RE: psfgen error protein-ligand

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Mon Oct 07 2019 - 10:03:53 CDT

Hi Pavel,

PDB files are delimited by columns. I'm betting that you are changing the number of columns the atom names take up, and as a result the x, y, and z coordinate columns are misplaced, resulting in a docking pose offset relative to the rest of the molecule. What does your full psfgen script look like?

-Josh



On 2019-10-05 10:24:15-06:00 owner-namd-l_at_ks.uiuc.edu wrote:

Dear João,
Thank you for your reply. To me, the problem seems to come from the pdb file. This is how to correct docking pose looks like: https://photos.app.goo.gl/qbjvP1JNAvpyyM249
My pdb file looks like this:
...
ATOM 4634 C GLY A 584 18.684 9.185 49.432 1.00130.85 C
ATOM 4635 O GLY A 584 19.885 8.917 49.405 1.00125.62 O
TER
HETATM 4636 O UNK 0 14.107 -5.022 28.492 0.00 0.00 O
HETATM 4637 C UNK 0 12.656 -5.392 28.849 0.00 0.00 C
HETATM 4638 N UNK 0 11.204 -5.762 29.207 0.00 0.00 N
HETATM 4639 C UNK 0 10.910 -6.594 30.468 0.00 0.00 C
HETATM 4640 C UNK 0 11.184 -8.110 30.471 0.00 0.00 C
HETATM 4641 C UNK 0 11.647 -8.812 29.181 0.00 0.00 C
HETATM 4642 C UNK 0 13.133 -9.165 28.988 0.00 0.00 C
HETATM 4643 C UNK 0 13.780 -9.037 27.596 0.00 0.00 C
HETATM 4644 C UNK 0 14.996 -8.114 27.393 0.00 0.00 C
HETATM 4645 N UNK 0 16.316 -8.688 26.848 0.00 0.00 N
HETATM 4646 C UNK 0 17.653 -8.309 27.513 0.00 0.00 C
HETATM 4647 O UNK 0 18.990 -7.931 28.178 0.00 0.00 O
CONECT 4636 4637
CONECT 4637 4636 4638
CONECT 4638 4637 4639
CONECT 4639 4638 4640
CONECT 4640 4639 4641
CONECT 4641 4640 4642
CONECT 4642 4641 4643
CONECT 4643 4642 4644
CONECT 4644 4643 4645
CONECT 4645 4644 4646
CONECT 4646 4645 4647
CONECT 4647 4646
END

When I try to change the 3rd row with the HETATM names (in which case the psfgen seems to work) with the names as I see them in the .str file from CGenFF, the docking pose is lost. The wrong docking pose looks like this: https://photos.app.goo.gl/eWnGVMgt2xV4urA49
It seems rather illogical to me why this would happen, I`m not changing any of the numbering of the pdb file, just the atom names, and the docking pose is lost somehow (I can see it the moment I load the molecule into VMD). How do you explain this to yourself? Why by altering this column VMD does not recognize the right topology of the ligand?
Best regards,
Pavel.

>-------- Оригинално писмо --------
>От: Joao Ribeiro jribeiro_at_ks.uiuc.edu
>Относно: Re: namd-l: psfgen error protein-ligand
>До: <namd-l_at_ks.uiuc.edu>, Pavel Kostadinov <pkcre_at_abv.bg>
>Изпратено на: 04.10.2019 18:29


Dear Pavel,



The residue and atoms’ names of the pdb and topology file have to match. When renaming the atoms, please make sure that they correspond. But, what do you mean that the docking pose is lost? Psfgen have only two ways to get the coordinates, from the pdb that you loaded, or from the IC cards. If the names of the pdb match the topology, it should not be using the IC cards.

If the hydrogen atoms are not present in the pdb, and their coordinates are guessed, that message is printed. This guess is purely geometry-based.



Best,



João


From: <owner-namd-l_at_ks.uiuc.edu> on behalf of Pavel Kostadinov <pkcre_at_abv.bg>
Reply-To: <namd-l_at_ks.uiuc.edu>, Pavel Kostadinov <pkcre_at_abv.bg>
Date: Friday, October 4, 2019 at 2:46 AM
To: <namd-l_at_ks.uiuc.edu>
Subject: namd-l: psfgen error protein-ligand

Dear NAMD users,



I have been trying to run psfgen for a protein in question (albumin) and a ligand (diisocyanate) from a docking pose. I have parameterized my ligand with CGenFF and downloaded the CHARMM topology and parameter files required for the psfgen run. However, on running the plug-in I obtained a warning that psfgen failed to set coordinates for the atoms:



psfgen) Warning: failed to set coordinate for atom O UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom N UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom N UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom C UNK:0 XO1

psfgen) Warning: failed to set coordinate for atom O UNK:0 XO1



Then I decided to change the names of the heteroatoms in the .pdb file (from C,O,N to C1,O2,N1 - the way I find them in the .str file) since they might not be CHARMM appropriate. On doing that, however, the correct docking pose is lost, BUT!!! I managed to create .psf file for this wrong pose. Still, I receive warning that coordinates for hydrogen atoms, which are not included in the pdb, are poorly guessed.



psfgen) Warning: poorly guessed coordinate for atom H1 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H2 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H3 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H4 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H5 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H6 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H7 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H8 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H9 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H10 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H11 UNK:0 XO1

psfgen) Warning: poorly guessed coordinate for atom H12 UNK:0 XO1



I have tried many things: from writing separate .str files from the one I obtained from CGenFF to changing the residue names in the pdb file, but nothing seems to work. How to change the pdb file in such a way so I don't lose my docking pose and the CHARMM files (and NAMD) recognize the ligand?



Thank you in advance, I appreciate your willingness to help!



Greetings,

Pavel.

This archive was generated by hypermail 2.1.6 : Sat Dec 07 2019 - 23:20:52 CST