GenClust

GenClust clusters gene expression data using a genetic algorithm to identify gene groups and support analysis of functional relationships and regulatory networks.

Key Features:

• Novel Coding of Search Space: A compact coding scheme represents candidate cluster configurations to facilitate efficient exploration of the search space.
• Integration with Data-Driven Internal Validation Methods: Compatible with data-driven internal validation and specifically tested with FOM (Fuzzy Overlap Measure) for cluster quality assessment in gene expression datasets.
• Rapid Convergence to Local Optima: Experimental validation indicates rapid convergence to local optima with the ability to identify meaningful clusters comparable to Average Link, CAST, CLICK, and K-means.

Scientific Applications:

• Gene expression clustering: Grouping genes with similar expression patterns to reveal functional relationships and regulatory networks.
• Comparative benchmarking: Comparative assessment of clustering performance against Average Link, CAST, CLICK, and K-means on real datasets.

Methodology:

GenClust employs a genetic algorithm framework with a compact search-space coding and evolutionary operators including selection, crossover, and mutation, and can integrate the FOM (Fuzzy Overlap Measure) for internal validation.