addStrat: Add a BiclusterStrategy to a BiclusterExperiment

In jonalim/mfBiclust: Pipelines for Bi-Clustering Using Matrix Factorization

Description

Returns a BiclusterExperiment identical to bce with an additional BiclusterStrategy accessible using strategies() or getStrat().

Usage

addStrat(bce, k, method = c("als-nmf", "svd-pca", "snmf", "nipals-pca",
  "plaid", "spectral"), silent = FALSE, ...)

## S4 method for signature 'BiclusterExperiment,numeric'
addStrat(bce, k,
  method = c("als-nmf", "svd-pca", "snmf", "nipals-pca", "plaid",
  "spectral"), silent = FALSE, ...)

Arguments

bce	A BiclusterExperiment-class object to analyze
k	The number of biclusters to find
method	The biclustering algorithm
silent	Suppresses warnings and messages
...	Additional parameters to pass to the biclustering back-end

Details

One of the biclustering algorithms is used to compute a number of biclusters. Matrix factorization methods will store matrix factors in the factors slot of the BiclusterStrategy. The matrix factors will be thresholded to yield the binary matrices in the clusteredSamples and clusteredFeatures slots of the BiclusterStrategy.

Value

A copy of bce with a new BiclusterStrategy added to BiclusterExperiment@strategies

Methods (by class)

• bce = BiclusterExperiment,k = numeric: Default method

Additional arguments

addStrat() automatically manipulates certain back-end arguments as needed to return the desired number of biclusters (k). Back-end arguments besides those automatically manipulated can be provided by name to addStrat(). See bicluster-methods for a description of back-end arguments.

Potential side effects

Due to requirements of various biclustering methods, this function may override user parameters, with warning. Also, if any elements of abund(BiclusterExperiment) are missing, the row and column containing those elements may be removed, with warning.

Examples

bce <- BiclusterExperiment(yeast_benchmark[[1]])
# Set seed for reproducibility
set.seed(12345)
bce <- addStrat(bce, k = 2)

jonalim/mfBiclust documentation built on May 4, 2019, 4:13 a.m.