Hierarchical modular structure identification with its applications in gene coexpression networks.

Area of Science:

• Network science
• Graph theory
• Computational biology

Background:

• Network module (community) structure is a key area of research.
• Hierarchical modularity is observed in various real-world networks, including biological and social networks.
• Existing methods primarily focus on partitional module identification, with less emphasis on inferring hierarchical structures.

Purpose of the Study:

• To propose a novel method for constructing hierarchical modular structures in networks.
• To address the gap in research concerning the inference of hierarchical modularity.
• To validate the method using both artificial and real-world network data.

Main Methods:

• Utilizing the stochastic block model as the foundation for hierarchical structure construction.
• Applying statistical tests to rigorously evaluate hierarchical relationships between identified modules.
• Demonstrating the method's performance on synthetic networks and real biological data.

Main Results:

• The proposed method successfully constructs hierarchical modular structures.
• Statistical tests confirm the validity of hierarchical relationships between modules.
• Application to a yeast gene coexpression network identified hierarchical modules linked to specific gene functions.

Conclusions:

• The stochastic block model provides a robust framework for inferring hierarchical network modularity.
• The developed method effectively identifies hierarchical structures and their functional relevance in biological networks.
• This research contributes to a deeper understanding of complex network organization and function.