ChemServerFree online molecular docking
Dock small molecules into protein targets with AutoDock Vina, Vina-GPU and GNINA, straight from your browser. ChemServer prepares the receptor and ligands, builds the grid box and runs the docking on its compute servers. You get ranked scores, poses and interaction views.
Docking engines
| Engine | Runs on | Good for |
|---|---|---|
| AutoDock Vina 1.2 | CPU | Standard docking and reproducing published Vina protocols |
| Vina-GPU 2.1 | GPU | Fast docking of larger ligand sets |
| GNINA 1.3 | GPU | Vina search with CNN rescoring, which often ranks the right pose higher |
| Vina-GPU with GNINA CNN | GPU | GPU speed for the search, CNN scores for every pose |
All engines read the same inputs, so you can dock a set with AutoDock Vina and repeat it with GNINA to compare.
This is the same AutoDock Vina you would run on the command line; download the config.txt of any grid
to reproduce a run on your own computer.
Three ways to define the binding site
- Reference ligand. Pick the co crystal ligand in the PDB file and the box is centred and sized on it. The best choice when the structure already has a ligand in the site you care about.
- Blind docking. The box covers the whole protein. Use it when the site is unknown, then dock again in the best region it finds.
- Custom grid. Enter the centre and size yourself, or apply a pocket found by pocket analysis. An interactive 3D grid explorer shows the protein, the ligands and the box before you start.
Preparation is built in
- Receptors are fetched from RCSB by PDB ID (several at once) or uploaded as PDB files.
- Optional automatic repair with PDBFixer adds missing atoms and hydrogens. See protein preparation and repair for full control.
- Buffers, solvents and other heteroatoms that do not take part in binding can be removed in one click.
- Receptors and ligands are written as PDBQT with polar hydrogens and Gasteiger charges by AutoDock Tools (MGLTools 1.5.7) or Meeko. Choose the one your reference papers used.
- Ligands can be SMILES, SDF, MOL, MOL2, PDB or PDBQT files, or ZIP archives of ready PDBQT files. SMILES are turned into 3D structures for you.
Settings you control
- Exhaustiveness, number of poses and energy range.
- Flexible side chains for residues in the pocket.
- Vina-GPU threads, set automatically for the GPU your run lands on.
- A separate chain and target ligand for every receptor in a batch.
Results
- Ranked score table for every receptor, best pose first.
- All poses as PDBQT and as PDB with correct bond orders, ready for PyMOL, Chimera or Discovery Studio.
- 3D view of each pose inside the pocket with its interactions.
- Complete docking logs and an Excel analysis workbook.
- Send the whole run to My Files to keep it, or download it as a ZIP.
Limits
A free account docks up to 50 ligand receptor pairs per run, one run at a time. Busy servers queue your run and show when it will start. Need more? Email chris@chemserver.online.
Related
- Virtual screening of ligand libraries and receptor panels
- How to do molecular docking, step by step
- Protein viewer for publication images of your best poses