Directed random walks and constraint programming reveal active pathways in hepatocyte growth factor signaling

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Pathway databases offer limited specificity for analyzing mRNA expression data.
• Canonical pathways may not fully capture complex gene interactions across different pathways.
• Few methods exist for generating and analyzing causal gene regulatory networks with qualitative logic.

Purpose of the Study:

• To develop and present an algorithm (MCWalk) for discovering context-specific gene regulatory subgraphs.
• To infer gene-regulation patterns from expression data within specific biological contexts.
• To identify key regulator nodes and predict expression shifts in signaling networks.

Main Methods:

• Developed MCWalk, an algorithm integrating logic programming and random walks for subgraph discovery in signaling networks.
• Applied MCWalk to analyze hepatocyte growth factor-stimulated cell migration and proliferation using gene expression data.
• Utilized perturbation experiments and Answer Set programming to validate discovered networks against experimental data.

Main Results:

• Uncovered signal transduction mechanisms in a gene interaction network relevant to cell migration and proliferation.
• Identified associations between hepatocyte growth factor receptor c-Met, differentially expressed genes, and cellular states.
• Discovered key regulator nodes and accurately predicted expression shifts consistent with experimental measurements.

Conclusions:

• MCWalk provides an automated pipeline for discovering context-specific signaling networks.
• The method effectively integrates gene expression data with network analysis to reveal biological mechanisms.
• This approach enhances understanding of complex cellular processes like migration and proliferation.