Sequence-based Ensemble Prediction with Attention Learning
for Protein-Protein Interactions
Explore predicted protein-protein interactions across multiple species with residue-level attention interpretability.
Select a species to browse its protein-protein interaction predictions.
SEPAL-PPI is a deep learning framework that predicts protein-protein interactions using protein sequence information. It employs an ensemble of specialized models with attention mechanisms to provide both accurate predictions and interpretable results at the residue level.
No structural data required β predictions are made directly from protein sequences using ESM-2 embeddings.
Combines multiple specialized models for robust and accurate interaction predictions.
Residue-level attention weights reveal which parts of each protein contribute to the predicted interaction.
Pre-computed predictions across multiple plant and model organisms, ready for exploration.