Updated project metadata. To analyse protein-protein interactions on a global scale, we devised a methodology termed COLA, which uses subcellular fractionation combined with quantitative proteomics to generate multivariate subcellular localisation signatures for cellular proteins. Bootstrapped clustering was then used to match proteins with significant similarity in their localisation signatures. We utilised SILAC labeled human Retinal Pigment Epithelial-1 (RPE1) cells. Heavy and light labelled cells were fractionated using 4 parallel subcellular fractionation procedures, resulting in a total of 12 fractions. The majority of fractions come from two procedures, one using serial solubilisation (resulting in fractions for cytosol, total membrane, nuclear lumen,chromatin-bound nuclear, as well as two cytoskeletal fractions) and the other using centrifugation coupled with aqueous biphasic extraction (resulting in a total nuclear fraction, intracellular membranes fraction, plasma-membrane fraction, and a cytosol + microsomes fraction). The third fractionation procedure separated cellular protrusions using porous transwell membranes, and finally the fourth procedure separated the extracellular compartment by collecting conditioned media. Fractions from each SILAC label were then mixed with equal amounts of total cell lysate from the oposite label to generate fraction/lysate SILAC SILAC mixes. Two repicrocally labelled biological replicate experiments were performed. In addition to untreated RPE1 cells, we also analysed RPE1 cells pre-teated with the Myosin-II inhibitor Blebbistatin for 4 hours prior to fractionation to reveal changes in the protein localisation profiles upon Myosin-II inhibition, as Myosin-II is a key regulator of the cytoskeleton as well as intra-cellular trafficking (Note that the blebbistatin set is not included in the associated manuscript).