callDEs: Testing the differential expression by counts difference

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

Description

Using NB distribution to calculate p-value for each gene as well as adjust p-value

Usage

callDEs(object, adjmethod = "BH", useaFold = FALSE)

Arguments

object	an ABSDataSet object.
adjmethod	the method for adjusting p-value, default is 'BH'. For details, see p.adjust.methods.
useaFold	switch for DE detection through fold-change, which will use a normal distribution (N(0,sd)) to test the significance of log2 fold-change. The sd is estimated through a quantile function of gamma distribution at callParameter.

Details

This function firstly calls p-value used pnbinom to call pvalue based on sum of counts difference between two groups or used pnorm to call pvalue via log2 fold-change, then adjusts the pvalues via p.adjust method. In addition, it also shrink the log2 fold-change towards a common dispersion after pvalue calling.

Value

an ABSDataSet object with additional elements: shrinked log2 fold-change, pvalue and adjusted p-value, denoted by foldChange pvalue and adj-pvalue, respectively. Use the results method to get access it.

Note

this function should run after callParameter

Examples

data(simuN5)
obj <- ABSDataSet(counts=simuN5$counts, groups=factor(simuN5$groups))
obj <- normalFactors(obj)
obj <- callParameter(obj)
obj <- callDEs(obj)
head(results(obj))

wtaoyang/ABSSeq documentation built on May 27, 2019, 8:46 a.m.