It is now possible to formulate and simulate models that account comprehensively for the large numbers of molecules and molecular interactions during that typically comprise a cell sig naling system, which raises the issue of how to annotate and visualize large scale rule based models. Visualization of the elements of a rule based model is natural to some extent because rule based modeling, at least in some realizations, is based on or can be inter preted as being based on an underlying graphical form alism, which serves as the foundation for the BioNetGen language. This model specifi cation language is supported by anumber of software tools. Another model specification languageisKappa, which is closely related to BNGL.

In the BNGL formalism, which is briefly summarized in this section and described in greater detail below, graphs are used to represent molecules, and graph rewriting rules are used to represent mole cular interactions. In a rule based model for a cell signaling system, the graphs of a model typically represent proteins, which are taken to be the building blocks of most chemical species in the system. These graphs can be visualized according to the conventions of Faeder et al. A graph representing a protein includes a colored vertex for each functional component of the protein. The color represents the type of protein being represented by a graph, i. e. the protein name is essentially the color of the graph representing the protein. The vertices of graphs can be associated with variable attributes to represent so called internal states of components.

An internal state is an abstraction that is often useful for representing, for example, the phosphorylation status of an amino acid residue. In graphs for molecular com plexes, edges are used to represent bonds between mole cular components. Thus, the composition and the connectivity of a molecular complex are tracked explicitly in a BNGL encoded rule based model. In general, the graph rewriting rules in a BNGL encoded model specify simple operations on graphs, which define the outcomes of molecular interactions, the addition of an edge to represent an association event, the removal of an edge to represent a dissociation event, or the change of a vertex attribute to represent an internal state change, such as a post translational modification event. Rules can also be specified for synthesis and degradation reactions.

Two important features of a rule are the specification of a reaction cen ter and the specification of the molecular context in GSK-3 which a molecular interaction occurs, i. e. the necessary and sufficient conditions that must be satisfied for a reaction to occur. Another feature of a rule is an associated rate law, which is used to characterize all reactions implied by the rule up to statistical factors, which are derived from the properties of reactants.