ChemServerFix protein and receptor structures online
Crystal structures are rarely ready for docking or simulation. ChemServer repairs missing atoms and residues, adds hydrogens at the pH you choose and prepares the receptor file your docking program needs.
Why a protein needs fixing
- Side chain atoms that were not resolved in the electron density are missing.
- Flexible loops and termini are often absent, leaving gaps in the chain.
- Crystal structures have no hydrogens, and docking needs at least the polar ones.
- Selenomethionine (MSE) and other nonstandard residues confuse force fields and docking tools.
- Waters, buffers, salts and crystallisation additives sit in or near the binding site.
What the repair does
| Option | What it changes |
|---|---|
| Add missing atoms | All missing heavy atoms and hydrogens, heavy atoms only, or none |
| Target pH | Protonation of His, Asp, Glu, Lys and other titratable residues (pH 0 to 14) |
| Add missing residues | Builds short missing loops and gaps in the chain |
| Replace nonstandard residues | For example MSE to MET |
| Keep heterogens | Keeps ligands and cofactors, needed when the grid is centred on a pocket ligand |
The repair uses PDBFixer from the OpenMM project, the same tool many MD and docking pipelines use.
Ready for docking
In molecular docking the repaired protein is converted to PDBQT with polar hydrogens and Gasteiger charges by AutoDock Tools (MGLTools) or Meeko. You can also clean the receptor of heteroatoms that do not take part in binding, keep only the chains you need, and add flexible side chains.
Clean up by hand
The protein viewer deletes waters, ions, all heteroatoms or chosen ligands with a click and downloads the edited PDB file.
How to fix a protein on ChemServer
- Fetch the structure by PDB ID or upload your PDB file.
- Open Protein Prep and Repair and choose the options above.
- Run the repair, then open the result in the protein viewer to check it.
- Download the fixed PDB or use it directly for docking.
Good to know
- Short loops are rebuilt well; long missing domains are better modelled with homology modelling first.
- Check the protonation of histidines and catalytic residues in the binding site; pH matters there most.
- Remove waters unless one clearly bridges the ligand and the protein in your reference structure.