Causal Disturbance Analysis: A Novel Graph Centrality Based Method for Pathway Enrichment Analysis.

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Pathway enrichment analysis (PEM) models interpret gene expression data using prior biological knowledge.
• Standard PEM often simplify analyses by ignoring gene interactions, potentially leading to incomplete or inaccurate biological insights.
• Interactions between genes are crucial for understanding pathway functionality and biological mechanisms.

Purpose of the Study:

• To introduce Causal Disturbance Analysis (CADIA), a novel graph-based pathway enrichment analysis model.
• To leverage gene interaction networks for more accurate interpretation of high-throughput gene expression data.
• To develop a new method that addresses the limitations of traditional PEM by incorporating topological gene importance.

Main Methods:

• Developed Causal Disturbance Analysis (CADIA), a graph-based pathway enrichment analysis model.
• Introduced a novel graph centrality model, Source/Sink Centrality, to quantify gene topological importance within pathways.
• Integrated Source/Sink Centrality with classical over-representation analysis to infer pathway enrichment scores.

Main Results:

• CADIA uniquely identifies critical pathway enrichments not detectable by other PEM.
• The Source/Sink Centrality model effectively prioritizes genes likely to disturb pathway functionality based on their topological importance.
• Evaluations using real-world and synthetic data demonstrate CADIA's sensitivity to topologically central gene expression changes.

Conclusions:

• CADIA provides a more informative framework for interpreting high-throughput data by incorporating gene interactions.
• The method enhances the accuracy and completeness of pathway enrichment analysis.
• CADIA's approach offers a sensitive and robust tool for uncovering key biological insights from complex datasets.