PRIDE (PRoteomics IDEntifications database), as it's name suggests, is currently focused on identification data. It's main advantage are structured data and meta data, powerful query capabilities and, recently, nice visualisation functionality through PRIDE Inspector. It is also worth noting that they also allow private submissions that are accessible to reviewers and authors only prior to paper/data publication. Since NCBI's Peptidome has been discontinued, PRIDE is, AKAIK, the main generalist structured data repository.
Tranche is very different. It's main goal is the unstructured cloud storage or raw/binary data.
Another big player is PeptideAtlas, from ISB's Proteome Center. They reprocess data sets with their pipeline and then make this processed data available to the community.
Another significant player is the GPMDB. There are many other specialised data bases out. This review from 2009 by Mead, Bianco and Bessant might be a good read.
Finally, these data bases do not work in glorious isolation. They do interact quite a lot, as illustrated by the ProteomExchange consortium, that will provide a single point of submission to proteomics repositories.
Update: you did not specify whether you were interested in MS repositories only. Protein Atlas, for instance, provides protein expression profiles based on immunohistochemistry and sub-cellular localisation data.