ECCB 2002 Poster sorted by: Author | Number

Next | Previous poster (in order of the view you have selected)
Title: Analysis of the hierarchical structure of large scale metabolic networks constructed from genome data
P103
Ma, H.-W.; Aeng, A.-P.

aze@gbf.de
German Research Center for Biotechnology

Genome projects have provided us with a complete list of parts of life. However, how these isolated parts are integrated into complex functional systems such as metabolic network is still not well understood. Organism-specific metabolic networks have been constructed from genome data. Analysis of the large scale structure property of these networks is important for understanding the design and evolution of cellular metabolism. Previous structure analysis of metabolic networks has shown that they have a power law connection degree distribution and thus are small world networks. However, the connection degree distribution is not enough to describe the great diversity of cellular metabolism. In this work, we reveal that a hierarchical structure as shown in Fig.1 exists in the metabolic networks of 65 fully sequenced organisms. Four subsets of metabolites, i.e. a fully connected giant strong component (GSC), a substrate subset (S), a product subset (P) and an isolated subset (IS) are identified in the whole network. GSC is the most complex part in which all the metabolites can be converted to each other. All the metabolites in the subset S can be converted to the metabolites in GSC, while the metabolites in the subset P can be produced from the metabolites in GSC. The metabolites in the isolated subset are not connected with GSC. We calculated the average path lengths of GSC and the whole network respectively and found a linear relationship between them, indicating that the average path length of the whole metabolic network is mainly determined by that of GSC. Furthermore, the term ?overall closeness centralization index? is introduced to describe the centrality distribution of metabolites in GSC. It is found that the average path length decreases with the centralization index. For the three domains of organisms, eukaryotes and archaea have lower values of centralization index than bacteria and hence longer average path length. This shows that the centralization index is a good structure parameter to show the structure difference between metabolic networks of different organisms. In addition to this theoretical significance, unraveling this hierarchical structure can also give some useful hints on decomposition of the whole network into small subsystems which can then be further analyzed by pathway analysis method to improve the cellular properties for industrial and medical applications.

(For figures see web representation of the poster abstract)