ISAModules-class: A set of ISA modules

In eisa: Expression data analysis via the Iterative Signature Algorithm

Description

An ISAModules object stores the results of one ISA run. It contains a set of biclusters (=modules or transcription modules) and some meta information about the ISA run and the input data.

Usage

## S4 method for signature 'ISAModules'
dim(x)
## S4 method for signature 'ISAModules'
featureNames(modules)
## S4 method for signature 'ISAModules'
sampleNames(modules)
## S4 method for signature 'ISAModules'
annotation(modules)
## S4 method for signature 'ISAModules'
getOrganism(modules)
## S4 method for signature 'ISAModules'
pData(modules)

## S4 method for signature 'ISAModules'
seedData(modules)
## S4 method for signature 'ISAModules'
runData(modules)
## S4 method for signature 'ISAModules'
featureThreshold(modules, mods)
## S4 method for signature 'ISAModules'
sampleThreshold(modules, mods)

## S4 method for signature 'ISAModules'
length(x)
## S4 method for signature 'ISAModules'
getNoFeatures(modules, mods)
## S4 method for signature 'ISAModules'
getNoSamples(modules, mods)

## S4 method for signature 'ISAModules'
getFeatures(modules, mods)
## S4 method for signature 'ISAModules'
getSamples(modules, mods)
## S4 method for signature 'ISAModules'
getFeatureNames(modules, mods)
## S4 method for signature 'ISAModules'
getSampleNames(modules, mods)
## S4 method for signature 'ISAModules'
getFeatureScores(modules, mods)
## S4 method for signature 'ISAModules'
getSampleScores(modules, mods)
## S4 method for signature 'ISAModules'
getFeatureMatrix(modules, binary = FALSE,
      sparse = FALSE, mods)
## S4 method for signature 'ISAModules'
getSampleMatrix(modules, binary = FALSE,
      sparse = FALSE, mods)
## S4 method for signature 'ISAModules'
getFullFeatureMatrix(modules, eset, mods)
## S4 method for signature 'ISAModules'
getFullSampleMatrix(modules, eset, mods)

## S4 method for signature 'ISAModules,ANY,ANY'
x[i, j, ..., drop = FALSE]
## S4 method for signature 'ISAModules,ANY,ANY'
x[[i, j, ..., drop = FALSE]]

Arguments

x,modules	An ISAModules object.
mods	An optional numeric index vector for the modules. If given, the information is only returned only for the specified modules.
binary	Logical scalar. Whether to binarize the feature or sample scores.
sparse	Logical scalar. Whether to return a sparse matrix. The Matrix package is required for sparse matrices.
eset	An ExpressionSet or ISAExpressionSet object. This is needed for calculating the scores of the features/samples that are not in the module. If an ExpressionSet object is supplied, then it is normalised by calling ISANormalize on it.
i	For ‘[’ an index vector for selecting features (=probes, genes). For ‘[[’ an index vector for selecting modules.
j	For ‘[’ an index vector for selecting samples. It is ignored for ‘[[’.
...	Additional indexing arguments, they are not used, just ignored.
drop	This argument is currently not used, just silently ignored.

Details

An ISAModules object contains a set of biclusters, obtained using one run of the Iterative Signature Algorithm.

Various operations are defined such an object, here we document all of them, in several groups.

Value

dim returns a numeric vector of length two. featureNames and sampleNames return a character vector. annotation and getOrganism return a character vector of length one. pData returns a data frame.

seedData returns a data frame, see more below. runData returns a named list, see more below. featureThreshold and sampleThreshold return a numeric vector.

length returns a numeric scalar. getNoFeatures and getNoSamples return a numeric vector.

getFeatures and getSamples return a list of named numeric vectors. getFeatureNames and getSampleNames return a list of character vectors. getFeatureScores and getSampleScores return a list of named numeric vectors. getFeatureMatrix, getSampleMatrix, getFullFeatureMatrix and getFullSampleMatrix return a numeric matrix.

Information about the input data.

dim returns the dimension of the input expression matrix, number of features times number of samples.

featureNames returns a character vector, the names of the features in the original input matrix. I.e. in the input was an ExpressionSet for an Affymetrix array, then the Affymetrix probe IDs are returned.

sampleNames returns a character vector, the names of the samples in the original expression set.

annotation returns a character scalar, the name of the array for the input expression set. More precisely, the annotation slot of the input ExpressionSet is returned, this is the name of the annotation package to use for the ExpressionSet.

getOrganism returns the scientific name of the organism for which the input expression data was measures. This is obtained by loading the annotation package of the input ExpressionSet object, so that must be installed.

pData returns the phenotypic data attached to the input ExpressionSet object, in a data frame, samples as rows and various phenotypic variables as columns.

