CorGO: An Integrated Method for Clustering Functionally Similar Genes.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Identifying co-expressed gene groups is crucial for understanding disease mechanisms, particularly in cancer.
• Existing clustering algorithms using gene expression data often fail to capture essential functional similarities between genes.

Purpose of the Study:

• To introduce CorGO, a new algorithm designed to identify functionally similar gene clusters.
• To integrate gene expression correlation with Gene Ontology (GO) semantic similarity for improved biological relevance.

Main Methods:

• CorGO calculates gene relationships using both correlation from gene expression data and GO-based semantic similarity.
• A novel similarity measure is defined by integrating these two components.
• The algorithm was applied to gene expression profiles from four cancer types.

Main Results:

• CorGO successfully identified functionally similar gene clusters across different cancer types.
• Validation through pathway enrichment, disease enrichment, and network analysis confirmed significant functional relatedness within identified clusters.
• Comparative analysis demonstrated CorGO's superior efficacy over commonly used clustering algorithms.

Conclusions:

• CorGO effectively identifies functionally homogeneous gene clusters by combining expression correlation and GO semantic similarity.
• The algorithm provides enhanced biological insights into disease mechanisms, outperforming existing methods.