Functional diversity reflects the immense chemical diversity of living organisms that produce hundreds of thousands of small molecule compounds, most of which remain chemically and functionally characterized, hence the dark matter. Because small molecules rarely work on their own but rather via interactions with proteins, following the proverbial "tell me who your friends are, and I will tell you who you are," identification of protein interactors can be used to unravel the function of a metabolite. The complex and dynamic protein-metabolite interactions (PMI) network underlies all biological processes but remains under-characterized. In my group, we adapted co-fractionation mass-spectrometry (CF-MS), a well-established approach to map protein assemblies, for proteome and metabolome-wide identification of the protein-metabolite complexes. CF-MS experiments combine the separation of native complexes with MS analysis of the obtained fractions and use the similarity of elution profiles, referred to as co-elution or co-fractionation, to delineate interactors. CF-MS enables the untargeted identification of complexes without needing a protein or a metabolite bait. The PMI networks we generated in the group comprise tens of annotated metabolites and hundreds of unknown metabolic features across the four model organisms: E. coli, S. cerevisiae, C. elegans, and A. thaliana. During my talk, I will discuss how we use the PMI data to uncover novel and often unexpected regulatory functions of small molecule compounds.