VisualizeABiologicalModel

This space needs to be filled out with thoughts on what this actually means. The prerequisite is that we have a CellML model of a biological process. We want a 2D visualization of this. What needs to be considered is :

  • what biological entities and processes are we trying to represent
  • what standards are already out there that provide visual languges for this?
  • how do we represent the graphical information within a CellML model?
  • since we can reuse models in other models, do we want to be able to reuse graphical representations too? i.e. is there some process to build more complicated graphical represenations that use existing ones?

Discussion

CellMLOntologies discusses the representation of biological models that are associated with a CellML model. Associated with a biological model is often graphical representations. These may be simple representations of the biological entities and relations, or more complex representations of states in a dynamical system. A CellML model can contribute to both of these representations.

The graphical representation of biological models can be defined many different ways. Some groups have formed visual languages that suit specific domains of biology. For example the signaling gateway and the SBML layout extension.

The main problem is how to associate a particular graphical representation with a CellML model. The two main ways a graphical representation could be formed is either:

  1. associating graphical objects directly with CellML elements
  2. associating graphical objects with biological terms describing physical entities and relations and then associating these biological terms with the CellML elements. In this case there could be many different graphical representations already associated with these terms.

There doesn't seem to be any reason not to allow both. The discussion on CellMLOntologies asks the question about one to one relations between CellML elements and biological model elements. If some guideline is established to help this, then this would also serve the graphical representation aspect, especially where reuse of graphical elements is sought after.

When describing a particular graphical representation, it does make sense to associate a CellML element directly with a graphical element, and not simply a biological term(where we could have many choices for graphical elements).

A basic mechanism for associating graphical elements would be give each the equivalent of a cmeta:id within their own model representation. A CellML model is then free to form an association between a cmeta:id of a CellML model element with the id of a graphical element in some other representation system. There would need to be some way of identifying where this id came from, much like the base uri of a CellML model identifies where a particular cmeta:id originated, since there is no rule specifying that these ids are unique across all models.

A development proposal that addresses this is CellMLGraphicalRepresentation.