Aids to model annotation exist, but rely extensively on the expert knowledge of the modeller for identification of appropriate additions. Saint (http://mygrid.ncl.ac.uk/saint), for example, is a lightweight integration environment wrapped in an easy-to-use web interface. The Saint interface allows modellers to easily upload an initial model, peruse and select appropriate annotation from that suggested by Saint, and then download the newly-annotated model. Similarly, Taverna workflows can be used to pull annotation from data sources to inform models [12]. semanticSBML (http://sysbio.molgen.mpg.de/semanticsbml/) focuses only on MIRIAM annotations, and does not add new model elements or any other type of information.
All of these systems have limitations in their understanding of the biology underlying the models themselves. While they resolve a certain amount of syntactic heterogeneity in the data sources, they are unable to process semantic heterogeneity. Semantic heterogeneity can occur when more than one data source uses the same word for a concept while defining that concept differently. For instance, a protein in BioPAX (http://www.biopax.org) is strictly defined as having only one polypeptide chain, while a protein in UniProtKB can consist of multiple chains.
The rule-based mediation approach has been implemented as a tool for model annotation via semantic data integration which pulls information for the modeller from data sources in a semantically as well as syntactically integrated way. Such an approach may return information more tailored to their needs and the biology of their models. Figure 2 shows rule-based mediation in the context of SBML model annotation.
RBM Home