Uncover disease genes by maximizing information flow in the phenome-interactome network

Area of Science:

• Genomics
• Computational Biology
• Systems Biology

Background:

• Identifying genes responsible for human inherited diseases is crucial for understanding disease pathogenesis.
• Existing methods for prioritizing candidate genes often use phenotype similarities and protein-protein interactions, but lack combinatorial approaches.

Purpose of the Study:

• To introduce the first combinatorial approach for prioritizing candidate genes underlying human inherited diseases.
• To develop a computational method (MAXIF) for maximizing information flow in a constructed network to uncover disease-related genes.

Main Methods:

• Constructed a phenome-interactome network integrating phenotype similarity, protein-protein interactions, and disease-gene associations.
• Developed the MAXIF computational method to maximize information flow within this network for gene prioritization.
• Validated the method through cross-validation experiments and comparison with state-of-the-art techniques.

Main Results:

• Demonstrated the effectiveness of the MAXIF method in prioritizing candidate disease genes.
• Showcased the method's capability to identify potential disease associations for query genes.
• Successfully applied the method to predict driver genes in melanoma copy number aberration regions, revealing new biological insights.

Conclusions:

• The proposed combinatorial approach offers a powerful new tool for identifying genes underlying inherited diseases.
• The MAXIF method provides competitive performance and robustness, advancing the field of disease gene discovery.
• The application in melanoma highlights the method's potential for uncovering novel driver genes and understanding their regulatory mechanisms.