Peptide Resources

Ensemble-AHTPpred is a robust ensemble machine learning model integrating Random Forest (RF), Support Vector Machine (SVM), and eXtreme Gradient Boosting (XGB), combined with multi-dimensional features (physicochemical properties, amino acid composition, etc.) and logistic regression-based composite features for antihypertensive peptide identification.

EnsembleQS is a stacked generalization ensemble model with Gradient Boosting Machine (GBM)-based feature selection for high-accuracy prediction of quorum sensing peptides.

FFMAVP is a two-stage deep learning tool for antiviral peptide (AVPs) prediction and virus-target classification, leveraging fusion features and multitask learning for efficient functional annotation. The first stage uses a dual-channel neural network (amino acid descriptor branch + sequence feature branch) to discriminate AVP activity. The second stage constructs two multiclass tasks for 6 virus families (e.g., Coronaviridae, Herpesviridae) and 8 virus species (e.g., HIV, SARS-CoV), using a parameter-sharing multitask framework to capture task correlations. Innovatively, network parameters from the first stage initialize the second stage to enhance feature transfer. 

FIRM-AVP (Feature-Informed Reduced Machine Learning for Antiviral Peptide Prediction) is a machine learning model for antiviral peptide (AVPs) prediction, enhancing accuracy by focusing on physicochemical and structural features of amino acid sequences. The model first extracts multi-dimensional features (secondary structure, hydrophobicity, charge distribution), uses feature importance analysis to identify key discriminators, and confirms that secondary structure features (α-helix/β-sheet propensities) are most critical for AVP activity.

A predictive webserver for anti-inflammatory peptide (AIEs) classification, integrating tripeptide composition features with motif information to build SVM and Random Forest models. Analyzing experimentally validated epitopes from Immune Epitope Database reveals AIEs are enriched in Leu, Ser residues and hydrophobic/polar motifs.