RWRNET: A Gene Regulatory Network Inference Algorithm Using Random Walk With Restart.

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Gene regulatory network inference is crucial for understanding gene interactions and disease mechanisms.
• Existing computational methods often overlook global network topology, leading to redundant relationships.
• Optimizing network inference requires considering both local and global structural properties.

Purpose of the Study:

• To introduce a novel algorithm, Random Walk with Restart Network Inference (RWRNET), for gene regulatory network construction.
• To integrate local and global network topology information for more accurate inference.
• To improve the identification of complex regulatory relationships and disease mechanisms.

Main Methods:

• RWRNET captures local topology using random walk principles.
• Global topology is incorporated by applying Random Walk with Restart.
• The Markov Blanket discovery algorithm addresses isolated genes within the network.

Main Results:

• The proposed RWRNET method was evaluated against state-of-the-art algorithms.
• Performance was assessed using six benchmark gene expression datasets.
• Experimental results confirmed the superior effectiveness of RWRNET.

Conclusions:

• RWRNET offers an effective approach for gene regulatory network inference by integrating diverse topological information.
• The algorithm enhances the identification of gene regulatory relationships and aids in understanding disease mechanisms.
• This method provides a valuable tool for systems biology research.