From: Pavel Kostadinov (pkcre_at_abv.bg)
Date: Tue Oct 08 2019 - 08:17:46 CDT
Hi Joshua, That was indeed the problem. I figured it out a few hours after I posted the email, it had to be from the pdb file. Simple mistake but unless you know the pdb file is that column-sensitive, you wonder where the problem comes from. Thank you all for the responses, hopefully we have helped other people experiencing the same issue.
Greetings, Pavel.
>-------- Оригинално писмо --------
>От: "Vermaas, Joshua" Joshua.Vermaas_at_nrel.gov
>Относно: RE: namd-l: psfgen error protein-ligand
>До: "jribeiro_at_ks.uiuc.edu" , "pkcre_at_abv.bg" , "namd-l_at_ks.uiuc.edu"
>Изпратено на: 07.10.2019 18:03
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: , Pavel Kostadinov
>Изпратено на: 04.10.2019 18:29
\.abv-omExternalClass\ \\\.abv-omExternalClass\\\ \\\\\\\.abv-omExternalClass\\\\\\\ p\\\\\\\.MsoNormal { margin: 0in; margin-bottom: 1.0E-4pt; font-size: 11pt; font-family: 'Calibri', sans-serif; } \.abv-omExternalClass\ \\\.abv-omExternalClass\\\ \\\\\\\.abv-omExternalClass\\\\\\\ a\\\\3Alink { color: #0563c1; text-decoration: underline; } \.abv-omExternalClass\ \\\.abv-omExternalClass\\\ \\\\\\\.abv-omExternalClass\\\\\\\ a\\\\3Avisited { color: #954f72; text-decoration: underline; } \.abv-omExternalClass\ \\\.abv-omExternalClass\\\ \\\\\\\.abv-omExternalClass\\\\\\\ p\\\\\\\.msonormal0 { margin-right: 0in; margin-left: 0in; font-size: 11pt; font-family: 'Calibri', sans-serif; } \.abv-omExternalClass\ \\\.abv-omExternalClass\\\ \\\\\\\.abv-omExternalClass\\\\\\\ span\\\\\\\.EmailStyle18 { font-family: 'Calibri', sans-serif; color: windowtext; } \.abv-omExternalClass\ \\\.abv-omExternalClass\\\ \\\\\\\.abv-omExternalClass\\\\\\\ \\\\\\\.MsoChpDefault { font-size: 10pt; }
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: on behalf of Pavel Kostadinov
Reply-To: , Pavel Kostadinov
Date: Friday, October 4, 2019 at 2:46 AM
To:
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 : Thu Dec 31 2020 - 23:17:12 CST