Affinity labelling is a popular method used for the study of macromolecules and their interactions with ligands. The method is based on the targeted delivery of a chemically cross-linkable group, attached to a reactive molecule with affinity for a particular site in the biopolymer of interest. In complex multicomponent systems, the applications of affinity labelling are restricted by the tendency of the reagents to randomly label nontargetted molecules. This review highlights techniques developed to minimize non-specific cross-linking and to achieve high selectivity for the labelling of target protein. Such techniques might be termed ?superselective labelling?, as opposed to traditional, less selective approaches.