Knowledge-based generalization of metabolic networks: An applicational study

Anna Zhukova 1, 2 David James Sherman 1, 2
1 MAGNOME - Models and Algorithms for the Genome
Inria Bordeaux - Sud-Ouest, UB - Université de Bordeaux, CNRS - Centre National de la Recherche Scientifique : UMR5800
Abstract : Genome-scale metabolic networks are complex systems that describe thousands of reactions participating in the organism's metabolism. During the process of genome network reconstruction these reactions are automatically inferred from pathway and reaction databases, and existing models for similar organisms, using genomic data [1]. The inferred draft network is then refined during several iterations of error detection, gap filling, analysis and improvement [2]. Although automatic tools for model inference and analysis are becoming more and more powerful, they may still miss some reactions or add those ones that should not belong to the network of the target organism. That is why the analysis by a human expert is needed during the network refinement process. However, being tailored for a computer simulation, and thus including all the reactions thought to participate in the organism's metabolism, genome-scale networks can be too complicated and detailed for a human. The errors may be hidden in the multitude of reactions. To help a human expert understanding these detailed networks, we developed a method for knowledge-based generalization that focusses on the higher-level relationships in the network, while omitting the details [3]. The generalization process groups chemical species present in the network into semantically equivalent classes, based on their hierarchical relationships in the ChEBI ontology [4], and merges them into a generalized chemical species. For instance, butyryl-CoA, hexanoyl-CoA and octanoyl-CoA species can be generalized into fatty acyl-CoA. After the species generalization, reactions that share the same generalized reactants and the same generalized products, are factored together into a generalized reaction. This provides a higher-level view of the network. In this poster, we show the application of this generalization method to the network of the yeast Y. lypolitica [5]. We analyze the generalized network, and illustrate how it can be used for easier error detection: We show the changes that both initial and generalized networks undergo if the catalyzing enzyme for some of the reactions is missing. The application of the generalization procedure also facilitates network comparison, which we show by comparing the generalized Y. lypolitica network to the networks of several other organisms.
Type de document :
Communication dans un congrès
Moscow Conference on Computational Molecular Biology, Jul 2013, Moscow, Russia. 2013
Liste complète des métadonnées

https://hal.inria.fr/hal-00859440
Contributeur : David James Sherman <>
Soumis le : dimanche 8 septembre 2013 - 10:00:49
Dernière modification le : jeudi 11 janvier 2018 - 06:22:12
Document(s) archivé(s) le : lundi 9 décembre 2013 - 03:05:12

Identifiants

  • HAL Id : hal-00859440, version 1

Citation

Anna Zhukova, David James Sherman. Knowledge-based generalization of metabolic networks: An applicational study. Moscow Conference on Computational Molecular Biology, Jul 2013, Moscow, Russia. 2013. 〈hal-00859440〉

Partager

Métriques

Consultations de la notice

311

Téléchargements de fichiers

209