Abstract
A large scale structural analysis of metabolic networks is presented focusing on neighbourhood relationships between individual reactions. We define two reactions to be neighbored if one of them provides the necessary set of substances for the other to proceed. A method is developed which allows determining all possible neighborhood relationships categorized as interaction patterns. These patterns differ in the types of participating reactions and in the way they share their reactants. The method is applied to a set of 4795 metabolic reactions contained in the KEGG database. We show that from the 1547 theoretically possible types of interactions 282 patterns are found in metabolism. More than 55% of all interactions occur between reactions with at most two reactants on one side. In these interactions only 25 different patterns play a role. We propose to use these neighborhood relationships as a concept of adjacency in large scale graph theoretical analyses of metabolism.
Original language | English |
---|---|
Pages (from-to) | 208-218 |
Number of pages | 11 |
Journal | Genome Informatics |
Volume | 17 |
Issue number | 1 |
Publication status | Published - 2006 |
Keywords
- Databases, Factual
- Glucokinase
- Metabolic Networks and Pathways
- Phosphotransferases
- Protein Interaction Mapping
- Substrate Specificity