Docking & Binding · Koes Lab / University of Pittsburgh
Gnina
Rescores protein–ligand docked poses using convolutional neural networks to improve virtual screening enrichment.
What Gnina is
Gnina is a molecular docking and pose rescoring tool developed by the Koes Lab at the University of Pittsburgh. It evaluates structural binding poses between small-molecule ligands and protein targets using a convolutional scoring function.
The tool is designed to improve candidate enrichment in virtual screening campaigns. Researchers routinely pair pose generation from AutoDock Vina with Gnina rescoring to identify promising chemical hits more reliably.
How it works
Gnina applies a convolutional scoring function directly to docked ligand poses against a receptor structure. For each evaluated pose, the model calculates specific predictive metrics, namely CNNscore and CNNaffinity.
Key facts
- Version
- 1.1
- Domain
- Docking
- Task
- Docking rescoring
- Input modalities
- Protein structure (PDB), Small molecule (SMILES)
- Output modalities
- Scores table (CNNscore, CNNaffinity)
- GPU memory
- 16 GB
What you can use it for
- Virtual screening enrichment
- Rescoring AutoDock Vina poses to enhance the ranking and identification of active compounds in screening libraries.
- Pose assessment
- Evaluating receptor–ligand binding conformations to produce CNNscore and CNNaffinity values for each pose.
- Structure-based ligand ranking
- Prioritising small-molecule SMILES candidates screened against target receptor PDB files.
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
Gnina
Strengths and limits
Good at
- Reliably improves hit enrichment when applied to docked poses generated by AutoDock Vina.
- Provides dual evaluation metrics consisting of CNNscore and CNNaffinity for every pose.
Watch out for
- Requires a dedicated 16 GB of GPU memory to execute.
- Input is restricted to PDB receptor structures and SMILES ligand formats.
Inputs and outputs
You provide
Receptor (PDB)
Ligand SMILES
Canonical SMILES for the small molecule.
You get back
CNNscore and CNNaffinity per pose.
Licence
Commercial use permitted · Apache-2.0
The Apache-2.0 licence allows use in commercial research and products.
Questions
Can I use Gnina commercially?
Commercial use permitted. The Apache-2.0 licence allows use in commercial research and products.
How do I run Gnina 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 16GB, typically in 8.0s, and a run costs about 42 credits. Visitors without an account get 1,000 credits a month.
What inputs does Gnina require?
Gnina requires a target protein receptor supplied in PDB format and one or more small-molecule ligands supplied as SMILES strings.
What does Gnina output?
The model outputs a table listing the CNNscore and CNNaffinity calculated for each docked pose.
How is Gnina typically used in computational docking?
It is commonly used as a rescoring step directly following pose generation with AutoDock Vina to improve screening enrichment.
What are the hardware requirements?
Gnina requires 16 GB of GPU memory.
Written from Gnina's own documentation. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.