Fusion of expression values and protein interaction information using multi-objective optimization for improving gene...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Functional genomics aims to identify genes controlling cellular mechanisms.
• Gene clustering relies on functional similarities and biological relevance.
• Existing methods may not fully capture complex gene relationships.

Purpose of the Study:

• To develop a multi-objective optimization genetic clustering technique for gene classification.
• To introduce a new quality measure, the protein-protein interaction confidence score, for gene clusters.
• To integrate gene expression data with protein-protein interaction information for biologically relevant gene selection.

Main Methods:

• Utilized a multi-objective optimization based genetic clustering approach.
• Developed a protein-protein interaction confidence score to assess gene cluster quality.
• Integrated microarray gene expression values and protein-protein interaction confidence scores.
• Optimized biological (homogeneity index, PPI confidence score) and traditional (fuzzy partition coefficient, Pakhira-Bandyopadhyay-Maulik-index) cluster validity indices.

Main Results:

• The proposed method demonstrated improved gene clustering compared to existing techniques.
• The integration of protein-protein interaction information enhanced the biological relevance of gene clusters.
• Experimental results on three real-life datasets validated the effectiveness of the approach.

Conclusions:

• The novel multi-objective genetic clustering approach effectively classifies genes based on functional and biological relevance.
• Incorporating protein-protein interaction data significantly improves gene clustering outcomes in functional genomics.
• This technique offers a robust solution for identifying key genes in complex biological pathways.