Reducing the complexity of complex gene coexpression networks by coupling multiweighted labeling with topological...

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Gene coexpression networks are valuable for identifying disease biomarkers but often contain numerous false positives.
• High complexity and connectivity in these networks necessitate effective filtering and feature selection.

Purpose of the Study:

• To develop an efficient multivariate filtering algorithm for analyzing coexpression network topology.
• To identify potentially relevant genes associated with specific diseases by reducing network complexity.

Main Methods:

• The study proposes a novel multivariate filtering approach to analyze network topological properties.
• The algorithm was tested on expression data from three diseases: acute myeloid leukemia, breast cancer, and diffuse large B-cell lymphoma.

Main Results:

• The developed algorithm successfully filtered coexpression networks, reducing complexity and identifying relevant genes.
• Validation using literature mining confirmed the biological relevance of the identified genes for the studied diseases.

Conclusions:

• The proposed filtering method offers an efficient way to extract meaningful biological insights from complex gene coexpression networks.
• This approach aids in discovering potential disease biomarkers and advancing our understanding of disease mechanisms.