ACP-ADA: A Boosting Method with Data Augmentation for Improved Prediction of Anticancer Peptides
Sadik Bhattarai1, Kyu-Sik Kim2, Hilal Tayara3
1Department of Electronics and Information Engineering, Jeonbuk National University, Jeonju 54896, Korea.
International Journal of Molecular Sciences
|October 27, 2022
Summary
Identifying anticancer peptides (ACPs) computationally is crucial for drug development. A new Ada-boosting algorithm, ACP-ADA, effectively predicts ACP candidates using integrated features and data augmentation, outperforming existing methods.
Area of Science:
- Biotechnology
- Computational Biology
- Oncology
Background:
- Cancer is a leading global cause of death.
- Therapeutic peptides targeting cancer cells are of significant interest.
- Identifying novel anticancer peptides (ACPs) via traditional experiments is costly and inefficient.
Purpose of the Study:
- To develop an effective computational approach for identifying anticancer peptide candidates.
- To propose a novel machine learning model for ACP prediction.
Main Methods:
- An Ada-boosting algorithm (ACP-ADA) was developed using random forest as the base learner.
- Peptides were represented using a 210-dimensional feature vector integrating binary profile, amino acid index, and amino acid composition.
- Training samples were augmented to enhance model performance with limited data.
Main Results:
- ACP-ADA demonstrated superior performance compared to existing methods.
- Five-fold cross-validation was used to optimize model parameters.
- The model achieved high accuracy (e.g., 86.4% on ACP740) and Mathew's correlation coefficient (e.g., 74.01% on ACP740).
Conclusions:
- ACP-ADA is a highly effective computational tool for identifying anticancer peptides.
- The integration of diverse features and data augmentation significantly improves prediction accuracy.
- This approach can accelerate drug development and advance biomedical research.
