Unfortunately, charge state information is not available in the RAW data, so charge state would be a post-processing step. There is also no vendor-provided
peak picking, so any centroiding has to be post-processing as well. Waters's best solution for this is the not-free ProteinLynx Global Server (PLGS) which does
both charge state determination and centroiding. MassLynx itself can do some denoising and peak picking, but not charge state inference IIRC.
There may be some good post-processing tools for inferring charge state and centroiding out there, but I'm not aware of ones that that both do a good job AND
read mz[X]ML AND write mz[X]ML. Perhaps someone else can clue me in on that.
-Matt