Class to contain data from multiple channel array technologies
Container for high-throughput assays and experimental meta-data. Data are from experiments where a single ‘chip’ contains several (more than 1) different ‘channels’. All channels on a chip have the same set of‘features’. An experiment consists of a collection of several N-channel chips; each chip is a ‘sample’.
An NChannelSet provides a way to coordinate assay data
(expression values) with phenotype information and references to chip
annotation data; it extends the eSet class.
An NChannelSet allows channels to be extracted (using the
channels method, mentioned below), and subsets of features or
samples to be selected (using [<features>,
<samples>]). Selection and subsetting occur so that relevant
phenotypic data is maintained by the selection or subset.
Objects can be created by calls of the form NChannelSet(
assayData, phenoData, ...). See the examples below.
assayData:Object of class AssayData, usually
an environment containing matrices of identical size. Each matrix
represents a single channel. Columns in each matrix correspond to
samples, rows to features. Once created, NChannelSet manages
coordination of samples and channels.
phenoData:Object of class AnnotatedDataFrame.
The data component of the AnnotatedDataFrame is
data.frame with number of rows equal to the number of
samples. Columns of the data component correspond to measured
covariates.
The varMetadata component consists of mandatory columns
labelDescription (providing a textual description of each column
label in the data component) and channel. The channel
of varMetadata is a factor, with levels equal to the names
of the assayData channels, plus the special symbol
_ALL_. The channel column is used to indicate which
channel(s) the corresponding column in the data component
of AnnotatedDataFrame correspond; the _ALL_ symbol
indicates that the data column is applicable to all
channels. varMetadata may contain additional columns with
arbitrary information.
Once created, NChannelSet coordinates selection and subsetting
of channels in phenoData.
featureData:Object of class AnnotatedDataFrame, used to contain
feature data that is unique to this experiment; feature-level
descriptions common to a particular chip are usually referenced
through the annotation slot.
experimentData:Object of class MIAME containing descriptions of the
experiment.
annotation:Object of class "character". Usually a length-1 character
string identifying the chip technology used during the
experiment. The annotation string is used to retrieve information
about features, e.g., using the annotation package.
protocolData:Object of class "character". A character vector
identifying the dates the samples were scanned during the
experiment.
.__classVersion__:Object of class Versions, containing
automatically created information about the class definition
Biobase package version, and other information about the user
system at the time the instance was created. See
classVersion and updateObject for
examples of use.
Class "eSet", directly.
Class "VersionedBiobase", by class "eSet", distance 2.
Class "Versioned", by class "eSet", distance 3.
Methods with class-specific functionality:
channel(object, name, ...)signature(object="NChannelSet", name="character").
Return an ExperessionSet created from the channel and
corresponding phenotype of argument name. name must
have length 1. Arguments ... are rarely used, but are
passed to the ExpressionSet constructor, for instance
to influence storage.mode.
channelNames(object), channelNames(object) <- value
signature(object = "NChannelSet").
Obtain, reorder, or rename channels contained in
object. See channelNames.
selectChannels(object, names, ...signature(object = "NChannelSet", names = "character").
Create a new NChannelSet from object, containing
only channels in names. The ... is not used by this
method.
object[features, samples]signature(object = "NChannelSet",
features = "ANY", samples = "ANY").
Create a new NChannelSet from object, containing
only elements matching features and samples; either
index may be missing, or a character, numeric, or logical vector.
sampleNames(object) <- valuesignature(object = "NChannelSet", value = "list") assign
each (named) element of value to the sampleNames of
the correspondingly named elements of assayData in
object.
Methods with functionality derived from eSet:
annotation, annotation<-, assayData,
assayData<-, classVersion, classVersion<-,
dim, dims, experimentData, experimentData<-,
featureData, featureData<-, phenoData,
phenoData<-, protocolData, protocolData<-,
pubMedIds, pubMedIds<-, sampleNames,
sampleNames<-, storageMode, storageMode<-,
varMetadata, varMetadata<-, isCurrent,
isVersioned, updateObject.
Additional methods: coerce (‘as’, to convert between
objects, if possible), initialize (used internally for
creating objects), show (invoked automatically when the
object is displayed to the screen)
Martin Morgan <mtmorgan@fhcrc.org>
## An empty NChannelSet
obj <- NChannelSet()
## An NChannelSet with two channels (R, G) and no phenotypic data
obj <- NChannelSet(R=matrix(0,10,5), G=matrix(0,10,5))
## An NChannelSet with two channels and channel-specific phenoData
R <- matrix(0, 10, 3, dimnames=list(NULL, LETTERS[1:3]))
G <- matrix(1, 10, 3, dimnames=list(NULL, LETTERS[1:3]))
assayData <- assayDataNew(R=R, G=G)
data <- data.frame(ChannelRData=numeric(ncol(R)),
ChannelGData=numeric(ncol(R)),
ChannelRAndG=numeric(ncol(R)))
varMetadata <- data.frame(labelDescription=c(
"R-specific phenoData",
"G-specific phenoData",
"Both channel phenoData"),
channel=factor(c("R", "G", "_ALL_")))
phenoData <- AnnotatedDataFrame(data=data, varMetadata=varMetadata)
obj <- NChannelSet(assayData=assayData, phenoData=phenoData)
obj
## G channel as NChannelSet
selectChannels(obj, "G")
## G channel as ExpressionSet
channel(obj, "G")
## Samples "A" and "C"
obj[,c("A", "C")]Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.