A multi-objective gene clustering algorithm guided by apriori biological knowledge with intensification and...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Understanding genetic basis of diseases requires analyzing high-throughput microarray data.
• Clustering is vital for grouping genes with similar expression profiles, but standard methods struggle with multi-dimensional data.
• Existing cluster validity indexes may not ensure biological coherence alongside expression similarity.

Purpose of the Study:

• To develop a Multi-Objective Clustering algorithm Guided by a-Priori Biological Knowledge (MOC-GaPBK).
• To identify gene clusters exhibiting high co-expression, biological coherence, compactness, and separation.
• To improve upon existing microarray clustering techniques by integrating biological knowledge.

Main Methods:

• The MOC-GaPBK algorithm optimizes multiple cluster quality indexes simultaneously, considering both expression levels and biological functionality.
• It employs intensification and diversification strategies to enhance the search process for optimal gene groupings.
• Cluster quality is assessed using criteria for compactness, separation, co-expression, and biological coherence.

Main Results:

• The MOC-GaPBK algorithm demonstrated superior performance on four public microarray datasets.
• Comparative analyses showed significant improvements over widely used microarray clustering techniques.
• Statistical, visual, and biological significance tests confirmed the algorithm's effectiveness.

Conclusions:

• Integrating a-priori biological knowledge into multi-objective clustering significantly enhances solution quality.
• The proposed MOC-GaPBK algorithm outperforms existing methods in co-expression, biological coherence, compactness, and separation.
• The study highlights the benefit of combining computational strategies with biological insights for gene expression analysis.