I am new to RNA sequencing and unfortunately I don't know much about bioinformatics neither. I did send 6 samples to AGRF containing 2 experimental groups and a control (2 samples each group), which have been sequenced in one lane. Later we did another run with ten samples in one lane. There has been added a new experimental group (4 samples) and 2 new samples added two the 3 groups of the first sequencing (So from RNAseq1 plus RNAseq2: 3 experimental groups and a control group all containing 4 samples). I tested for DE with EdgeR, Voom and DEseq2 using Galaxy. I have been told that I shouldn't normalize my Data beforehand. If I put all the samples together I barely get any significant changes. But if I compare samples from the second run with the controls of the first run only, I get thousands of deferentially expressed genes. So there must be a difference between Run 1 and 2. So I am wondering now if it is just not a good idea to compare samples of different runs? Or can I do some kind of Normalization to make them more comparable?
Thaks in advance, Lena