CellMix 1.6

Main Interface for Gene Expression Deconvolution Methods

Description

Main Interface for Gene Expression Deconvolution Methods

Usage

ged(object, x, method, ...)

S4 (MatrixData,ANY,GEDStrategy)
`ged`(object, x, method, data = NULL, maxIter = 1, ..., verbose = cellmix.getOption("verbose"), 
  wrap = TRUE)

S4 (ANY,ANY,function)
`ged`(object, x, method, ..., name = NULL)

Arguments

object
global gene expression matrix-like data object (e.g., matrix or ExpressionSet)
x
input data used by the algorithm to deconvolve global gene expression.
method
specification of a gene expression deconvolution method. If missing, a default suitable algorithm is selected, based on the type and/or dimensions of x and data.
...
extra arguments to allow extension, most of which are passed down to the deconvolution algorithm itself.
data
optional data, typically a marker list, specified in a format that is supported by the factory function MarkerList.
maxIter
maximum number of iterations to perform. If method is missing, the value of this argument can influence which method is selected. See section Details.
verbose
logical that toggles verbosity. A number (integer) can be passed to specify the verbosity level (the higher the more messages are output). Passing verbose=Inf toggles debug mode (all messages). Note that because it appears after ... it must be fully named.
wrap
logical that indicates the result returned by the method should be wrapped into an NMF object. If FALSE, it is returned as is. Note that because it appears after ... it must be fully named.
name
optional name for the custom algorithm.

Methods

  1. gedsignature(object = "MatrixData", x = "ANY", method = "GEDStrategy"): Default ged method apply the auto-selection scheme for determining which method is suitable for the type of input data.

  2. gedsignature(object = "ANY", x = "ANY", method = "character"): Applies a deconvolution algorithm registered in the CellMix registry.

  3. gedsignature(object = "ExpressionSet", x = "ANY", method = "GEDStrategy"): Defined to handle an issue in S4 method dispatch in union classes, and makes ged works with objects from other packages that inherit from ExpressionSet-class, e.g. LumiBatch objects defined in the lumi package.

    See R-devel thread: https://stat.ethz.ch/pipermail/r-devel/2013-May/066609.html.

  4. gedsignature(object = "ANY", x = "ANY", method = "missing"): This method deconvolves the target expression matrix using the algorithm selected by the automatic selection strategy implemented by selectGEDMethod, which choose a suitable algorithm whose data requirements match the provided input data (i.e. arguments x and optionally data). See some more details in the selectGEDMethod man page.

  5. gedsignature(object = "ANY", x = "ANY", method = "function"): Applies a custom gene expression algorithm.

Examples


# random global expression data: 3 cell types, 20 samples, 100 features
X <- rmix(3, 100, 20, markers=5)
dim(X)

##   Features    Samples Components 
##        100         20          3

# extract signature/proportion/markers
sig <- basis(X)
prop <- coef(X)
m <- getMarkers(X)
summary(m)

##      Length Class  Mode   
## CL_1 5      -none- numeric
## CL_2 5      -none- numeric
## CL_3 5      -none- numeric

#--------------------------------------------
# Automatic selection of a suitable algorithm
#--------------------------------------------
# expression data only: fastdeconf
res <- ged(X, 3)
# with markers only, non iterative: qprog
res <- ged(X, m)
# with markers only, iterative: ssKL
res <- ged(X, m, maxIter=5)
# with signatures: lsfit
res <- ged(X, sig)
# with proportions: csSAM
res <- ged(X, prop)
# with proportions, iterative: DSection
## Not run: 
##D if( require.quiet(RcppOctave) ){
##D  # requires octave package statistics to be installed (for gamfit)
##D     res <- ged(X, prop, maxIter=5)
##D }
## End(Not run)

See also

selectGEDMethod