I use the DESeq2 package for many projects in which the number of samples is quite small. But, sometimes i have some projects with almost 700 samples and i want to do the same kind of analysis for all the projects.
When the number of samples is huge the estimation of the dispersion is very time consuming. In this specific case, i think that using the estimation providing by the parametric or local regression is not required, the gene-wise dispersion can be used (or something close to this estimation).
So my question is, is there a way to by pass the estimation and directly use the gene-wise dispersion ?