Normalization on protein ID duplicates

Question

[jumagarilla]

Hi everyone!

I have a list of gene IDs with their raw count values from a counting step in a RNA-Seq experiment. When converting the gene IDs to UniProt IDs, duplicate IDs appear at the protein level. If a normalization step is required to treat this data, would it be better to handle these duplicates by, for example, taking their mean or median value of their raw count values, before normalization or is it more correct to do so after normalization?

Thanks in advance! Regards,

Juan

EDIT:

I am also wondering if it is correct to group the duplicate values as below by their mean, median, min or max values.

A1L3X0  0
A1L429  0
A1L429  0
A1L429  0
A1L443  0