Top-down network analysis to drive bottom-up modeling of physiological processes.

Area of Science:

• Systems Biology
• Computational Biology
• Molecular Interactomics

Background:

• Top-down systems biology approaches identify correlations but struggle with experimental design.
• Bottom-up approaches build detailed models for simulation but are labor-intensive.
• Existing methods for integrating these approaches are limited.

Purpose of the Study:

• To present LINKER, an automated, data-driven method for analyzing molecular interactomes.
• To propose extensions for simulated biological models based on interactome data.
• To bridge the gap between correlation discovery and experimental design in systems biology.

Main Methods:

• LINKER employs teleporting random walks and k-shortest path computations.
• It identifies connections from a source protein to a set of proteins involved in a specific cellular process.
• The method analyzes molecular interactomes to extend existing biological models.

Main Results:

• LINKER successfully proposed extensions to a dynamic model of the cell division cycle in Saccharomyces cerevisiae.
• Subnetworks identified by LINKER were significantly enriched in cell cycle-related Gene Ontology (GO) terms.
• The method elucidated the role of protein kinase Cdc5 in the mitotic exit network.

Conclusions:

• LINKER provides an efficient and automated way to generate testable hypotheses from large-scale biological data.
• The method enhances the predictive power of systems biology models by integrating interactome information.
• LINKER facilitates experimental design by suggesting specific molecular interactions and pathways for investigation.