ACP-DA

ACP-DA leverages data augmentation and machine learning to improve prediction accuracy of anticancer peptides (ACPs) for computational identification of potential therapeutic peptides.

Key Features:

• Data Augmentation: Employs augmentation in the feature space to generate synthetic samples and increase diversity and size of training datasets.
• Feature Representation: Encodes peptide sequences using binary profile features and AAindex features derived from the AAindex database that capture amino acid presence and physicochemical/evolutionary properties.
• Machine Learning Integration: Trains machine learning models on the augmented feature-space dataset to predict anticancer peptides (ACPs) with improved generalization.

Scientific Applications:

• Oncology Research: Supports computational identification and prioritization of anticancer peptides (ACPs) for cancer biology studies.
• Drug Discovery and Peptide Screening: Enables screening of peptide libraries and prioritization of candidates for experimental validation in anticancer therapeutic development.

Methodology:

Peptide sequences are encoded with binary profile and AAindex features. The training dataset is expanded by generating synthetic samples in the feature space. Machine learning models are trained on the augmented dataset to predict ACPs.