RiboWeb [13,14] is a resource whose primary aim is to facilitate the construction of three-dimensional models of ribosomal components and compare the results to existing studies. The knowledge that RiboWeb uses to perform these tasks is captured in four ontologies: The physical-thing ontology; the data ontology; the publication ontology and the methods ontology. The physical-thing ontology describes ribosomal components and associated `physical things'. It has three principle conceptualisations: Molecules, Molecule-Ensembles and Molecule-Parts. The first describes covalently bonded molecules and includes the main biological macromolecules. Molecule-ensembles captures non-covalently bonded collections of molecules, such as enzyme complexes. The molecule-part ontology holds knowledge about regions of molecules that do not exist independently, but need to be talked about by biologists. These would include amino acid side chains and the 3' and 5' ends of nucleic acid molecules. The data ontology captures knowledge about experimental detail as well as data on the structure of physical-things. The methods ontology contains information about techniques for analysing data. It holds knowledge of which techniques can be applied to which data, as well as the input and outputs of each method.
Instances are added to RiboWeb that correspond to these concepts. For example, a publication in a peer-reviewed article describes the three-dimensional structure of the 30s ribosomal subunit. This means linked instances need to be created in the publication, data and physical-thing ontologies. A user may want to see if this structure is consistent with others captured within RiboWeb [14]. The constraints described within RiboWeb can highlight conflicts with current knowledge to the biologist.