OrderEV: Find the optimal arrangement of biclusters for visualization...
In ExpressionView: Visualize biclusters identified in gene expression data

Description Usage Arguments Details Value Author(s) See Also Examples

Finds the optimal arrangement of possibly overlapping biclusters that maximizes the areas of the largest contiguous parts of the biclusters. The reordering is necessary to obtain a visually appealing layout of the biclusters.

## S4 method for signature 'ISAModules'
OrderEV(biclusters, initialorder, maxtime, debuglevel)
## S4 method for signature 'Biclust'
OrderEV(biclusters, initialorder, maxtime, debuglevel)
## S4 method for signature 'list'
OrderEV(biclusters, initialorder, maxtime, debuglevel)

`biclusters`	An `ISAModules` object, a `Biclust` object or a named list. The last one is probably coming from the `isa2` package.
`initialorder`	A list containing the initial order. Usually the output of a previous ordering.
`maxtime`	The maximal computation time in seconds. The default value is one minute (`maxtime=60`).
`debuglevel`	The level of information provided during the ordering. By default, the debug output is turned off (`debuglevel=0`).

OrderEV performs a brute-force ordering of the biclusters, treating the rows and the columns of the matrix independently. The ordering algorithm is described in more detail in the accompanying vignette of the package.

A named list is returned with the following elements:

`rows / genes`	A list containing the maps between the rows of the initial and the optimally ordered gene expression matrix. The first element represents the map of the complete data set, while the subsequent entries contain the row maps of the data sets projected onto the individual clusters. This entry is called ‘`rows`’ if the function is called with a simple list as the first argument, and ‘`genes`’ otherwise.
`cols / samples`	A list containing the maps between the columns of the initial and the optimally ordered gene expression matrix. The first element represents the map of the complete data set, while the subsequent entries contain the column maps of the data sets projected onto the individual clusters. This entry is called ‘`cols`’ if the function is called with a simple list as the first argument, and ‘`samples`’ otherwise.
`status`	A list containing the status of the ordering. The list has two entries, named `genes` and `samples` (or `rows` and `cols` if the function was called with a simple list as the first argument). Each entry is a numeric vector of ones and zeros. A `1` indicates that the map is fully optimized, whereas a `0` signals that the ordering could not be completely within the given time frame.

Andreas Lüscher andreas.luescher@a3.epfl.ch

ExportEV, LaunchEV, ISA, biclust

## We generate some noisy in-silico data with biclusters,
## scramble the initially ordered arrangement
## identify the bicluster with the Iterative Signature Algorithm (ISA)
## and order the results with the OrderEV function
library(isa2)
data.in.silico <- isa.in.silico(noise=0.1)[[1]]
data.in.silico <- data.in.silico[sample(c(1:dim(data.in.silico)[1])),
                                 sample(c(1:dim(data.in.silico)[2]))] 
isa.results <- isa(data.in.silico)
optimalorder <- OrderEV(isa.results)
str(optimalorder)

## Create a plot for the scrambled and the optimal orderings
## Not run: 
layout(rbind(1:2))
image(data.in.silico)
image(data.in.silico[optimalorder$rows[[1]],
                     optimalorder$cols[[1]]])

## End(Not run)