I have a question regarding pre-filtering of genes in RNA-Seq data for MaSigPro.
I have filtered out genes that the counts in all samples was low. In some data sets I did not get significant genes. I thought to further filter the data, to keep only genes had a difference above of 1.5 between any 2 samples:
keep only genes with: (max (counts in all samples)/min(counts in all samples) > 1.5
I then managed to get significant genes.
I wanted to ask whether it is a good workflow to have that per-filtering? And if so, would you suggest to do it always, or only in cases that no significant genes are obtained.