Ranking differential hubs in gene co-expression networks

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Identifying differential gene expression is crucial for understanding disease mechanisms.
• Differential network analysis detects changes in gene interaction networks.
• Existing methods often miss critical topological network changes due to pairwise comparisons.

Purpose of the Study:

• To propose a novel algorithm, DiffRank, for ranking genes based on their contribution to differential network structures.
• To address limitations of existing methods in capturing topological changes in biological networks.

Main Methods:

• Developed DiffRank, a gene ranking algorithm utilizing differential connectivity and differential betweenness centrality.
• Implemented score propagation through network structures for enhanced gene ranking.
• Utilized a synthetic differential scale-free network simulator for algorithm validation.

Main Results:

• DiffRank demonstrated superior performance compared to existing methods on synthetic datasets.
• The algorithm successfully identified biologically relevant results when applied to real gene expression datasets.
• The proposed method effectively captures essential topological changes in network structures.

Conclusions:

• DiffRank offers a powerful approach for identifying differentially expressed genes by considering network topology.
• The algorithm provides biologically meaningful insights into disease mechanisms.
• DiffRank advances differential network analysis by incorporating global and local structural measures.