DLFF-ACP

DLFF-ACP predicts anticancer peptides (ACPs) by combining a dual-channel deep neural network ensemble with multi-view feature fusion to identify peptide sequences with anticancer potential.

Key Features:

• Dual-Channel Deep Neural Network Ensemble: A dual-channel deep neural network ensemble integrates two distinct channels to improve prediction accuracy.
• CNN-Based Spatial Feature Extraction: One channel uses a convolutional neural network (CNN) to automatically extract spatial features from peptide sequences.
• Handcrafted Feature Processing: The second channel processes handcrafted features, including composition-based features derived from k-spaced amino acid group pairs.
• Multi-View Feature Fusion: A feature fusion strategy integrates CNN-extracted spatial features with composition-based k-spaced amino acid group pair features for combined analysis.
• Performance Validation: The model was validated on two independent test sets, achieving an area under the curve (AUC) of 0.90 on the first test set and outperforming most other methods on the second.

Scientific Applications:

• Anticancer drug discovery and development: Computational prediction of ACPs to support discovery efforts in anticancer therapeutics.
• Exploration of novel peptide sequences: Integration of diverse feature sets to prioritize and explore peptide sequences with potential therapeutic properties.

Methodology:

The methodology includes CNN-based extraction of spatial features from peptide sequences, processing of handcrafted features including k-spaced amino acid group pairs, a dual-channel deep neural network ensemble, and multi-view feature fusion.