isa2 (version 0.3.4)

isa.biclust: Convert ISA modules to a Biclust class, as defined by the biclust package

Description

This function converts the object with ISA modules to a Biclust object, so all the functions in the biclust package can be used on it.

Usage

isa.biclust(modules)

Arguments

modules

The ISA modules, as returned by the isa or some other function.

Value

A Biclust object.

Details

biclust is an R package that implements many biclustering algorithms in a unified framework. This function converts a set of ISA biclusters to a Biclust object, this class is used to store all biclustering results by the biclust package.

The Biclust class only supports binary biclusters, so the ISA modules are binarized during the conversion.

Examples

## You need the biclust package for this
if (require(biclust)) {
  set.seed(1)
  data <- isa.in.silico()
  modules <- isa(data[[1]])
  bc <- isa.biclust(modules)

  ## A heatmap
  drawHeatmap(data[[1]], bc, 1)

  ## A "bubble" plot
  bubbleplot(data[[1]], bc)

  ## Compare values inside and outside the bicluster
  plotclust(bc, data[[1]])

  ## Plot profiles of bicluster elements
  parallelCoordinates(data[[1]], bc, number=1)

  ## Coherence measures vs. ISA robustness
  cV <- sapply(seq(bc@Number), function(x)
     constantVariance(data[[1]], bc, x, dimension="both"))
  aV <- sapply(seq(bc@Number), function(x)
     additiveVariance(data[[1]], bc, x, dimension="both"))
  mV <- sapply(seq(bc@Number), function(x)
     multiplicativeVariance(data[[1]], bc, x, dimension="both"))
  sV <- sapply(seq(bc@Number), function(x)
     signVariance(data[[1]], bc, x, dimension="both"))

  rob <- robustness(isa.normalize(data[[1]]), modules$rows,
     modules$columns)

  cor( cbind(cV, aV, mV, sV, rob) )
}

Run the code above in your browser using DataCamp Workspace