A Dual Level Analysis with Evolutionary Computing and Swarm Models for Classification of Leukemia.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Cancer, particularly leukemia, is a major cause of mortality, necessitating accurate diagnostic tools.
• Microarray data is vital for cancer diagnosis but suffers from high dimensionality, requiring effective gene selection.
• Identifying the most informative genes is critical for improving data classification and diagnostic accuracy.

Purpose of the Study:

• To evaluate and integrate multiple feature selection techniques for identifying informative genes from high-dimensional microarray data.
• To enhance gene selection by incorporating evolutionary optimization algorithms for improved cancer diagnosis.
• To determine the optimal combination of feature selection and optimization methods for leukemia classification.

Main Methods:

• Initial gene selection using Minimum Redundancy Maximum Relevance (MRMR), Signal to Noise Ratio (SNR), Multivariate Error Weight Uncorrelated Shrunken Centroid (EWUSC), and multivariate correlation-based feature selection (CFS).
• Application of five evolutionary optimization techniques: African Buffalo Optimization (ABO), Artificial Bee Colony Optimization (ABCO), Cockroach Swarm Optimization (CSO), Imperialist Competitive Optimization (ICO), and Social Spider Optimization (SSO).
• Classification of optimized gene sets using Probabilistic Neural Network (PNN).

Main Results:

• The combination of multivariate correlation-based feature selection (CFS) with Social Spider Optimization (SSO) proved most effective.
• This optimal gene selection strategy, when classified with a Probabilistic Neural Network (PNN), achieved a high classification accuracy of 95.70%.
• The study demonstrates the efficacy of integrating advanced feature selection and evolutionary optimization for cancer diagnosis.

Conclusions:

• The proposed hybrid approach of multivariate CFS and SSO offers a powerful method for selecting informative genes in high-dimensional microarray data.
• This gene selection strategy significantly enhances the accuracy of leukemia classification.
• The findings support the use of computational methods for improving cancer diagnostic capabilities.