Peptide Resources

DAC-AIPs is a deep learning model for anti-inflammatory peptides (AIPs) identification based on deep variational autoencoder (VAE) and contrastive learning, capturing sequence features via multi-hot encoding, constructing autoencoders with convolutional and linear layers for feature reconstruction, enhancing latent feature representation through variational inference, and optimizing classification performance with contrastive learning.

DAMPD (Dragon Antimicrobial Peptide Database) is an updated replacement for the ANTIMIC database, containing 1,232 manually curated antimicrobial peptides (AMPs). It supports queries by taxonomy, species, AMP family, citation, etc., and integrates tools like Blast, ClustalW, Hydrocalculator, SignalP, and an AMP predictor to facilitate biological analysis of AMPs for new anti-infective drug development.

DAMPD is an updated replacement for the ANTIMIC database, manually curating 1,232 antimicrobial peptides (AMPs) with integrated search functions based on taxonomy, species, AMP family, citations, and keywords (Advanced Search). It integrates sequence analysis tools (Blast, ClustalW, HMMER), Hydrocalculator, SignalP, AMP predictor, and other modules to support natural AMP template mining and new anti-infective drug development.

DeepACP is a BiLSTM-based tool for anticancer peptide (ACPs) prediction, systematically comparing convolutional, recurrent, and hybrid architectures to identify BiLSTM as the optimal solution. Leveraging only peptide sequence information, it achieves state-of-the-art accuracy, outperforming existing methods. Model interpretability is enhanced through visualization, supporting proteomics research and drug discovery.

Deep-AmPEP30 is a short antimicrobial peptide (≤30 aa) predictor using PseKRAAC reduced amino acid composition and convolutional neural network (CNN), achieving 77% accuracy, 85% AUC-ROC/PR on 188-sample dataset, superior to existing ML methods, supporting sequence prediction and genome screening.