Glycan breaks after topoaa

HADDOCK
1

Hi,
I’m trying to dock a series of glycans to a glycan receptor. Some of these are chemically modified so I cannot use the glycam webserver to construct these and I have to use rdkit to construct them. My input glycan looks good, but after the topoaa step all of the pdbs generated split my trisaccharide into its single sugar units. Any idea what could be happening here?
Thank you!

2

Hi @marcosgly
When is this happening?
if your monosaccharides do not belong to the list of supported carbohydrates (HADDOCK Web Server - Settings), it’s better to treat the full glycan as a ligand (see an example here haddock3/examples/docking-protein-ligand/docking-protein-ligand-test.cfg at main · haddocking/haddock3 · GitHub), i.e., providing topology and parameters.

Best,
Marco

3

Hi Marco,
Thanks for the quick reply! The trisaccharide is already broken up in the output of topoaa, the very first step in my config file. I have pseudo set the glycan label for the modified monosaccharide unit to match the HADDOCK-accepted labels. The glycosidic bond between the other two,unmodified glycans is also broken up however. I thought the strength of this glycan approach was to better model the flexibility of glycans within haddock, not sure I can do the same elsewhere and then feed topology and parameter; or do you think that’s quite easily done?
Anyways, I have attached PDB files for the starting glycan (glycan_ready) and the topoaa output (glycan_ready_haddock).
glycan_ready.pdb (7.8 KB)
glycan_ready_haddock.pdb (7.5 KB)

Thank you!
Marcos

4

hi, I checked the PDB file..for HADDOCK it’s impossible to model such a molecule: although the residue names are part of the library of supported libraries, the atom names do not match, so HADDOCK cannot guess the linkages or calculate any energy..the naming should match the standard OPLS/CHARMM topology (see for example https://charmm-gui.org/?doc=open_toppar&filename=top_all36_carb.rtf or directly the HADDOCK topology haddock3/src/haddock/cns/toppar/carbohydrate.top at main · haddocking/haddock3 · GitHub)

To answer to your question, HADDOCK’s ability to model the flexibility of the glycan will not be influenced by whether you provide or not additional topology/parameters. Those are simply used within HADDOCK to calculate forces and energies for molecules that HADDOCK doesn’t recognize :slight_smile:

Best
Marco