iACP-GAEnsC: Evolutionary genetic algorithm based ensemble classification of anticancer peptides by utilizing hybrid...

Area of Science:

• Computational Biology
• Bioinformatics
• Drug Discovery

Background:

• Cancer is a leading cause of death, with conventional therapies like chemotherapy and radiation being costly and having severe side effects.
• There is a critical need for accurate and efficient computational models to identify anticancer peptides, aiding in the development of novel cancer treatments.

Purpose of the Study:

• To propose an intelligent computational model, iACP-GAEnsC, for the accurate identification of anticancer peptides.
• To evaluate the effectiveness of different feature representation methods and ensemble techniques for enhancing classification accuracy.

Main Methods:

• Developed an ensemble model (iACP-GAEnsC) utilizing a genetic algorithm for optimizing classifier combinations.
• Employed three feature representation methods: amphiphilic Pseudo amino acid composition, g-Gap dipeptide composition, and Reduced amino acid alphabet composition.
• Investigated individual and hybridized feature spaces, combining predictions using genetic algorithm optimization and majority voting.

Main Results:

• The iACP-GAEnsC model achieved a high accuracy of 96.45% on a hybrid feature space.
• The genetic algorithm-based ensemble classification significantly outperformed individual classifiers and simple majority voting.
• Demonstrated remarkable improvements in various performance metrics compared to existing techniques.

Conclusions:

• The iACP-GAEnsC model shows significant potential as a leading tool in drug design and proteomics.
• The proposed model offers a more accurate and efficient approach for identifying anticancer peptides.
• Hybridization of feature spaces and ensemble learning enhances predictive performance for anticancer peptide identification.