Macromolecular interactions, such as those between proteins, form the basis of all biological activity. Here, we propose to map the interactions between 800 proteins involved in early embryonic development of the nematode Caenorhabditis elegans. These proteins are involved in core processes common to all multicellular organisms. To generate the network, a novel application of the yeast two-hybrid system will be used, that yields several improvements compared to current interaction datasets. First, it will lower the false negative rate, increasing the fraction of interactions identified. Second, it will identify the domains required for each interaction. Most protein interactions occur between specific protein domains. For an accurate representation of interaction networks, this information should be captured. It will enable us, for example, to know whether two interactions may take place simultaneously, or are mutually exclusive because they are mediated by the same domain. Finally, the domain interaction map will be integrated with existing phenotypic characterizations and expression profiles to be able to better predict the biological functions of the proteins in the network.The resulting combined network will be used to gain a better understanding of multicellular development. We are using C. elegans as a model system because it is a relatively simple, tractable multicellular organism, and powerful genetic tools are available to study components of interaction networks in vivo. We will focus on candidate novel components involved in asymmetric cell division and polarity establishment, and on proteins with human homologs that are implicated in disease. The interaction network will also be publicly available. Because of the conserved nature of the proteins involved, the interaction network and future results derived from it will be valuable not only to C. elegans researchers but to researchers in other fields as well.