clustering (hclust). The
subtrees in the resulting dendrogram are sorted based on the
average distance of subtrees at every merging point. The
tighter cluster, in other words the cluster with smaller
average distance, is placed on the left side of branch.
When a leaf merge with a cluster, the leaf is placed on the
right side.
dendsort(d, isReverse = FALSE, type = "min")dendsort(d, isReverse = FALSE, type = "min")
d |
a dendrogram or hclust object. |
isReverse |
logical indicating if the order should be reversed.Defaults to FALSE |
type |
character indicating the type of sorting. Default to "min" |
output A sorted dendrogram or hclust.
#generate sample data set.seed(1234); par(mar=c(0,0,0,0)) x <- rnorm(10, mean=rep(1:5, each=2), sd=0.4) y <- rnorm(10, mean=rep(c(1,2), each=5), sd=0.4) dataFrame <- data.frame(x=x, y=y, row.names=c(1:10)) #calculate Euclidian distance distxy <- dist(dataFrame) #hierachical clustering "complete" linkage by default hc <- hclust(distxy) #sort dendrogram dd <- dendsort(as.dendrogram(hc)) hc_sorted <- as.hclust(dd) #sort in reverse, you can also pass hclust object plot(dendsort(hc, isReverse=TRUE)) #sort by average distance plot(dendsort(hc, type="average")) #plot the result par(mfrow = c(1, 3), mai=c(0.8,0.8,2,0.8)) plot(x, y, col="gray", pch=19, cex=2) text(x, y, labels=as.character(1:10), cex=0.9) plot(hc,main="before sorting", xlab="", sub="") plot(hc_sorted, main="after sorting", xlab="", sub="")#generate sample data set.seed(1234); par(mar=c(0,0,0,0)) x <- rnorm(10, mean=rep(1:5, each=2), sd=0.4) y <- rnorm(10, mean=rep(c(1,2), each=5), sd=0.4) dataFrame <- data.frame(x=x, y=y, row.names=c(1:10)) #calculate Euclidian distance distxy <- dist(dataFrame) #hierachical clustering "complete" linkage by default hc <- hclust(distxy) #sort dendrogram dd <- dendsort(as.dendrogram(hc)) hc_sorted <- as.hclust(dd) #sort in reverse, you can also pass hclust object plot(dendsort(hc, isReverse=TRUE)) #sort by average distance plot(dendsort(hc, type="average")) #plot the result par(mfrow = c(1, 3), mai=c(0.8,0.8,2,0.8)) plot(x, y, col="gray", pch=19, cex=2) text(x, y, labels=as.character(1:10), cex=0.9) plot(hc,main="before sorting", xlab="", sub="") plot(hc_sorted, main="after sorting", xlab="", sub="")
a multivariate table obtained from the integrated pathway analysis of gastric cancer from the Cancer Genome Atlas (TCGA) study. In this data set, each column represents a pathway consisting of a set of genes and each row represents a cohort of samples based on specific clinical or genetic features. For each pair of a pathway and a feature, a continuous value of between 1 and -1 is assigned to score positive or negative association, respectively.
data(sample_tcga)data(sample_tcga)
A data frame with 215 rows and 117 variables
We would like to thank Sheila Reynolds and Vesteinn Thorsson from the Institute for Systems Biology for sharing this sample data set.