summarizeMolecularProfiles-PharmacoSet-method: Takes molecular data from a PharmacoSet, and summarises them...

In bhklab/PharmacoGx: Analysis of Large-Scale Pharmacogenomic Data

Takes molecular data from a PharmacoSet, and summarises them into one entry per drug

Description

Given a PharmacoSet with molecular data, this function will summarize the data into one profile per cell line, using the chosen summary.stat. Note that this does not really make sense with perturbation type data, and will combine experiments and controls when doing the summary if run on a perturbation dataset.

Usage

## S4 method for signature 'PharmacoSet'
summarizeMolecularProfiles(
  object,
  mDataType,
  cell.lines,
  features,
  summary.stat = c("mean", "median", "first", "last", "and", "or"),
  fill.missing = TRUE,
  summarize = TRUE,
  verbose = TRUE,
  binarize.threshold = NA,
  binarize.direction = c("less", "greater"),
  removeTreated = TRUE
)

Arguments

object	PharmacoSet The PharmacoSet to summarize
mDataType	character which one of the molecular data types to use in the analysis, out of all the molecular data types available for the pset for example: rna, rnaseq, snp
cell.lines	character The cell lines to be summarized. If any cell.line has no data, missing values will be created
features	caracter A vector of the feature names to include in the summary
summary.stat	character which summary method to use if there are repeated cell.lines? Choices are "mean", "median", "first", or "last" In case molecular data type is mutation or fusion "and" and "or" choices are available
fill.missing	boolean should the missing cell lines not in the molecular data object be filled in with missing values?
summarize	A flag which when set to FALSE (defaults to TRUE) disables summarizing and returns the data unchanged as a ExpressionSet
verbose	boolean should messages be printed
binarize.threshold	numeric A value on which the molecular data is binarized. If NA, no binarization is done.
binarize.direction	character One of "less" or "greater", the direction of binarization on binarize.threshold, if it is not NA.
removeTreated	logical If treated/perturbation experiments are present, should they be removed? Defaults to yes.

Value

matrix An updated PharmacoSet with the molecular data summarized per cell line.

Examples

data(GDSCsmall)
GDSCsmall <- summarizeMolecularProfiles(GDSCsmall, mDataType = "rna", cell.lines=sampleNames(GDSCsmall), summary.stat = 'median', fill.missing = TRUE, verbose=TRUE)
GDSCsmall

bhklab/PharmacoGx documentation built on April 18, 2024, 3:13 a.m.