callParameterwithoutReplicates: Calculate parameters for differential expression test base on...

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

Description

Calculate parameters for each gene (the moderating basemean and dispersions), without replicates

Usage

callParameterwithoutReplicates(object)

Arguments

object

a ABSDataSet object.

Details

buliding a pseudo group to esitimate parameter by mean difference. shifted and calculate a set of parameters from normalized counts table before callDEs

Value

A ABSDataSet object with absolute differences, basemean, mean of each group, variance, log2 of foldchange, named as 'absD', 'baseMean', 'Amean', 'Bmean', 'Variance' and 'foldChange', respectively. Use the results to get access it

Note

This function should run after normalFactors or providing size factors. This function firstly constructs an expression level depended fold-change cutoffs and then separate the data into two groups. The group with fold-change less than cutoffs is used to training the dispersion. However, the cutoff might be too small when applied on data set without or with less DEs. To avoid it, we set a prior value (0.5) to it.

Examples

data(simuN5)
obj <- ABSDataSet(counts=(simuN5$counts)[,c(1,2)], groups=factor(c(1,2)))
obj <- normalFactors(obj)
obj <- callParameterwithoutReplicates(obj)
obj <- callDEs(obj)
head(results(obj))

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