Information about the ISA run

seedData returns information about the modules. Each row of the returned data frame corresponds to one module, the columns are various variables:

iterations

The number of ISA iterations needed to find the module.

oscillation

The length of the oscillation cycle for oscillating modules, zero for others.

thr.row

The feature (=gene) threshold used for finding the module.

thr.col

The sample (=condition) threshold used for finding the module.

freq

The number of times the module was found. This is always one, unless ISAUnique was performed.

rob

The robustness score of the module. See ISARobustness for details.

rob.limit

The robustness limit that was used for filtering the modules. As this depends of the feature and sample thresholds, it may be different for different modules.

runData returns information about the ISA runs, it is a named list with elements:

annotation

The annotation package corresponding to the input expression set.

organism

The scientific name of the organism.

direction

The direction parameter of the ISA. Please see ISAIterate for details.

convergence

The method to determine ISA convergence, a character scalar. Please see ISAIterate for details.

cor.limit

Correlation limit for the “cor” convergence criterium, see ISAIterate for details.

eps

Difference limit for the “eps” convergence criterium, see ISAIterate for details.

corx

Size of the time window for the “corx” convergence criterium, see ISAIterate for details.

maxiter

The maximum number of ISA iterations that was allowed.

oscillation

Logical, whether oscillating modules were considered during the ISA iteration.

N

Numeric scalar, the number of input seeds that were used to find the modules.

unique

Logical scalar, whether ISAUnique was run on the modules.

prenormalize

Logical scalar, whether the input data was prenormalized during ISA normalization, see ISANormalize.

hasNA

Logical scalar, whether the normalized input data contained some NA or NaN values.

rob.perms

Numeric scalar, the number of times the input data was scrambled when the modules were filtered according to robustness.

Note that some of these might be missing, i.e. rob.perms is only present if ISAFilterRobust was performed.

featureThreshold returns the feature thresholds that were used to find the modules.

sampleThreshold returns the sample thresholds that were used to find the modules.

Information about the modules

length returns the number of modules.

getNoFeatures returns the number of features (=genes) in the input data. The number of features after filtering is returned if the input data was filtered.

getNoSamples returns the number of samples (=conditions) in the input data.

Retrieve the modules

getFeatures returns the indices of the features included in the modules. It returns a list, with one entry for each module. Each entry contains the indices of the features (=genes) in the corresponding module.

getSamples does the same as getFeatures, but for samples.

getFeatureNames is similar to getFeatures, but returns feature names instead of feature indices.

getSampleNames is similar to getSamples, but returns sample names instead of sample indices.

getFeatureScores returns the feature scores for the selected modules (all modules by default). It returns a list, with one entry for each module. Each list entry contains the feature scores for one module, in a named numeric vector.

getSampleScores is similar to getFeatureScores, but for samples and sample scores.

getFeatureMatrix returns feature scores for the specified modules (all modules by default) in a matrix form. The number of rows is the number of features and the number of columns is the number of modules requested. It can optionally binarize the values.

getSampleMatrix is similar to getFeatureMatrix, but for sample scores.

getFullFeatureMatrix is similar to getFeatureMatrix, but is also calculates scores for the features that were not included in the module. For this it performs one ISA iteration and omits the thresholding step. You need to supply the normalized (or the original) expression data to make this possible.

getFullSampleMatrix is the same as getFullFeatureMatrix, but for sample scores.

Indexing

A couple of indexing operations were defined to make it easier selecting subsets of modules, features or samples from an ISAModules object.

The ‘[[’ double bracket indexing operator can be used with a single index vector to select a subset of modules.

The ‘[’ single bracket indexing operator can be used to restrict an ISAModules object to a subset of features and/or samples. The first index corresponds to features, the second to samples. Indices can be numeric, logical or character vectors, for the latter feature and sample names are used.

Author(s)

Gabor Csardi [e-mail]

References

Bergmann S, Ihmels J, Barkai N: Iterative signature algorithm for the analysis of large-scale gene expression data Phys Rev E Stat Nonlin Soft Matter Phys. 2003 Mar;67(3 Pt 1):031902. Epub 2003 Mar 11.

See Also

The vignette included in the eisa package.

Examples

data(ALLModulesSmall)
ALLModulesSmall

length(ALLModulesSmall)
dim(ALLModulesSmall)
annotation(ALLModulesSmall)
getOrganism(ALLModulesSmall)

seedData(ALLModulesSmall)

getNoFeatures(ALLModulesSmall)
getNoSamples(ALLModulesSmall)

getFeatureScores(ALLModulesSmall, 1)[[1]]

eisa documentation built on Nov. 8, 2020, 6:47 p.m.