Modern techniques in molecular biology produce enormous amounts of biological data. Entire genomes have been sequenced, expression data for almost complete sets of genes of many organisms is now available. In order to cope with the data and to extract a maximal amount of information from it, sophisticated tools for the comprehensive and partly automatical analysis of the data have to be developed.
As the focus shifts from genes to genomes, the analysis of metabolic pathways becomes important. Systems like KEGG [Ogata et al., 1999] provide a nearly complete but static view of the current knowledge in this eld. A system for a dynamic modelling of metabolic pathways [Fellenberg and Mewes, 1999] would overcome many of the limitations of such static encyclopaedias. Here, we present an objectoriented data model for a pathway database. A system for dynamic modelling of metabolic pathways heavily relies on such a data model.
Objectoriented modelling allows a realistic description of biological entities and their relationships. Interconnections between the classes are modelled via dependencies like heritage, association and aggregation. These capabilities make an object oriented approach ideally suited for the described task. The data objects can be stored persistantly in an objectoriented database that extends the functionality of the data model naturally [Eldridge and Mancuso, 1998].
We provide a set of classes that represent the molecules found in cells and organisms and the way they interact. The system integrates the GAMS [Bents and Mewes, 1999] data model and extends it with classes that represent biochemical pathways and their reactions (Fig. 1).
|
The class Pathway contains a set of Reaction objects, where each Reaction consists of several MolecularElements that represent the reactants and products of the reaction as well as cofactors.
A typical reaction in a metabolic pathway is the EnzymaticReaction which subclasses Reaction. The EnzymaticReaction is organism specic and corresponds directly to a Protein while Reactions represent reaction mechanisms related to ECnumbers.