Hi all,

I have an rna seq dataset for which I used Deseq2 to find differentially expressed genes between two groups. I got many genes (in the thousands) with adjp <0.1, but the base mean varies from 0.5 to 5000 depending on the gene.

Is there a common threshold people use to filter based on the base mean counts?

Thanks 

DESeq2 uses a data adaptive approach to find a threshold which maximizes the number of detected DE genes. You can read about this in the vignette, under "independent filtering".

If you want to reduce the number of gene candidates, I would suggest to use lfcThrehsold, rather than removing genes with low-ish counts, which still demonstrate statistical evidence of differential expression.