SSGA and MSGA: two seed-growing algorithms for constructing collaborative subnetworks.

Area of Science:

• Systems Biology
• Bioinformatics
• Genomics

Background:

• Collaborative networks of transcription factors (TFs) are crucial for controlling biological processes and complex traits.
• Previous work established methods for genome-wide TF coexpression network construction and one algorithm (Triple-Link) for subnetwork building.

Purpose of the Study:

• To develop and evaluate novel algorithms for constructing collaborative TF subnetworks.
• To compare the performance of new algorithms against an existing method in terms of functional cohesion and network properties.

Main Methods:

• Developed two new algorithms: Single Seed-Growing Algorithm (SSGA) and Multi-Seed Growing Algorithm (MSGA).
• These algorithms identify triple-node seeds from decomposed networks and grow them into subnetworks using distinct strategies.
• Comparative appraisal of subnetworks based on functional cohesion, intra-subnetwork association strength, and inter-subnetwork association strength.

Main Results:

• SSGA and MSGA outperformed the Triple-Link Algorithm in generating more functional and cohesive TF subnetworks.
• The new algorithms incorporated more systemic analyses of edge weights and network connectivity during construction.
• A framework was presented for utilizing these algorithms to identify TF sets governing biological processes.

Conclusions:

• SSGA and MSGA offer improved approaches for acquiring collaborative TF subnetworks.
• These algorithms provide valuable tools for dissecting the regulatory mechanisms underlying biological processes and complex traits.
• The developed methods enhance the ability to identify functionally relevant groups of TFs.