Mass spectrometry-based proteomic experiments generate massive amounts of complex data (multiple experimental modalities, instrument types and data analytic workflows). Large collections of such publicly available datasets from past experiments can be a valuable information asset for both informatics methods' developers and biomedical researchers. However, their effective utilization depends on the ability to quickly explore/query these complex data for insights. For these data, traditional relational databases present certain challenges. Therefore we built a scalable framework exploiting modern, open-source "Big Data" technologies, to harness these data in novel ways. In addition to the general NOSQL landscape, MongoDB, a document-oriented store, will be discussed in the context of applications in computational proteomics. Along with a powerful query engine that supports efficient querying over massive distributed datasets, MongoDB offers flexible and rich data representations, which are particularly suitable for many scientific domains.