Entering edit mode
I'm trying to wrap my head around the nondetects package.
Looking at the oncogene2013 data set, they say they used 2 cell lines, three treatments, and 4 replicates. That's 24 samples, which is the number of columns in exprs().
What's not discussed is the number of technical replicates, which I assume they must have done.
Can nondetects impute a dataset that includes technical replicates? Or do you just toss the replicates that don't work and use the geometric mean of anything remaining?
Thanks -Ed