I am investigating the RUVseq right now and have noticed that they are using glm model of the edgeR to estimate the common and tag wise dispersion. From edgeR manual I read that one would use GLM model for general experiments with multiple factors. However, in RUVseq tutorial in the example they have only one factor control and treated. Why do they still use GLM model and not normal estimateCommonDisp() and estimateTagwiseDisp()?
design <- model.matrix( ̃x, data=pData(set))
y <- DGEList(counts=counts(set), group=x)
y <- calcNormFactors(y, method="upperquartile")
y <- estimateGLMCommonDisp(y, design)
y <- estimateGLMTagwiseDisp(y, design)