A protein complex in this context is meant as a stable set of interacting proteins which can be co-purified by an acceptable method, and where the complex has been shown to exist as an isolated, functional unit in vivo. Acceptable experimental methods include stringent protein purification followed by detection of protein interaction. The following methods should be considered non-acceptable: simple immunoprecipitation, pull-down experiments from cell extracts without further purification, colocalization and 2-hybrid screening. Interactions that should not be captured as protein complexes include: 1) enzyme/substrate, receptor/ligand or any similar transient interactions, unless these are a critical part of the complex assembly or are required e.g. for the receptor to be functional; 2) proteins associated in a pull-down/co-immunoprecipitation assay with no functional link or any evidence that this is a defined biological entity rather than a loose-affinity complex; 3) any complex where the only evidence is based on genetic interaction data; 4) partial complexes, where some subunits (e.g. transmembrane ones) cannot be expressed as recombinant proteins and are excluded from experiments (in this case, independent evidence is necessary to find out the composition of the full complex, if known). Interactions that may be captured as protein complexes include: 1) enzyme/substrate or receptor/ligand if the complex can only assemble and become functional in the presence of both classes of subunits; 2) complexes where one of the members has not been shown to be physically linked to the other(s), but is a homologue of, and has the same functionality as, a protein that has been experimentally demonstrated to form a complex with the other member(s); 3) complexes whose existence is accepted based on localization and pharmacological studies, but for which experimental evidence is not yet available for the complex as a whole.