Doubt about the nonbond_params section and redundancy in the top.file

User discussions
,
1

GROMACS version:2018
GROMACS modification: No

Dear members of the community,

I am a newbie user of GROMACS, so, please, forgive me if my questions sound obvious.
I would like to run a test MD simulation for a molecule of methyl-triphosphate, but I have a few doubts concerning the topology file I produced.

At the top of the file, I defined the [ default ] section, where I chose the Lennard-Jones potential with geometric combination rules for both epsilon and sigma and I selected the scaling option for the 1-4 interaction as well as the values of scaling factors.

[ defaults ]
; nbfunc comb-rule(both geometric) gen-pairs(1-4 scaling) fudgeLJ fudgeQQ
1 3 yes 0.5 0.5

Then, I defined all the " type sections" (atomtypes, bondtypes, angletypes, dihedraltypes), where all parameters and type functions are reported.

[ atomtypes ]
;name at.ID mass charge ptype sigma eps
opls_802 O802 15.9990 0.000 A 2.90000E-01 5.85760E-01
[…]

[ bondtypes ]
; i j func b0 kb
opls_800 opls_801 1 0.1090 284512.000
[…]

[ angletypes ]
; ai aj ak funct th0 cth
opls_810 opls_813 opls_814 1 108.230 836.800
[…]

[ dihedraltypes ]
; ai aj ak al funct c1 c2 c3 c4
opls_814 opls_813 opls_810 opls_809 5 0.0000 0.0000 -0.4184 0.0000
[…]

QUESTION 1:
Should I report explicitly also the [ nonbond_params ] section cointaining all LJ homo- and heteroatomic interaction parameters or are these automatically computed by GROMACS through the keywords I’ve already switched on in the [ default ] section ( I mean nbfunc = type 1 and comb-rule = type 3)?

Then, in the same file I specified the connectivity, so [ moleculetype ], [ atoms ], [ bonds ], [ angles ], [ dihedrals ] and [ pairs ] sections. In [ bonds ], [ angles ], [ dihedrals ] sections, I reported only the atom number of atoms involved.

[ bonds ]
2 1
2 5
[…]

[ angles ]
17 14 16
17 14 11
[…]

[ dihedrals ]
15 14 11 10
16 14 11 10
[…]

QUESTION 2:
Should I specify again all parameters and type functions explicitly in these sections ([ bonds ], [ angles ], [ dihedrals ])? I guess not since they have been already defined in the previous “type sections” ([bondtypes], [angletypes], [dihedraltypes]).

I hope I explained myself well. Please, let know if you have any suggestions.
I would be very grateful to anyone who would help.
Thank you very much in advance for your time and your support.

Best regards,
Emma Rossi

P.S. I am sorry for copying and pasting a few lines of the top.file in the chat, but I was not able to attach the document.

2

All LJ pairwise interactions are automatically computed according to the combination rules. The [nonbond_params] directive is for specifying parameters that will override the combination rule values.

You need to specify the function type, but not the explicit parameters. grompp references them from the force field files. It will only throw an error if it can’t find a match for the sequence of atom types listed.

2 Likes