distorted conformation

Hi,


I am running Marvin 5.1.2 under Linux. I was trying to generate conformers for a peptide-like ligand with a guanidinium side chain with the conformer plugin (command-line version). The input file is in mol2 format which was generated from an x-ray structure of a protein-ligand complex from the PDB using insightII manualy and it is correct. The guanidinium group is nice and planar in the input structure but it becomes distorted in the conformers because the conformer plugin changes all the bonds in the guanidinium group into single bonds. Is there a way to prevent the program from changing bond types? Also, this is just a minor issue, but the columns of the output mol2 files are not aligned. It looks bad and makes column-based editing, when necessary, a pain.





Another issue: all the conformers in the output mol2 file have identical molecule id-s following the @<TRIPOS>MOLECULE line. This is bad because the docking program for which these conformers serve as input will overwrite them. Is there a way to have the conformer plugin add conformer numbers to the ligand name? For example, if the input name is ligand, the conformers would be called ligand_1, ligand_2, etc.





Thanks,


Maria

Hi,

Maria Zavodszky wrote:

I am running Marvin 5.1.2 under Linux. I was trying to generate conformers for a peptide-like ligand with a guanidinium side chain with the conformer plugin (command-line version). The input file is in mol2 format which was generated from an x-ray structure of a protein-ligand complex from the PDB using insightII manualy and it is correct. The guanidinium group is nice and planar in the input structure but it becomes distorted in the conformers because the conformer plugin changes all the bonds in the guanidinium group into single bonds. Is there a way to prevent the program from changing bond types?

Could you send us the input structure in PDB and mol2 format? The plugin does not modify any bond order information in the input structures, so I think this information was missing from the PDB, or it was lost during conversion to mol2. We need the input files to be able to check it.

Quote:

Also, this is just a minor issue, but the columns of the output mol2 files are not aligned. It looks bad and makes column-based editing, when necessary, a pain.

We will check this issue. Could you send the output file which looks bad?

Quote:

Another issue: all the conformers in the output mol2 file have identical molecule id-s following the @<TRIPOS>MOLECULE line. This is bad because the docking program for which these conformers serve as input will overwrite them. Is there a way to have the conformer plugin add conformer numbers to the ligand name? For example, if the input name is ligand, the conformers would be called ligand_1, ligand_2, etc.

cxcalc adds a property field named _MOLCOUNT to each molecule, it can be used as molecule ID. E.g. to the conformers of the first input molecule the following filed is added, if the output format is SDF:

Code:

>  <_MOLCOUNT> 1

Unfortunately our mol2 export does not generate this field. Using SDF as output format could help. Let us know if this is not a suitable option for you.





Best regards,


Zsolt

Hi Zsolt,





Thanks for looking into this. I am attaching both the input mol2 and the output mol2 file of the molecule generated with cxcalc. When I use the pdb file cut out of the original protein-ligand complex and I feed it into cxcalc, I get the same distortion of the guanidinium group.





Regarding adding a conformer count: using sdf files is not an option for me. The docking program reads only mol2 format. Adding such an option to your mol2 export should not be hard to implement and it might be useful for


many users.





Maria

Hi Maria,





Do you use molconvert to generate the conformers? What is the exact command you execute?





Zsolt

The exact command I used:


cxcalc leconformer -f mol2 1a2c_ligand.mol2 > 1a2c_ligand_leconformer_from_mol2.mol2





Maria

Hi Maria,





apologies for not getting back sooner to you: the situation is fairly complex and we wanted to thoroughly investigate all aspects of it.





1. The primary concern you raised was the planarity of the guanidinium group.





I confirm that at present we do not properly handle this group, for various reasons the hybridisation of the C atom is not recognised as sp2.


In your case it is also a cation therefore all three bonds should be aromatic, precisely speaking, delocalised (according to Wikipedia: http://en.wikipedia.org/wiki/Guanidine). Based on this information one possible workaround could be to modify the bond types in your original mol2 input file: bonds 74, 76 and 79 should be changed from type 1 to ar. The portion of your file should look sg like this:

Code:

72   68   69 1     73   70   86 1     74   70   89 ar     75   70   71 1     76   72   89 ar     77   72   73 1     78   72   74 1     79   75   89 ar     80   75   76 1     81   75   77 1

This should solve the planarity problem.








2. mol2 output format





You are right, columns should be aligned in the output file, we fill fix that problem in the next major release (5.2) the latest.








3. Individual names per conformations





We understand that need and we agree that many other users may require unique names. We will implement some viable solution to support individual names for multiple conformers, though at present the actual way of doing it has not been decided.








4. PDB related problems





The distorted conformation has the same reason as in case of mol2.


The other problem, that the molecule falls to 4 fragments is related to a non-standard PDB file format. The hetero group should have the same residue id and residue name in order to consider it as one structure. At least, this is the PDB definition, though many programs take different approach and assign residue names to atoms in HETATM records just as in ATOM records in polymer chains.


We will extend the implementation of the PDB file reader to properly recognise such situations (most likely again in the next major release 5.2).





If you need a workaround for this problem, then you need to edit your PDB file, I'm afraid. Change all residue names and residue id's to the same value. (Please find attached one such file.)





I hope this help, please come back to the forum if there are any further problems.





Regards


Miklos