AMBIENT: Active Modules for Bipartite Networks - using high-throughput transcriptomic data to dissect metabolic response

作者：Bryant William A^*; Sternberg Michael J E; Pinney John W

来源：BMC Systems Biology, 2013, 7(1): 26.

DOI：10.1186/1752-0509-7-26

• Metabolic networks
• Simulated annealing
• High-throughput data
• Stress response
• Network analysis

Background: With the continued proliferation of high-throughput biological experiments, there is a pressing need for tools to integrate the data produced in ways that produce biologically meaningful conclusions. Many microarray studies have analysed transcriptomic data from a pathway perspective, for instance by testing for KEGG pathway enrichment in sets of upregulated genes. However, the increasing availability of species-specific metabolic models provides the opportunity to analyse these data in a more objective, system-wide manner.
Results: Here we introduce ambient (Active Modules for Bipartite Networks), a simulated annealing approach to the discovery of metabolic subnetworks (modules) that are significantly affected by a given genetic or environmental change. The metabolic modules returned by ambient are connected parts of the bipartite network that change coherently between conditions, providing a more detailed view of metabolic changes than standard approaches based on pathway enrichment.
Conclusions: ambient is an effective and flexible tool for the analysis of high-throughput data in a metabolic context. The same approach can be applied to any system in which reactions (or metabolites) can be assigned a score based on some biological observation, without the limitation of predefined pathways. A Python implementation of ambient is available at http://www.theosysbio.bio.ic.ac.uk/ambient.

• 出版日期2013-3-25