Incorporating information of control samples in sva, ComBat or models in DGE analysis (DESeq2/limma, etc)

Question

[marionette.kent]

Hi all,

I understand that it is always good to have across-batch control samples to look for batch effect and check if the correction method did its job. However, is there anyway to incorporate that information when doing batch effect correction such as with ComBat(), psva(), removeBatchEffect() (for visualisation), sva(), model in DESeq2 and limma/voom (for differential expression analysis)?

I read the vignette and only know sva library can use control genes, but not control samples, into the model. And I don't think I have seen many discussions on that in DESeq2/limma experiment design. It would be nice if someone can let me know if I can make better use of the information from these controls or not. Thank you.