sAMPpred-GAT

sAMPpred-GAT predicts antimicrobial peptides (AMPs) by integrating predicted peptide structures with sequence and evolutionary data into graph-based models processed by a Graph Attention Network and full connection networks to improve AMP identification.

Key Features:

• Integration of Predicted Peptide Structures: Incorporates predicted peptide structures alongside sequence and evolutionary data for combined analysis.
• Graph-Based Modeling: Constructs graphs from integrated structural, sequence, and evolutionary information to capture relationships within peptide data.
• Graph Attention Network (GAT): Employs a Graph Attention Network to learn discriminative features from the constructed graphs.
• Full Connection Networks: Uses full connection networks as the output module to classify peptides as AMP or non-AMP.

Scientific Applications:

• AMP identification: Predicts antimicrobial peptides (AMPs) from peptide sequence and predicted structural data.
• Benchmarking and evaluation: Demonstrated performance across eight independent test datasets, outperforming state-of-the-art methods in Area Under the Curve (AUC).
• Innate immunity and peptide-function studies: Supports research into innate immunity therapies and investigations of peptide functions.

Methodology:

Combine predicted peptide structures with sequence and evolutionary data to form graphs, apply Graph Attention Networks for feature learning from these graphs, and use full connection networks to produce final AMP predictions.