ACP_MS

ACP_MS predicts anticancer peptide sequences using monoMonoKGap feature extraction, AdaBoost feature selection, and stochastic gradient descent classification to support identification of antitumor peptides.

Key Features:

• Feature Extraction (monoMonoKGap): monoMonoKGap captures sequence-specific characteristics of anticancer peptides and converts them into digital features.
• Feature Selection (AdaBoost): AdaBoost refines the extracted features by selecting the most discriminative features for subsequent modeling.
• Sequence Identification (Stochastic Gradient Descent): Stochastic gradient descent is used to identify/classify anticancer peptide sequences by optimizing model parameters.
• Validation Techniques: Model performance is assessed using 7-fold cross-validation and independent test set validation.
• Performance Metrics: Main dataset accuracies are 92.653% (7-fold cross-validation) and 91.597% (independent test); alternate datasets show 98.678% (cross-validation) and 98.317% (independent test).
• Comparative Performance: Reported improved identification ability relative to other advanced prediction models.

Scientific Applications:

• Anticancer peptide identification: Detecting candidate anticancer peptides from sequence datasets.
• Peptide-based drug development: Supporting discovery and development of peptide therapeutics with antitumor activity.
• Mechanistic and translational research: Facilitating studies of molecular mechanisms underlying antitumor activity and supporting personalized medicine approaches in oncology.

Methodology:

Feature extraction with monoMonoKGap, feature selection via AdaBoost, sequence identification using stochastic gradient descent, evaluated by 7-fold cross-validation and independent test set validation.