makeCompleteDEUAnalysis {DEXSeq}R Documentation

Complete differential exon usage analysis

Description

This function performs a complete differential exon usage analysis, calling all the necessary functions and giving back an ExonCountSet object with p values and p adjusted values.

Usage

makeCompleteDEUAnalysis(ecs, 
        formulaDispersion=count ~ sample + condition*exon, 
        minCount=10, maxExon=50, formula0=NULL, formula1=NULL, 
        FDR=0.1, fitExpToVar="condition", nCores=1, path=NULL, 
        color=NULL, color.samples=NULL, quiet=FALSE, file="")
     

Arguments

ecs

An ExonCountSet object.

formulaDispersion

Formula used in the glm to calculate the dispersion values. The factors on the formula must be present in the design columns of the ExonCountSet object.

minCount

Minimum number of counts on an exon for it to be considered in the tests. This significantly increases the speed of the dispersion estimations and testing for differential exon usage. This is supported by the fact that small count exons are less likely of being called significant, so it should not affect the results.

maxExon

Genes with more exons than this value will be discarded from the analysis. This is a speed issue. Currently, time of dispersion estimations and testing for differential exon usage increases with number of exons.

formula0

Formula for the NULL model to fit a the glm. The factors must be present in the design columns of the ExonCountSet object. As it is tested for each of the exons, a factor exonID can be added to the formula, so that it will iterate over the exons of the gene fitting the glm for each of them. If it is left in NULL, the default formula is "count~sample+exon+condition" for the NULL model.

formula1

Same as formula0, but for the test model. If it is left in NULL, the default formula will be "count~sample+exon+condition*I(exon==exonID)". If added a factor "exonID", it will iterate over each of the exons of the geneID, e.g. If a geneID contains exons E01, E02, E03,...,EN, and the function is left in the default formula, the function will fit N glms, the last part of the formula will change in the iterations as follows: I(exon==E01), I(exon==E02), I(exon==E03),...,I(exon==EN).

FDR

A false discovery rate used to indicate the significant exons.

fitExpToVar

A variable contained in the design annotation of the ExonCountSet, the expression values will be fitted to this variable using the formula count~fitExpToVar*exon using a model frame obtained from the function modelFrameForGene.

nCores

Number of cores to be used to estimate the dispersions. multicore package must be loaded in order to split the job in several cores.

path

A path in the system where to write the report from DEXSeqHTML. If NULL, no report will be created.

color

A vector of colors for each of the levels from the factor in the design of the ExonCountSet object indicated by "fitExpToVar". If path is NULL, this parameter will be ignored.

color.samples

A vector of colors for each of the samples. If NULL, the colors of each sample will be asigned according to its corresponding level from "fitExpToVar". This option is useful to visualize complex experimental designs. If path is NULL, this parameter will be ignored.

quiet

If TRUE, no progress report is shown. In case the session is not an interactive session and progress report is wanted, include a file name in the parameter "file".

file

A file name to write the progress reports. If file="", output will be written in the standart output connection.

Value

An object of class ExonCountSet.

Examples

data("pasillaExons", package="pasilla")
formuladispersion <- count ~ sample + ( exon + type ) * condition
formula0 <- count ~ sample + type * exon + condition
formula1 <- count ~ sample + type * exon + condition * I(exon == exonID)
pasillaExons <- makeCompleteDEUAnalysis(pasillaExons, 
   formulaDispersion=formuladispersion, 
   formula0=formula0, 
   formula1=formula1)
   

[Package DEXSeq version 1.0.2 Index]