What's the current standard practice in removing noise (in terms of probe values)? These noises will typically affect the interpretation of fold change.
We could just pick all the values above certain threshold, what would be the best way to determine this threshold?
Any R package for doing that?