[BUG] about biosimspace HOT 10 CLOSED

Thanks for reporting. I can't see the attachment containing the ligands. Were you meant to attach, or are they available from the paper link?

from biosimspace.

Hi Lester, Sorry I forgot to attach the zip file...
issue.tar.gz

from biosimspace.

hi @jasmin-guven , can you also share the BSS script you used to generate a merged molecule so we can check the syntax used to generate atom mappings ?
It would be good to use https://github.com/OpenBioSim/biosimspace/issues for reporting future issues

from biosimspace.

I can't even parameterise the ligands using both GAFF or OpenFF, so can't perform the merge from the SDF files alone. That said, I've checked the mappings manually and all involving ligand11 include the P element in the mapping, e.g. mapping atom index 15 in ligand11 to atom index 16 in ligand5 and ligand4.

mapping_11_5
Out[23]:
{0: 0,
 1: 1,
 2: 2,
 3: 3,
 4: 4,
 5: 5,
 6: 6,
 7: 7,
 8: 8,
 9: 9,
 10: 11,
 11: 12,
 12: 13,
 13: 14,
 14: 15,
 15: 16,
 16: 18,
 17: 19,
 18: 22,
 19: 23,
 20: 24,
 21: 25,
 22: 26,
 23: 10}

mapping_11_4
Out[24]:
{0: 1,
 1: 2,
 2: 3,
 3: 4,
 4: 5,
 5: 6,
 6: 7,
 7: 8,
 8: 9,
 9: 10,
 10: 11,
 11: 12,
 12: 13,
 13: 14,
 14: 15,
 15: 16,
 16: 20,
 17: 21,
 18: 22,
 19: 23,
 20: 24,
 21: 25,
 22: 0,
 23: 26}

from biosimspace.

hi @jasmin-guven , can you also share the BSS script you used to generate a merged molecule so we can check the syntax used to generate atom mappings ? It would be good to use https://github.com/OpenBioSim/biosimspace/issues for reporting future issues

Hi @jmichel80,
Here's my code:
https://github.com/meyresearch/metalloenzymes/blob/main/scripts/prepafe.py

And sorry, will report these in the above repo in the future!

from biosimspace.

No problem. Could you share the equlibrated prm/rst files for the ligands as well so that I can try to merge. It may be that the mapping is fine with the original SDF file, but is failing for the topology loaded from AMBER format. We now have a direct RDKit conversion, which should make things like this more robust.

from biosimspace.

Hi, @lohedges
Here are equilibrated prm7 and rst7 files for ligand 11 and ligand 3, 4, 12 and 14.

ligand_3.zip
ligand_11.zip
ligand_4.zip
ligand_14.zip
ligand_12.zip
ligand_5.zip

Thanks!

from biosimspace.

Hi again. I've tested locally and with the latest BioSimSpace from OpenBioSim I get the same mapping using either the SDF input or the PR7/RST7 AMBER files, which doesn't have dummies for the phosphonate group. I would expect this to be the case since we now use the native Sire to RDKit conversion behind the scenes, which has some tools to recover stereochemistry.

Prior to this addition, conversion to RDKit format went via an intermediate file format, so information could be lost depending on the format used. If molecules were loaded from SDF, then we would try use that format as the intermediate in order to preserve stereochemistry. (Which assumes that the original SDF file was well formatted.) If it wasn't loaded from SDF, then we go via PDB format and hope that RDKit can recover missing information during sanitisation. In your script you are performing the merge using the ligands loaded from AMBER format files, not the original SDFs, so are going via PDB format. I can only assume that some of the ligands aren't being loaded into RDKit in an optimal way, hence the strange MCS mappings. (Perhaps the equilibrated coordinates are sub-optimal, or some stereochemistry is lost.)

I'd advise the following:

1. Try with the current BioSimSpace from OpenBioSim:

mamba create -n openbiosim -c conda-forge -c openbiosim/label/dev biosimspace ambertools gromacs
mamba activate openbiosim

(You can replace mamba with conda if you don't use it.)

2. Try changing your script to load the SDF ligands to perform the MCS mappings.

from biosimspace.

For reference, this is what I see using your AMBER files with the latest code:

The two phosphonate groups map onto each other. The only things that don't map are:

1. A hydrogen attached to atom index 10 in ligand11 becomes a dummy in the lambda = 1 end state.
2. Three hydrogens attached to atom index 10 in ligand5 become dummies in the lambda = 0 end state.

from biosimspace.

Hi Lester,

Sorry for my delayed response.

I tried equilibrating the systems again and now it seems I don't get weird mappings. I also installed openbiosimspace, but was met with other errors along my workflow (which I will try to raise issues for on the correct site, if they persist).

If I get this error again, I will probably try using the SDF files to set up the mappings, unless I get my openbiosimspace to work.

Thanks again for helping, I will close this issue now.

from biosimspace.