nbinomGLMTest {DESeq}R Documentation

Perform chi-squared tests comparing two sets of GLM fits

Description

For each gene, the function calculates a chi-square p value by simply calculating: 1 - pchisq(resReduced$deviance - resFull$deviance, attr(resReduced, "df.residual") - attr(resFull, "df.residual"))

Usage

nbinomGLMTest(resFull, resReduced)

Arguments

resFull, resReduced

GLM fit data frames, as returned by fitNbinomGLMs, first the full, then the reduced model.

Value

a vector of p values

Author(s)

Simon Anders, anders@embl.de

See Also

fitNbinomGLMs

Examples

cds <- makeExampleCountDataSet()[ 1:100, ]
cds <- estimateSizeFactors( cds )
cds <- estimateDispersions( cds, method="pooled" )
fit1 <- fitNbinomGLMs( cds, count ~ condition )
fit0 <- fitNbinomGLMs( cds, count ~ 1 )
nbinomGLMTest( fit1, fit0 )

[Package DESeq version 1.6.1 Index]