Hi I have 4 samples belonging to 2 groups, so 2 replicates for each group. I am using edgeR to do differential analysis, and just find that many variables which hold not small logFC values are not significantly changed (FDR > 0.05). But I checked these variables, they actually are suitable to be defined as differential.

I think the problem is that the variance between samples are large due to I only have 2 replicates, so it is hard to pass the significance test in edgeR.

How I lower the sample variance properly? Which math and statistic methods should I use?

Thanks for your attention!