I have been analysing some protoarray protein arrays in R and following quantile normalisation of log2 transformed data it looks like this:
Later down the line I wanted to compare the overall reactivates from the "ARF" individuals (as seen in image) compared to all the rest which are "controls"
So I un transformed the data using 2^(value) and then took the mean and standard error of "ARF" vs "controls" and got the following:
My question is, does this make sense as an analysis? Could I say that "ARF" have higher reactivates than "controls",
If so how does the quantile normalisation not mean the two groups are identical?