MultiAssayExperiment: Construct an integrative representation of multi-omic data...

In vjcitn/biocMultiAssay: Software for the integration of multi-omics experiments in Bioconductor

Construct an integrative representation of multi-omic data with MultiAssayExperiment

Description

The constructor function for the MultiAssayExperiment-class combines multiple data elements from the different hierarchies of data (study, experiments, and samples). It can create instances where neither a sampleMap or a colData set is provided. Please see the MultiAssayExperiment API documentation for more information.

Usage

MultiAssayExperiment(
  experiments = ExperimentList(),
  colData = S4Vectors::DataFrame(),
  sampleMap = S4Vectors::DataFrame(assay = factor(), primary = character(), colname =
    character()),
  metadata = list(),
  drops = list()
)

Arguments

experiments	A list or ExperimentList of all combined experiments
colData	A DataFrame or data.frame of characteristics for all biological units
sampleMap	A DataFrame or data.frame of assay names, sample identifiers, and colname samples
metadata	An optional argument of "ANY" class (usually list) for content describing the experiments
drops	A list of unmatched information (included after subsetting)

Value

A MultiAssayExperiment object that can store experiment and phenotype data

colData

The 'colData' input can be either 'DataFrame' or 'data.frame' with subsequent coercion to DataFrame. The rownames in the 'colData' must match the colnames in the experiments if no sampleMap is provided.

experiments

The 'experiments' input can be of class SimpleList or 'list'. This input becomes the ExperimentList. Each element of the input 'list' or 'List' must be named, rectangular with two dimensions, and have 'dimnames'.

sampleMap

The sampleMap can either be input as 'DataFrame' or 'data.frame' with eventual coercion to 'DataFrame'. The 'sampleMap' relates biological units and biological measurements within each assay. Each row in the 'sampleMap' is a single such link. The standard column names of the 'sampleMap' are "assay", "primary", and "colname". Note that the "assay" column is a factor corresponding to the names of each experiment in the 'ExperimentList'. In the case where these names do not match between the 'sampleMap' and the experiments, the documented experiments in the 'sampleMap' take precedence and experiments are dropped by the harmonization procedure. The constructor function will generate a 'sampleMap' in the case where it is not provided and this method may be a 'safer' alternative for creating the 'MultiAssayExperiment' (so long as the rownames are identical in the 'colData', if provided). An empty 'sampleMap' may produce empty experiments if the levels of the "assay" factor in the 'sampleMap' do not match the names in the 'ExperimentList'.

See Also

MultiAssayExperiment-class

Examples

## Run the example ExperimentList
example("ExperimentList")

## Create sample maps for each experiment
exprmap <- data.frame(
    primary = c("Jack", "Jill", "Barbara", "Bob"),
    colname = c("array1", "array2", "array3", "array4"),
    stringsAsFactors = FALSE)

methylmap <- data.frame(
    primary = c("Jack", "Jack", "Jill", "Barbara", "Bob"),
    colname = c("methyl1", "methyl2", "methyl3", "methyl4", "methyl5"),
    stringsAsFactors = FALSE)

rnamap <- data.frame(
    primary = c("Jack", "Jill", "Bob", "Barbara"),
    colname = c("samparray1", "samparray2", "samparray3", "samparray4"),
    stringsAsFactors = FALSE)

gistmap <- data.frame(
    primary = c("Jack", "Bob", "Jill"),
    colname = c("samp0", "samp1", "samp2"),
    stringsAsFactors = FALSE)

## Combine as a named list and convert to a DataFrame
maplist <- list(Affy = exprmap, Methyl450k = methylmap,
    RNASeqGene = rnamap, GISTIC = gistmap)

## Create a sampleMap
sampMap <- listToMap(maplist)
## Create an example phenotype data
colDat <- data.frame(sex = c("M", "F", "M", "F"), age = 38:41,
    row.names = c("Jack", "Jill", "Bob", "Barbara"))

## Create a MultiAssayExperiment instance
mae <- MultiAssayExperiment(experiments = ExpList, colData = colDat,
    sampleMap = sampMap)

vjcitn/biocMultiAssay documentation built on Feb. 7, 2024, 11:20 a.m.