AntiCP

AntiCP predicts anticancer peptides using Support Vector Machine (SVM) models that analyze amino acid composition and binary profile features to identify peptides with anticancer activity.

Key Features:

• Model Development and Input Features: AntiCP employs SVM models that use amino acid composition and binary profile features to predict anticancer properties.
• Training Data: The primary training set includes experimentally validated anticancer peptides and random peptides sourced from the SwissProt database.
• Alternate Dataset: An alternate dataset comprises antimicrobial peptides with reported anticancer activities, with a positive set of 225 antimicrobial peptides.
• Performance Metrics: A binary profile-based SVM model achieved 91.44% accuracy and a Matthews Correlation Coefficient (MCC) of 0.83.
• Functional Capabilities: Features include predicting minimal sequence mutations to enhance anticancer potency, virtual screening of peptides for novel candidates, and scanning natural proteins to locate potential anticancer peptides.

Scientific Applications:

• Peptide discovery: Computational prediction and virtual screening to identify novel anticancer peptide candidates.
• Sequence optimization: Prediction of minimal mutations to improve anticancer activity of peptide sequences.
• Proteome mining: Scanning natural proteins to detect regions with potential anticancer activity.

Methodology:

SVM classifiers were trained using features derived from amino acid composition and binary profiles on datasets of experimentally validated ACPs, random SwissProt peptides, and an alternate set of 225 antimicrobial peptides with anticancer activity; positional analysis identified enrichment of Cysteine (Cys), Glycine (Gly), Isoleucine (Ile), Lysine (Lys), and Tryptophan (Trp).