There are multiple roles assignable to each path. Input and output roles are used to define the flow of material. Parameters are objects influencing the process, e.g. as inhibitors or cofactors. Paths leading to locations determine where the processes can be observed. In the example we defined at least defined six paths. Different metabolites act as input, as output and as parameter. The same enzymes are defined as input and as output of the process, while organisms and pathways give us information about the location.

Actually, the construction of a path is very simple, but once defined such a pattern of a process can act as a rule to calculate instances of processes from object-oriented models. For each object of the class Reaction a single processes will be created and the paths defined in the pattern will be traced starting at the current object. All objects found at the end of the traced paths are then assigned to the processes in the role in which they have been defined in the model.

Fig. 11.12 Relationship between conceptual mode (1) and objects level (2).

The combination of classes of processes leads to conceptual process oriented models, which are bipartite graphs containing classes of processes and classes objects. Figure 11.12 shows the relationship between a conceptual model (1) and the instances defined by the model (2). The example contains two object classes Metabolite and Enzyme and two classes of processes A and B which are consuming and producing objects of the classes Metabolite and Enzyme in different directions. In fact, these two patterns could describe the catalysis as a reversible and discrete process.

In the second part of Figure 11.12 we see a process oriented network including different types of objects and processes. It is clear that a conceptual model with a relatively low number of elements can describe the structure behind a very complex network. The advantage of the approach consists in the automation of building the instances from conceptual models. The remaining task for the modeller is to identify the types of processes and to model the patterns. Depending on the size of the object oriented model, the automated reconstruction of processes produces a large number of element which are interconnected in large scale networks. For performance enhancements in the further analysis of the, we materialize all of the computed processes, which means that they are stored directly within the model in a database.

0 0

Post a comment