Inferring meaningful pathways in weighted metabolic networks

Area of Science:

• Metabolic network analysis
• Computational biology
• Systems biology

Background:

• Metabolic networks represent biochemical reactions within organisms.
• Understanding metabolic pathways is crucial for biological research.
• Previous methods often struggle with accuracy in pathway identification.

Purpose of the Study:

• To develop and evaluate a novel approach for computing meaningful metabolic pathways.
• To improve the accuracy of metabolic pathway inference.
• To investigate the properties of weighted metabolic networks.

Main Methods:

• Representing metabolic networks as weighted graphs, where compound weights reflect participation in reactions.
• Employing pathfinding algorithms to identify lowest-weight paths.
• Systematically evaluating computed pathways against annotated ones in raw, filtered, and weighted graphs.

Main Results:

• The weighted graph approach achieved approximately 85% correspondence with annotated pathways, significantly outperforming raw (<30%) and filtered (65%) graphs.
• Considering the five lightest paths increased correspondence to 92% on average.
• The weighted graph revealed a larger average distance between metabolites, challenging previous small-world network assumptions and suggesting pool metabolites create shortcuts.

Conclusions:

• A weighted graph approach provides a more accurate method for computing biologically meaningful metabolic pathways.
• The shortest path length in the weighted graph serves as a valid measure of metabolic distance between enzymes.
• This approach holds promise for inferring metabolic pathways in newly sequenced genomes.