GROMACS version:2021.5
GROMACS modification: Yes/No
Here post your question
I was going prepare a simulation and modelling residues with virtual site interactions using the charmm36-jul2021.ff, but with pdb2gmx I have gotten the following error:
“Fatal error:
A dummy mass for type CT3 bonded to type C is required, but no virtual site
database (.vsd) files where found”
Indeed there is no *.vsd file present in the ff directory.
Furthermore, in the charmm36-feb2021.ff release the ffbonded.itp has a section [ constrainttypes ] under which constraints imposed for dummy atoms. This section is also missing from the jul2021 release.
I wonder, would it be safe to copy over the constrainttypes section for the feb2021 to jul2021 and also copy merged.vsd to jul2021.ff/aminoacid.vsd?
Sandor
Dear Sándor,
Virtual site information was omitted from the release you mentioned, but in principle you can copy it from the previous one. The same applies for the constraints. We plan to reintroduce these in an upcoming release of the force field.
The reason: virtual site data and the corresponding are not present in the upstream CHARMM force field distribution. @jalemkul, please correct me if I’m wrong.
Kind regards,
András
Hello Andras,
Yeah, I have thought that I can copy over, the file and constraints, but I wanted to make sure if it would be correct and, sort of, draw attention to it.
thanks for the quick response
Sandor
It should be safe to copy it but just be aware that this file is entirely a GROMACS construct and has nothing to do with the actual CHARMM force field, hence why it got left out of the new port, which was generated in a more automated fashion from the CHARMM force field files. We used to write one as a boilerplate entry from the charmm27.ff file, but we don’t test it in any way.
This a year later reply:
I followed Andras’s suggestions and have copied the merged.vsd file to the charmm36-jul2022 release, added the [ constrainttypes ] from the feb2021 rlease to the new ffbonded.itp.
Using gmx 2021.5, pdb2gmx and solvation went well but for the minimization for grompp I have gotten the following errors:
“virtual site MCH1 (Res ACE-23) has non-zero mass 7.5175 (state A) /
7.5175 (state B)
Check your topology.
…
virtual site MNZ1 (Res LYS-27) has non-zero mass 8.5155 (state A) /
8.5155 (state B)
Check your topology.”
Why dummy atoms MCH1, MNZ1,… got masses, while in the atomtypes.atp Vsite atoms have 0 mass definition?
Interestingly, when I use the feb2021 release of the charmm36m ff and the same version of gmx, the virtual site atoms still get the same high masses, but grompp and minimization and md go without errors.
I have checked back my older simulations using Vsites with CHARMM36 all-atom force field (November 2018) release and gmx 2016, obviously, had no errors either.
Hi Slovas,
Did you by any chance manage to solve the issue?
I am getting the same error about non-zero masses with the charmm36-jul2022 release.
As per my understanding, the dummy atoms here are supposed to have masses because the masses from the connected heavy atom and the hydrogens are distributed to the dummy particles. So the error statement does not make much sense to me.
Some hints would be highly appreciated.
Abhishek
No responses yet
I was able to setup VS on the C36 feb2021 version. Both jul2021 and jul2022 version are giving the same error. It is surprising because even standard protein residues are giving this error. Not clear, what changed in the ff files thereafter.
Abhishek
Reviving an old topic, since was unable to set up vsites with the jul2022 gromacs port of charmm36. I can’t find the feb2021 port Abhishek mentioned above.
@awacha @jalemkul : I know these ports take a huge amount of work, and we (or at least I) can’t thank you enough for doing them! I noticed (and I believe Justin told me in a previous thread) that the layout changed substantially between the 2020 and 2021 ports. This coincided with the disappearance of the .vsd file (merged.vsd). After copying merged.vsd from the jul2020 release to the jul2022 release, I’m getting errors related to vsite atom types not being known atomtypes – presumably not found in atomtype.atp? The file atomtype.atp was also clearly reorganized between 2020 and 2021 releases. Is it possible to add the vsite atomtypes from the old version of atomtypes to the new one, or would it be easier to simply use the jul2020 release if I want to use vsites?
Thank you!
