Gene expression data mining by hybrid biclustering with improved GA and BA.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Clustering gene expression data is vital for understanding biological processes and diseases.
• High dimensionality, complexity, and noise in gene expression data pose significant clustering challenges.
• Conventional clustering methods often fall short in accuracy and efficiency.

Purpose of the Study:

• To develop an advanced dual clustering method for gene expression data analysis.
• To enhance heuristic algorithms (genetic and bat algorithms) for improved optimization.
• To overcome the limitations of existing clustering approaches in bioinformatics.

Main Methods:

• Integration of enhanced genetic and bat algorithms into a dual clustering framework.
• Improvement of the optimization processes within the genetic and bat algorithms.
• Evaluation of the method's performance on single-peak and multi-peak functions.

Main Results:

• The enhanced genetic and bat algorithms demonstrated superior convergence speed and accuracy compared to other heuristic algorithms.
• The proposed dual clustering method resulted in farther inter-cluster distances and closer intra-cluster distances.
• Key metrics (geometric mean 0.99, silhouette coefficient 1.0, Davies-Bouldin index 0.2, adjusted rand index 0.92) outperformed other dual clustering methods.

Conclusions:

• The novel dual clustering method effectively achieves high inter-cluster variability and intra-cluster similarity.
• This approach significantly enhances the efficiency and accuracy of gene expression data analysis.
• The method provides robust technical support for advancing bioinformatics research.