made4 (version 1.46.0)

prettyDend: Draw hierarchical tree of gene expression data with a colorbar for numerous class vectors

Description

Function which performs a hierarchical cluster analysis of data, drawing a dendrogram, with colorbars for different sample covariate beneath the dendrogram

Usage

prettyDend(dataset, labels = NULL, title = "", classvec = NULL,covars=1,returnTree=FALSE, getPalette=getcol,...)

Arguments

dataset

a matrix, data.frame, ExpressionSet or marrayRaw-class. If the input is gene expression data in a matrix or data.frame. The rows and columns are expected to contain the variables (genes) and cases (array samples) respectively.

labels

Vector, labels to be placed on samples in plots. Default is rownames(dataset).

title

Character, label to be placed on plots. Default is NULL.

classvec

A factor or vector or matrix or data.frame which describes the classes in columns of the dataset. Default is NULL.

covars

Numeric. The columns of the data.frame classve to be used as class vectors. These will be displayed as color bars under the dendrogram. The default is 1 (column 1).

returnTree

Logical. Return the hieracrhical cluster analysis results. Default is FALSE.

getPalette

Function, which generates a palette of colors. The default uses getcol function in made4. Other examples are provided below

...

further arguments passed to or from other methods.

Value

Details

The hierarchical plot is produced using average linkage cluster analysis with 1- Pearson's correlation metric. The default set of colors used to generate the color bars of the plots can be changed (see example below). By default, if there is only two levels in the factor, the colors will be black and grey.

References

See Also

See also as overview, hclust

Examples

  data(khan)
  logkhan<-log2(khan$train)

 
 # Get a character vector which defines which khan samples are cell lines or tissue sample
  khanAnnot= cbind(as.character(khan$train.classes),khan$cellType)
  print(khanAnnot[1:3,])

# Add 2 color bar, one for cancer subtype, another for cell type under dendrogram
  prettyDend(logkhan, classvec=khanAnnot, covars = c(1,2), labels=khan$train.classes)

# To change the palette of colors
# Use topo.colors(), see colors() for more help on inbuilt palettes

  prettyDend(logkhan, classvec=khanAnnot, covars = c(1,2), labels=khan$train.classes, getPalette=topo.colors)

# To use RColorBrewer Palettes
  library(RColorBrewer)

  # Use RColorBrewer Dark2 which contains 8 colors 
  prettyDend(logkhan, classvec=khanAnnot, covars = c(1,2), labels=khan$train.classes, getPalette=function(x) brewer.pal(8,"Dark2")[1:x])

  # Use RColorBrewer Set1 which contains 9 colors 
  prettyDend(logkhan, classvec=khanAnnot, covars = c(1,2), labels=khan$train.classes, getPalette=function(x) brewer.pal(9,"Set1")[1:x])