@awacha @jalemkul : I was able to work out this error I was getting with pdb2gmx, after copying over merged.vsd from one of the earlier ports:
Opening force field file ./charmm36.ff/aminoacids.arn
Program: gmx pdb2gmx, version 2024
Source file: src/gromacs/gmxpreprocess/gen_vsite.cpp (line 610)Fatal error:
Dummy mass type (MNH3) not found in atom type database
Turns out, the end of atomtypes.atp looks like this in both jul2021 and jul2022 ports:
; Atom types for virtual sites MNH3 0.00000 ; Vsite (Rigid tetrahedral NH3 group)
MNH2 0.00000 ; Vsite (Rigid NH2 group)
MCH3 0.00000 ; Vsite (Rigid CH3 group)
MCH3S 0.00000 ; Vsite
MW 0.00000 ; Vsite
After uncommenting MNH3, pdb2gmx runs fine.
However, I’m now getting the following errors with genion:
ERROR 1 [file solvated.top, line 6454]:
No default Constr. No Co. typesERROR 2 [file solvated.top, line 6455]:
No default Constr. No Co. typesERROR 3 [file solvated.top, line 6456]:
No default Constr. No Co. types
One error for every line in the [ constraints ] section of solvated.top. Any idea what’s going on?
In the ffbonded.itp file you need to add [ constrainttypes ]
following the feb2021 release as example
Adding the [ constrainttypes ] section from charmm36-feb2021/ffbonded.itp to charmm36-jul2022/ffbonded.itp did address the “charmm36-feb2021/ffbonded.itp” above. However, I’m now getting a different error with genion (same one you quoted above):
virtual site MN1 (Res GLU-444) has non-zero mass 8.5155 (state A) /
8.5155 (state B)
Check your topology.
And indeed, all of the virtual sites in the topology file have non-zero mass, e.g.:
; residue 446 VAL rtp VAL q 0.0
44 NH1 446 VAL N 14 -0.47 15.015
45 H 446 VAL HN 14 0.31 0
46 CT1 446 VAL CA 14 0.07 13.019
47 HB1 446 VAL HA 14 0.09 0
48 CT1 446 VAL CB 15 -0.09 13.019
49 HA1 446 VAL HB 15 0.09 0
50 MCH3 446 VAL MCG1 16 0 7.5175
51 MCH3 446 VAL MCG2 16 0 7.5175
52 CT3 446 VAL CG1 16 -0.27 0
53 HA3 446 VAL HG11 16 0.09 0
54 HA3 446 VAL HG12 16 0.09 0
55 HA3 446 VAL HG13 16 0.09 0
56 MCH3 446 VAL MCG1 17 0 7.5175
57 MCH3 446 VAL MCG2 17 0 7.5175
58 CT3 446 VAL CG2 17 -0.27 0
59 HA3 446 VAL HG21 17 0.09 0
60 HA3 446 VAL HG22 17 0.09 0
61 HA3 446 VAL HG23 17 0.09 0
62 C 446 VAL C 18 0.51 12.011
63 O 446 VAL O 18 -0.51 15.9994 ; qtot 1
However, running the same process with charmm36-feb2021 produces the same topology file:
; residue 446 VAL rtp VAL q 0.0
44 NH1 446 VAL N 42 -0.47 15.015
45 H 446 VAL HN 43 0.31 0
46 CT1 446 VAL CA 44 0.07 13.019
47 HB1 446 VAL HA 45 0.09 0
48 CT1 446 VAL CB 46 -0.09 13.019
49 HA1 446 VAL HB 47 0.09 0
50 MCH3 446 VAL MCG1 48 0 7.5175
51 MCH3 446 VAL MCG2 48 0 7.5175
52 CT3 446 VAL CG1 48 -0.27 0
53 HA3 446 VAL HG11 49 0.09 0
54 HA3 446 VAL HG12 50 0.09 0
55 HA3 446 VAL HG13 51 0.09 0
56 MCH3 446 VAL MCG1 52 0 7.5175
57 MCH3 446 VAL MCG2 52 0 7.5175
58 CT3 446 VAL CG2 52 -0.27 0
59 HA3 446 VAL HG21 53 0.09 0
60 HA3 446 VAL HG22 54 0.09 0
61 HA3 446 VAL HG23 55 0.09 0
62 C 446 VAL C 56 0.51 12.011
63 O 446 VAL O 57 -0.51 15.999 ; qtot 1
and yet with that version of charmm36 genion does not produce any errors. Any idea why? Were you ever able to get this non-zero mass error resolved?
Bumping this thread in case anyone can help with the issue of the non-zero masses for virtual sites. As mentioned before, despite all virtual sites looking identical in the topology file when generated under the july 2022 and feb 2021 versions of charmm36 ports, running genion with the feb 2021 version does not produce non-zero mass errors, which running it with the july 2022 version does. Any ideas as to why?
Thank you!