Identification, selection, validation and prioritization of targets for therapeutic intervention requires understanding of the biological role of individual proteins in cellular pathways. Unraveling the ways in which proteins interact with each other appears to be crucial in achieving that goal. A number of recently described high-throughput approaches for analyzing cellular protein-protein interactions and previously proposed prediction procedures are compared in this review. The relative advantages of each method are discussed in relation to reproducibility, comprehensiveness and biological significance. It is concluded that only a combination of complementary biochemical technologies supported by reliable algorithms, will provide exhaustive maps of protein interactions for a cellular interactome.