rnb.execute.imputation: rnb.execute.imputation

In RnBeads: RnBeads

Description

Removes missing methylation values in the methylation matrix of the given object

Usage

rnb.execute.imputation(
  rnb.set,
  method = rnb.getOption("imputation.method"),
  update.ff = TRUE,
  ...
)

Arguments

rnb.set	Dataset object inheriting from RnBSet.
method	Imputation method to be used, must be one of "mean.cpgs", "mean.samples", "random", "knn", "median.cpgs", "median.samples", or "none".
update.ff	flag indicating if the disk based matrices should be updated. Should be set to FALSE, if methylation matrix should only temporarly be changed. If this value is FALSE, the region level methylation values are not updated and only the site-wise matrix is changed temporarly.
...	Optional arguments passed to knn.imputation

Details

Imputes missing values by applying on the following methods:

mean.cpgs:

missing values are inferred as the average methylation value from all other (non-mising) CpGs in this sample

mean.samples:

missing values are inferred as the average methylation value from all other (non-mising) values at this CpG sites in all other samples

random:

missing values are inferred by randomly selecting a (non-missing) methylation value from any other sample at this CpG site

knn:

missing values are inferred by k-nearest neighbors imputation (see impute)

median.cpgs:

missing values are inferred as the median methylation value from all other (non-mising) CpGs in this sample

median.samples:

missing values are inferred as the median methylation value from all other (non-mising) values at this CpG sites in all other samples

none:

imputation should not be performed

Value

The modified rnb.set object without missing methylation values.

Author(s)

Michael Scherer

RnBeads documentation built on March 3, 2021, 2 a.m.