Topology-enhanced machine learning model (Top-ML) for anticancer peptide prediction

Joshua Zhi En Tan,JunJie Wee,Xue Gong,Kelin Xia

2024-07-12

Abstract:Recently, therapeutic peptides have demonstrated great promise for cancer treatment. To explore powerful anticancer peptides, artificial intelligence (AI)-based approaches have been developed to systematically screen potential candidates. However, the lack of efficient featurization of peptides has become a bottleneck for these machine-learning models. In this paper, we propose a topology-enhanced machine learning model (Top-ML) for anticancer peptide prediction. Our Top-ML employs peptide topological features derived from its sequence "connection" information characterized by vector and spectral descriptors. Our Top-ML model has been validated on two widely used AntiCP 2.0 benchmark datasets and has achieved state-of-the-art performance. Our results highlight the potential of leveraging novel topology-based featurization to accelerate the identification of anticancer peptides.

Quantitative Methods,Machine Learning,General Topology,Biomolecules

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address the bottleneck issue in the prediction of Anticancer Peptides (ACPs), specifically how to efficiently characterize peptides to improve the performance of machine learning models. Specifically, the research team proposes a topology-enhanced machine learning model (Top-ML) for predicting peptides with anticancer properties. ### Specific Problems and Methods 1. **Background and Challenges**: - Cancer remains one of the leading causes of death worldwide. Traditional treatment methods (such as radiotherapy and chemotherapy) have significant limitations, such as systemic toxicity and drug resistance. - Anticancer peptides have emerged as a novel alternative for cancer treatment due to their high specificity, low toxicity, and ease of chemical modification. - However, experimental methods for discovering and designing anticancer peptides are costly, time-consuming, and labor-intensive, making large-scale application difficult. 2. **Research Methods**: - **Topological Features**: Utilize Topological Data Analysis (TDA) to extract topological features from peptide sequences, including vector features and spectral features. - **Feature Types**: Combine four types of peptide features: Magnus vector, natural vector, terminal composition features, and spectral features. - **Machine Learning Model**: Train the model by combining these features and using an Extra Trees Classifier, validating the model's performance on benchmark datasets. 3. **Results and Contributions**: - The proposed Top-ML model achieved state-of-the-art performance on two widely used AntiCP 2.0 benchmark datasets. - The results indicate that utilizing topological features can significantly enhance the accuracy of anticancer peptide recognition. Through these methods, the research team demonstrates the potential of mathematically-assisted peptide features in improving the performance of anticancer peptide recognition.

Topology-enhanced machine learning model (Top-ML) for anticancer peptide prediction

ACP-ML: A sequence-based method for anticancer peptide prediction

ACP-DL: A Deep Learning Long Short-Term Memory Model to Predict Anticancer Peptides Using High-Efficiency Feature Representation

Efficient prediction of anticancer peptides through deep learning

Prediction of Anticancer Peptides Using a Low-Dimensional Feature Model

ACPNet: A Deep Learning Network to Identify Anticancer Peptides by Hybrid Sequence Information

Anticancer peptides prediction with deep representation learning features

Improved prediction of anti-angiogenic peptides based on machine learning models and comprehensive features from peptide sequences

Integrating In Silico and In Vitro Approaches to Identify Natural Peptides with Selective Cytotoxicity against Cancer Cells

To Assist Oncologists: An Efficient Machine Learning-Based Approach for Anti-Cancer Peptides Classification

Machine learning for antimicrobial peptide identification and design

MLASM: Machine learning based prediction of anticancer small molecules

ACP-check: An anticancer peptide prediction model based on bidirectional long short-term memory and multi-features fusion strategy

Prediction of anticancer peptides based on an ensemble model of deep learning and machine learning using ordinal positional encoding

Contrastive learning for enhancing feature extraction in anticancer peptides

Peptide-Based Drug Predictions for Cancer Therapy Using Deep Learning

ACP-MLC: A two-level prediction engine for identification of anticancer peptides and multi-label classification of their functional types

mACPpred 2.0: Stacked Deep Learning for Anticancer Peptide Prediction with Integrated Spatial and Probabilistic Feature Representations

ITP-Pred: an interpretable method for predicting, therapeutic peptides with fused features low-dimension representation

An efficient consolidation of word embedding and deep learning techniques for classifying anticancer peptides: FastText+BiLSTM

MA‐PEP: A novel anticancer peptide prediction framework with multimodal feature fusion based on attention mechanism