Hi,

I'm using DESeq2 and I noticed that the estimateDispersions function is slow when I run it on many samples. I have checked the function input arguments and there is no available option for multi-threading. I was wondering if there is a way to add multi-threading to the source code.

I am looking for ways to make the function run faster. Can you think of any way other than setting quiet=TRUE?

It would be great if you could give me some suggestions.

Thanks in advance, Hossein

First of all, upgrade to the current release of DESeq2 as the team has speed up the fitting of dispersion for large samples. You can use the default which should be quite faster than older versions of DESeq2 or you can try glmGamPoi dispersion fitting within DESeq2. This is likely the best way to speed things up.

We do have built in parallelization at the DESeq() level, but usually it won't be better than the above due to overhead from transferring data across cores.