Understanding protein–protein interactions within complex molecular networks can help us to understand many biological processes. A eukaryotic protein-protein interaction, or interactome, mapping effort has been initiated for Saccharomyces cerevisiae. However, many of the protein-protein interactions that are relevant for understanding human biology, disease and development only take place in multicellular organisms. Now, the interactome maps of two multicellular model organisms have been reported.

From a draft map of 7,048 proteins and 20,405 interactions, Jonathan Rothberg and colleagues have produced a high-confidence Drosophila melanogaster interactome map of 4,679 proteins and 4,780 interactions — to do this, they developed a computational method (a statistical model that incorporates experimental data) to assign confidence to the interactions. These authors found that the interactome map not only recreated known pathways, but that it also extended them and identified new pathway components. They also found that the protein–protein interactions were organized on both a local and a global level — this organization is thought to represent the formation of multiprotein complexes and intercomplex connections, respectively.

For Caenorhabditis elegans, Marc Vidal and colleagues identified 4,027 interactions, which they validated using a second, independent protein interaction assay. They then combined these interactions with previously identified interactions (found for specific processes, such as vulval development and germline formation) and those predicted by in silico searches for interactions that are known to exist for orthologues in other species. The result, Worm Interactome version 5 (WI5), contains 5,534 interactions and connects 15% of the C. elegans proteome. Interestingly, by showing that ancient, multicellular and worm-specific proteins interact with each other equally well, these authors also add weight to the theory that evolution creates new structures by modifying pre-existing ones.

The availability of these two different interactome maps is good news, because they not only provide “...functional hypotheses for thousands of uncharacterized proteins...” but are also “...a starting point for the systems biology modeling of multicellular organisms, including humans”. And, in an effort to make these interactomes public resources, Rothberg, Vidal and colleagues have deposited the interactions in various databases, including FlyBase, GRID (general repository of interaction datasets), BIND (biomolecular interaction network database) and DIP (database of interacting proteins).