Docking & Binding · MIT
DiffDock-PP
A diffusion-based model for docking protein–protein interfaces.
What DiffDock-PP is
DiffDock-PP is a protein–protein docking tool developed by researchers at MIT. It serves as the protein–protein counterpart to DiffDock, specifically designed for protein–protein interaction (PPI) discovery tasks where the ligand partner is itself a protein.
The model provides a computational method for assembling macromolecular complexes when the individual binding partners are known, offering an alternative to co-folding approaches when full co-folding is computationally prohibitive.
Key facts
- Version
- 1.0
- Task
- Protein–protein docking
- Modality
- Protein structure
- GPU memory
- 24 GB
What you can use it for
- Protein–protein complex prediction
- Predicting the quaternary interface and bound coordinates when given separate receptor and ligand protein structures.
- PPI-directed discovery
- Modelling protein interactions for discovery pipelines where candidates function as macromolecular protein ligands.
- Alternative to co-folding
- Generating docked structural assemblies of known interaction partners when joint de novo folding is computationally impractical.
Run it here
No installation and no GPU of your own: fill in the inputs and it runs on our servers. Results stay in your browser unless you sign in.
Not runnable yet
This model isn't available to run right now.Run this model
DiffDock-PP
Strengths and limits
Good at
- Generates all-atom coordinates for the docked complex in standard PDB format.
- Provides a ranked list of candidate poses accompanied by confidence scores.
- Enables complex assembly without the high computational cost of full co-folding.
Watch out for
- Requires 24 GB of GPU memory to run.
- Requires both receptor and ligand partners to be provided as pre-existing structures in PDB format.
Inputs and outputs
You provide
Receptor (PDB)
Ligand protein (PDB)
You get back
All-atom coordinates in PDB format, viewable and downloadable.
Ranked complexes with confidence.
Licence
Commercial use permitted · MIT
The MIT licence allows use in commercial research and products.
Questions
Can I use DiffDock-PP commercially?
Commercial use permitted. The MIT licence allows use in commercial research and products.
How do I run DiffDock-PP without installing anything?
Use the form under “Run it here” on this page: fill in the inputs and press Run model. It runs on our servers on 1× A100 40GB, typically in 1.2min, and a run costs about 350 credits. Visitors without an account get 1,000 credits a month.
What inputs does DiffDock-PP require?
It requires two protein structures in PDB format: one designated as the receptor and the other as the protein ligand.
What outputs does the model generate?
DiffDock-PP outputs all-atom structural coordinates in PDB format alongside a ranked table of predicted poses and confidence values.
When should this model be used instead of co-folding?
It is designed for cases where the interaction partners are already known or where co-folding methods are computationally too expensive.
Written from DiffDock-PP's own documentation. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.