Hi, I am studying RNAseq data obtained from human intestinal organoids treated with parasites derived material, so i have three biological replicates per condition (3 controls and 3 treated). I have performed reads count and normalization, and after DeSeq2 run with default parameters (padj<0.1 and FC>1), among over 16K transcripts included in the list of usable material, only 7 genes appear to be differentially expressed in term of padj value <0.05 (if i look at the pvalue not adjusted after correction for multiple testing, significant transcripts are arounf 600).

I can understand that maybe the treatment with parasites derived material may have not caused any crucial large significant changes in expression of genes, however i was wondering if i am missing something just because of the default parameters i have used as a threshold). Could you give please some advise on how to modify such parameters? I can imagine that if i change FC in a more stringent way, even less transcripts may appear to be significantly differentially expressed, however an opinion from experts may of course be of great help. Thanks in advance, SC

You don't say how you are setting the fold change criterion. Are you using the lfcThreshold and altHypothesis arguments to results, or simply filtering on the fold change 'by hand'? If the former, do note that you are using a pretty strong criterion for the fold change (you are requiring evidence that the gene more than doubles in expression, which means that the observed fold change will have to be much larger than that, given the degrees of freedom). If you are just filtering the results to those with a logFC > 1, you shouldn't do that, as it completely invalidates both your FDR and the p-values.

The main use-case for filtering on fold change is when you have so many differentially expressed genes that you are overwhelmed by the results. In your situation there isn't a compelling reason to do so.