Firstly, this is my first time posting to this website, so I hope that I don't miss anything important. This is more of a question for advice of what to try next from people experienced with this type of analysis. 

I have a moderately sized population cohort of RNA-Seq samples, ~400. As these have been processed over multiple batches and some have strong GC-bias, I chose to use salmon and tximport for processing in DESeq2. I have a number of covariates which I am interested in, as this is a control population with no sign of disease. However, from what I can understand, whilst DESeq2 will accept gender as a covariate, being a factor, it will not provide meaningful information for a continuous variable such as age. Is there a function in DESeq2 that would help to get around this, or would it be best to group ages and use a dummy variable for factoring? Or would it simply be best to use limma-voom for this purpose.

Many thanks,

Matt 

You can certainly provide a continuous factor, my point in the FAQ is just that by adding a continuous variable without any other specification, you are saying to expect equal FC per year of age. This is true of any gene expression software you use. Binning is an easy way for new users to allow for different effects over the range of a continuous variable. Another approach would be splines (this is actually super easy with design formula in R), but I don’t send my users down that path with code in hand because most DESeq2 users are doing statistical analysis for the first time and don’t know how to go about assessing the splines.