PPI
Protein-Protein Interaction
Predict whether two proteins physically interact using sequence-derived protein language model embeddings.
hub Open PPI builder arrow_forwardSequence-based molecular interaction modeling
Deep-Interact Studio
Build, train, compare, and apply deep learning models that predict protein, drug, and RNA interactions from biological sequence data, with no code required.
- Three interaction builders
- Live training metrics
- Downloadable models and predictions
Choose an interaction model
Each builder turns your sequence pairs into embeddings, trains a configurable classifier, and gives you metrics, predictions, and reusable model files.
DTPI
Drug-Target Protein Interaction
Model potential binding between a compound and a protein target from SMILES and target sequence features.
medication Open DTPI builder arrow_forward
RPI
RNA-Protein Interaction
Detect RNA-protein associations by combining RNA language model features with protein embeddings.
genetics Open RPI builder arrow_forwardFrom molecular pairs to predictions
Task-specific encoders, configurable neural classifiers, asynchronous GPU jobs, model comparison, and batch inference in one research workflow.
-
1
Prepare Pairs
Upload a CSV of protein, compound or RNA pairs with 0/1 labels.
-
2
Build Model
Select embeddings, layer structure, dropout, activation, and training settings.
-
3
Train and Compare
Track loss, accuracy, precision, recall, F1, ROC-AUC, and PR-AUC across runs.
-
4
Run Inference
Apply trained models to new molecular pairs and export prediction tables.