Filters


Type of Resource

Top Features

Machine learning

Deep learning

Anticancer peptides

Antimicrobial peptides

Anti-inflammatory peptides

Antiviral peptides

Ensemble learning

Feature selection

Random forest

Cell-penetrating peptides


mACPpred

Webserver
Anticancer peptides Support Vector Machine Feature Selection Optimal features Sequential forward search

mACPpred is a machine learning model for precise prediction of anticancer peptides (ACPs), employing a two-step feature selection protocol on seven encoding types covering sequence composition, physicochemical properties, and profile features. After extracting optimal features, it uses prediction probabilities as input vectors for support vector machine (SVM) integration.

mACPpred 2.0

Webserver
Anticancer peptides Stacked deep learning NLP pre-trained embeddings CNN Feature fusion

mACPpred 2.0 is a stacked deep learning (SDL)-based predictor for anticancer peptides (ACPs), integrating the latest benchmark dataset with hybrid feature representations (top 7 NLP pre-trained embeddings + 90-dimensional probabilistic features). It employs 1D CNN blocks to extract spatial features, pioneering the joint modeling of spatial and probabilistic representations.

mAHTPred

Webserver
Antihypertensive peptides Meta-predictor Ensemble learning Feature selection

mAHTPred is a sequence-based meta-predictor using 6 ML algorithms (ERT, RF, etc.) and 51 feature descriptors, developing meta-models via two-step feature selection and ensemble learning to improve antihypertensive peptide prediction.

mAHTPred 

Tool
Antihypertensive peptides CNN SVM Ensemble learning

An ensemble deep learning model combining Convolutional Neural Network (CNN) and Support Vector Machine (SVM), using Amino Acid Composition (AAC) and g-gap Dipeptide Composition (DPC) for feature extraction, achieving 95% and 88.9% accuracy on benchmark and independent datasets for antihypertensive peptide prediction.

MA-PEP

Tool
Anticancer peptides Multi-attention mechanism Feature fusion Deep learning

MA-PEP is a prediction tool for anticancer peptides (ACPs) based on multiple attention mechanisms, integrating molecular-level chemical features and sequence information via feature enhancement and fusion to address limitations in multi-modal feature integration. The model enhances key feature representations through attention mechanisms, demonstrating superior prediction performance on multiple benchmark datasets. Visual analysis and case studies validate feature extraction reliability, providing a new approach for ACP exploration with code and datasets available online.