Skip to contents

Generate Functional Spectrum

Source: R/FunctionalSpectra.R
getFunctionalSpectrum.Rd

This function generates functional spectrum for a single gene expression profile.

Usage

getFunctionalSpectrum(
  expressionProfile,
  geneSets = "MSigDB",
  refExp = NULL,
  logChange = FALSE,
  inverseRescale = FALSE,
  filter = -3
)

Arguments

expressionProfile

a named numeric vector containing gene expression profile

geneSets

gene sets to score. Either `"MSigDB"` for the bundled cache, `"MSigDB_<version>"` for an explicit bundled release, or a named list such as the result of [get_msigdbr()]

refExp

a character indicating cancer typer according to TCGA's indentifier, or a named vector reference expression

logChange

a logical flag indicating whether the input data is already in log change form, e.g., for two color microarray, you should turn it on. (default: FALSE)

inverseRescale

a logical flag indicating whether we rescale the reference to the scale of input data. If your single sample is microarray data and the reference is RNA-Seq, you should turn it on. (default: FALSE)

filter

a numeric indicating the cutoff value of expression. (default: -3)

Value

a numeric vector containing functional spectrum

Details

A single sample cannot be centred against a cohort, so a reference expression profile is required unless the input is already in log-change form. The reference is first restricted to genes expressed above `filter`, then the sample and reference are related by a linear fit and the score is computed on the difference between the fitted reference and the sample. The resulting spectrum is a comparison of the sample against that reference, not an enrichment of differentially expressed genes.

Set `inverseRescale = TRUE` when the sample is microarray and the reference is RNA-seq; this reverses the direction of the rescaling to account for the different expression scales. Set `logChange = TRUE` when the input is already a log-change vector, in which case no reference is used.

Scores are not comparable across different references, because the reference defines the baseline being compared against.

Note

You can generate the reference expression profile from your previous data or public data, which is the same(similiar) cancer type and platform. In DeepCC we also prepared average expression profiles of each cancer types in TCGA project as references. To use them, just use the TCGA identifier (COADREAD, BRCA, OV, etc.) to indicate the cancer type. If your single sample is microarray data, we strongly sugguest turn the parameter inverseRescale on, since TCGA is RNA-Seq, which has very small expression value for low expressed genes, compared with microarray.

See also

getFunctionalSpectra for a batch of gene expression profiles.

Examples

if (FALSE) { # \dontrun{
ep <- setNames(rnorm(100), paste0("G", seq_len(100)))
ref <- setNames(rnorm(100), paste0("G", seq_len(100)))
fs <- getFunctionalSpectrum(ep, geneSets=list(setA=c("G1","G5","G20")), refExp=ref)
} # }