Only run one normalization method when override_normalization is set

[bnovotny]

If override_normalization is set (not blank), it would be more efficient to turn off the processes for the other three normalization methods. We could add this logic in normalizations.nf and change characterize_normalizations.py to skip the comparison report if the overrideVar is found.

Might be worth adding a "normalization only" flag to the pipeline in case someone wants to compare the normalization methods before running classification with a specific one.

[raymond301]

This is eventually intended to be automatically selected when higher sophistication. This method is not fully designed out. For now override is default, can run multiple times using '-resume' without the need to redo normalization computation.