Entering edit mode
7.7 years ago
Sara
▴
260
I have 5 spikes (short sequence) and 4 samples (RNA-SEQ), one of them is control and 3 different treatment. I used different concentration for each spike but concentration of each spike in each sample is equal. now I counted the spikes in each sample. in two files the counts are almost the same but the other two samples have different spike counts. meaning in one of them they are 9 times less than control (comparing all 5 spikes between two samples) and the other one is 3 times less than control. now I am going to normalize the read counts using spikes? do you know how I can do that?
Trying to normalize with only 5 spike-ins sounds like a recipe for disaster. It there a reason that you actually need to normalize with the spike-ins? If so, which bioconductor package are you using for the analysis?
do you think one is enough? I am planning to use edgeR. what do you think? is that fine?
Without a good reason otherwise, I would aim for more like 100... Then again, one shouldn't use spike-ins without a good reason.