Cancer classification in high dimensional microarray gene expressions by feature selection using eagle prey...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Microarray gene expression data are crucial for cancer classification and diagnosis.
• High dimensionality of microarray data poses challenges for effective feature selection.
• Existing computational methods for gene selection require continuous improvement.

Purpose of the Study:

• To introduce Eagle Prey Optimization (EPO), a novel genetically inspired algorithm for microarray gene selection.
• To identify a minimal yet informative subset of genes for accurate cancer subtype classification.
• To enhance the efficiency and accuracy of gene selection in high-dimensional cancer datasets.

Main Methods:

• Utilized Eagle Prey Optimization (EPO), inspired by eagle hunting strategies, for gene selection.
• Employed a genetic mutation operator combined with an EPO fitness function to evolve gene subsets.
• Incorporated a fitness function that considers gene discriminative power, diversity, and redundancy.
• Implemented adaptive mutation rates for efficient search space exploration.

Main Results:

• EPO demonstrated superior performance in cancer classification compared to state-of-the-art methods.
• Achieved significant dimensionality reduction while maintaining high classification accuracy.
• Showcased robustness to noise in microarray gene expression data.
• Consistently identified compact and informative gene subsets across multiple cancer datasets.

Conclusions:

• Eagle Prey Optimization (EPO) is an effective and robust algorithm for microarray gene selection in cancer classification.
• EPO offers a promising approach for improving diagnostic accuracy and reducing the complexity of cancer data analysis.
• The algorithm's ability to balance gene discriminative power, diversity, and redundancy is key to its success.