The epitope tagging approach offers advantages of economy, universality, and precision over the use of antibodies raised directly against a protein of interest. The latter strategy promises a potentially greater diversity of reagents and obviates the need to modify the protein, but it may not yield sufficiently high-affinity, abundant, or specific antibodies. The major uncertainty in an epitope-tagging strategy, namely, the ability of the altered protein to function in vivo, is readily resolved in yeast by testing complementation of a null allele by the modified gene. Modification of the protein is easily accomplished by addition of the epitope coding sequence to the gene via oligonucleotide-mediated site-directed mutagenesis. The uniqueness of the epitope in the genome and the use of the monoclonal antibody assure a high-affinity, specific, and abundant antibody. Unrelated but identically modified proteins can be immunoprecipitated and affinity purified under the same conditions. Only extraction conditions and possibly a simple initial fractionation step need vary. Moreover, otherwise identical but differentially tagged proteins can be separated. Even proteins completely defective in an essential in vivo function can be purified and studied. Finally, polypeptides coprecipitating with the protein of interest are normally difficult to distinguish from those merely cross-reactive with the antibody used. As an alternative to defining a complex of proteins using a battery of antibodies, complexes are defined as a set of immunoprecipitable polypeptides present only in extracts containing the modified protein.