Parameter error between GMX 2020.1 and GMX 5.1.1

User discussions
1

Hi guys,

I have built a simulation system on my local computer using the 2020.1 release of gromacs. It includes the parameterisation of a novel DNA residue. This has been working fine, but now I am trying to run it on our clusters I am getting some parameter errors when i try to prepare the .tpr with grompp.

The system was built using the charmm-gui to prepare the membrane, and a DNA structure inserted by hand.

I can prepare files with grompp on my local system, using gmx 2020.1, but out cluster has gmx 5.1.1 installed, and this throws up parameter errors when I try to grompp the system together, I can replicate the errors if I install 5.1.1 on my home machine.
The error I am recieveing during grompp is as so;

“WARNING 1 [file ffparams.itp, line 40043]:
Too few gb parameters for type C”

[…]

WARNING 12 [file ffparams.itp, line 40692]:
Too few gb parameters for type C
"

There being a selection of errors, 10 total, all for gb parameters for type C

Here in an excerpt from my ffparams.itp, with line numbers included;
40009 : C NH1 CT C NH1 1 24 24
[…]
40041 : -6.2103112 0.1058552 -1.7041432 -8.798952 -12.8247968 -14.3419152 -13.4457024 -9.0064784 -2.0936736 3.0601776
40042 : 2.9200136 3.1283768 4.420396 5.4446392 5.1182872 3.2819296 1.4945248 0.874456 -5.8969296 -8.7981152
40043 : -7.4713688 -5.0609664 -2.7585112 -0.6313656 7.8374688 6.7220144 -10.3185808 -14.3243424 -16.2138368 -15.8598704
40044 : -12.5674808 -6.1040376 -0.1075288 1.5317624 2.5392696 4.353452 5.8266384 6.1659608 4.9258232 2.6007744
40045 : 0.7564672 -6.399428 -11.6248256 -9.3696496 -2.81374 3.123356 7.1830912 7.3893624 1.8915864 -4.6028184\

There doesn’t seem to be any less parameters on line 40043 than on it’s immediate neighbours, and the system will grompp and run just fine in gmx 2020.1. I can also use the gromologist tool to check my topology & coordinates match and that does not throw up any forcefield errors when asked ,

I’m not well versed in what the gb parameters are, and I was hoping someone may be able to shed some light on the issue.
Thank you in advance,
Jonah

2

grompp is interpreting CMAP parameters as GB parameters, so something is broken in the topology, but as to why this is version-dependent, I can’t say, unless one version of your force field does not invoke GB parameters (thus grompp never reads that directive) while the other does.

3

" grompp is interpreting CMAP parameters as GB parameteres"

If you have the time, would you be able to elaborate on what this means / link me to where I can find out.

“unless one version of your force field does not invoke GB parameters”

Apologies if I am exposing my ignorance here, but why would different gromacs versions result in different forcefields, surely the forcefields are the same?

On my local machine I installed the charmm36 ffs in the gromacs FF directory, not a local directory. However I was under the impression that is is only called in pdb2gmx, and my topology only calls local .itps (produced through previous pdb2gmx calls & charmm-gui when building the system) when using grompp.

Thank you for the advice

4

The snippet you posted are CMAP parameters, not GB. Since grompp is complaining about GB parameters, that means something is broken. Please note that any force field file from CHARMM-GUI is designed to be 100% self-contained, and any modification to it is likely to break it. If you’re making changes “by hand” as you say above, it is quite likely that your topology will cease to function as you expect it to.

This I can’t explain, honestly. There is nothing that strikes me as version-specific here, so if something works on one machine but fails on the other, it implies that your input files are different in some way, maybe in the parent force field or something else. Without access to all of your files, I’m only going to be able to guess, likely unproductively.

5

" The snippet you posted are CMAP parameters, not GB. Since grompp is complaining about GB parameters, that means something is broken"

Ah gotcha, yes that is quite a serious problem.

“Please note that any force field file from CHARMM-GUI is designed to be 100% self-contained, and any modification to it is likely to break it. If you’re making changes “by hand” as you say above, it is quite likely that your topology will cease to function as you expect it to.”

Thank you for the heads up. I have done it this way as I have not had any success in importing the coordinates for my nanopore (DNA, not protein) into the charmm-gui pdb reader. It has some modified terminal residues, (a cholesterol group + short linker). I parameterised the modified residue as a ligand using the charm-gui, and included that in my topology, then built the required bonds in using a tool called gromologist, reccomended to me by a user of this forum.

As I could not get the structure be read by the charmm-gui pdb reader, I just read in a short DNA oligo on its own and built a membrane simulation system. Then I removed the oligo and added my structure, and combined the topologies using include statements.

"so if something works on one machine but fails on the other, it implies that your input files are different in some way, maybe in the parent force field or something else. "

Maybe I was not clear, if I downgrade gromacs from 2020.1 to 5.1.1 on my local machine I can reproduce the error, input files are identical. I can then reinstall 2020.1 one again and proceed through grompp with no errors.

Are there parent forcefiled files that are called in grompp apart from the local (as in within the same dir) .itps called in my topology?

Again, your time and patience is much appreciated Justin.

6

I honestly can’t explain this. The structure of a topology file doesn’t differ between versions, so I can’t see how you get an error in 5.1.1 and not in 2020.1.

7

Wild. Well thank you very much for the help, I probably wouldn’t have been able to identify that error as cmap parameters otherwise. I think I will rebuild the system and see if is still present.
All the best,
Jonah