Topology-Enhanced Machine Learning Model (Top-ML) for Anticancer Peptide Prediction.

Area of Science:

• Biotechnology
• Computational Biology
• Drug Discovery

Background:

• Therapeutic peptides show promise for cancer treatment.
• Artificial intelligence (AI) aids in screening anticancer peptides.
• Efficient peptide featurization is a bottleneck for AI models.

Purpose of the Study:

• To propose a topology-enhanced machine learning (Top-ML) model for anticancer peptide prediction.
• To address the limitations of current peptide featurization methods in AI models.

Main Methods:

• Developed Top-ML model using peptide topological features from sequence connection information.
• Utilized spectral descriptors for characterizing peptide topology.
• Employed an Extra-Trees classifier for prediction.

Main Results:

• Validated Top-ML on AntiCP 2.0 and mACPpred 2.0 datasets.
• Achieved state-of-the-art or comparable performance to deep learning models.
• Demonstrated greater interpretability compared to existing methods.

Conclusions:

• Novel topology-based featurization accelerates anticancer peptide identification.
• Top-ML model shows significant potential for AI-driven cancer therapeutics.
• The approach enhances the efficiency and interpretability of machine learning in drug discovery.