edgeR - Run GLM on only one gene

[nnn]

I would like to use edgeR's GLM functionality to determine the influence of several factors on a gene's expression levels. For performance reasons I would like to estimate the dispersion using the entire count table, but then determine the fit just for the gene I am interested in.

Is there any way to do this, i.e. call glmFit() on just one row of the table? I already tried to create a new CDS for just the gene and copy over the dispersion from the full CDS, but that does not seem to work.

[Gordon Smyth]

If 'y' is your full DGEList object, with common or tagwise dispersions already estimated, you can simply use:

fit <- glmFit(y[i,], design)

where 'i' is any subset of genes you want, even a single gene. The dispersions are automatically subsetted for you.

Note that edgeR is a Bioconductor package, so you can get answers quickly by posting them to the Bioconductor mailing list.

Gordon