Optimized Gene Selection Using Nomadic People and Salp Swarm Algorithms for Cancer Detection.

Area of Science:

• Bioinformatics
• Computational Biology
• Machine Learning in Oncology

Background:

• Gene expression analysis is vital for cancer detection but faces challenges like high dimensionality and redundancy.
• These challenges lead to overfitting and computational inefficiency in traditional methods.
• Effective gene selection is crucial for accurate cancer diagnosis and personalized treatment.

Purpose of the Study:

• To propose a hybrid feature selection framework integrating filter-wrapper approaches with swarm intelligence for optimized gene selection.
• To enhance cancer detection accuracy and computational efficiency using a novel NPO-SSVM method.
• To address the limitations of high dimensionality and redundancy in gene expression datasets.

Main Methods:

• A hybrid filter-wrapper approach combining Nomadic People Optimizer (NPO) and Mutual Information (MI) for gene subset identification.
• An optimized Support Vector Machine (SVM) classifier with hyperparameters tuned by an enhanced Salp Swarm Algorithm (SSA) incorporating a crossover operator.
• Validation on five cancer gene expression datasets: LUAD, GSE2034 (breast cancer), GBM, GSE2109 (ovarian cancer), and COAD.

Main Results:

• The proposed NPO-SSVM achieved high classification accuracies (91.25%-97.02%) and AUC-ROC values (0.85-0.97) across diverse cancer datasets.
• Achieved an average feature reduction of 51%, significantly reducing data complexity.
• Outperformed existing state-of-the-art methods (GA, PSO, GWO, SSA) by 3-12% in accuracy and 15% in computational efficiency.

Conclusions:

• The NPO-SSVM framework offers a robust and efficient solution for gene selection in cancer detection.
• Demonstrates significant advancements in personalized medicine and early cancer diagnosis through improved biomarker identification.
• Highlights the potential of hybrid swarm intelligence approaches for complex biological data analysis.