Not all. Only when you are trying to define populations where you have low frequency and/or low expression. Then the total number in the FMO live cell gate should match that of the test sample.
Hi Yong, No the FMO cell number does not need to be the same as your actual sample. I would however keep the same number of cells between your FMO samples (i.e. split a full sample equally into the number of FMO samples you need). This should work just fine.