HybridGWOSPEA2ABC: a novel feature selection algorithm for gene expression data analysis and cancer classification.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• DNA micro-array technology is crucial for cancer research, gene function studies, and classification.
• Machine Learning (ML) techniques are increasingly used for cancer classification via gene pattern analysis.
• Efficient gene selection is a significant challenge in cancer diagnosis and treatment.

Purpose of the Study:

• To introduce a novel hybrid algorithm for optimizing gene selection in cancer classification.
• To enhance solution diversity, convergence efficiency, and exploration/exploitation in high-dimensional gene expression data.
• To validate the algorithm's effectiveness against existing bio-inspired methods.

Main Methods:

• Integration of Grey Wolf Optimizer (GWO), Strength Pareto Evolutionary Algorithm 2 (SPEA2), and Artificial Bee Colony (ABC) into a hybrid algorithm.
• Application of the hybrid algorithm for feature selection on various cancer datasets.
• Comparative analysis against five bio-inspired algorithms using five different classifiers.

Main Results:

• The HybridGWOSPEA2ABC algorithm outperformed conventional bio-inspired algorithms in identifying cancer biomarkers.
• The hybrid approach demonstrated enhanced capabilities in handling high-dimensional data for gene selection.
• Superior performance was observed in addressing the challenges of cancer classification using gene expression data.

Conclusions:

• The novel hybridization algorithm improves performance by maintaining solution diversity and efficient convergence.
• The study advances gene selection methodologies for improved cancer detection and classification.
• Provides a better understanding of relevant genes for accurate cancer classification.