ReplaceOutliersByMAD: Replacing outliers by moderated MAD

In ABSSeq: ABSSeq: a new RNA-Seq analysis method based on modelling absolute expression differences

Description

Function for replacing the outliers by MAD

Usage

ReplaceOutliersByMAD(object, replaceOutlier = TRUE, cutoff = 2,
  baseMean = 100, limitMad = 0.707, spriors = 2, Caseon = TRUE, ...)

Arguments

object	a ABSSeq object with element of 'counts' and 'normMethod', see the constructor functions ABSDataSet.
replaceOutlier	switch for replacing, default is TRUE.
cutoff	cutoff of moderating MAD for outliers, default is 2
baseMean	parameter for limiting the trimming at low expression level by baseMean/(sample size), default is 100.
limitMad	the minimal prior for moderating MAD, default is set to 0.707, which is usually the highest standard deviation at expression level of 1
spriors	prior weight size for prior MAD, default is 2
Caseon	switch for dealing with outlier trimming at sample size of 2
...	reserved parameters

Details

Given a matrix of count data, this function replacing the outliers by MAD. Noticely, this function also provides part of parameters for DEs calling. It is called by callParameter

Value

a ABSDataSet object with normalized counts after trimming (replaceOutlier=TRUE) or not (replaceOutlier=FALSE). Use the excounts to show it. Use results with name 'trimmed' to view the trimming status.

Examples

data(simuN5)
obj <- ABSDataSet(counts=simuN5$counts, groups=factor(simuN5$groups))
obj <- normalFactors(obj)
obj <- ReplaceOutliersByMAD(obj)
head(excounts(obj))
head(results(obj,c("trimmed")))

ABSSeq documentation built on Nov. 8, 2020, 5:07 p.m.