run.rnb.ICA: run.rnb.ICA

View source: R/ICA.R

run.rnb.ICAR Documentation

run.rnb.ICA

Description

Performs ICA to remove the effect of a confounding factor from the methylation matrix

Usage

run.rnb.ICA(
  rnb.set,
  conf.factor,
  ica.setting = NULL,
  nmin = 10,
  nmax = 30,
  ntry = 1,
  thr.sd = 0,
  alpha.fact = 1e-10,
  type = "remove",
  ncores = 1,
  out.folder = NULL,
  save.report = F,
  alpha.feat = 0.01
)

Arguments

rnb.set

An object of type RnBSet-class containing methylation information

conf.factor

A vector of column names in the sample annotation sheet of rnb.set representing confounding factors to be removed.

ica.setting

Optional argument. If specified a named vector of arguments to be used. Can be one of the following.

nmin

Minimum number of components to be used

nmax

Maximum number of components to be used

ntry

Further argument for runICA

thr.sd

Threshold for the standard deviation across samples. Only sites with a standard deviation larger than this threshold are kept. 0 deactivates filtering.

alpha.fact

Significance level for the factor

type

Analysis type to be performed. Can be either "remove" to remove the effect of the confounding factor or "keep" to export the sites that are linked to the confounding factor. Note that for "keep", the factor effect is not removed, but only reported.

ncores

The number of cores used

out.folder

Folder to store ICA's output and diagnostic plots

save.report

Flag indicating if a report is to be created and stored for ICA

alpha.feat

Significance level detecting features contributing to components

Value

The modified rnb.set object with updated methylation values. The effect of the confouding factor is removed using independent component analysis (ICA).

Author(s)

Michael Scherer. ICA code was generated by Peter Nazarov and Tony Kamoa


CompEpigen/DecompPipeline documentation built on Nov. 3, 2023, 5:35 p.m.