Adding non-standard amino acid topologies to CHARMM36 using CHARMM-GUI

[areyesromero]

GROMACS version: 2021
GROMACS modification: Yes/No

Dear all,

I have the include topology of my non-natural amino acid “aa-lig.itp” generated with Ligand Reader & Modeler. My goal is to add them to the CHARMM36 forcefield that I will use for my simulation.
From what I understand in the manual (Adding a Residue to a Force Field — GROMACS 2021.2 documentation) I should do as follow:

1. Readjust the topologies obtained from the external generator, in this case, CHARMM-GUI, to the form required by the .rtp format.
   Question n1: I have uploaded the .mol2 format coordinate to get the topologies without energy minimization. Would a minimization (e.g. with MMFF94s) of the non-natural amino acid before generating the topology be appropriate?

Question n2: Some differences are pretty clear between .itp and .rtp format. For example, in the former file the atom block contains an indexed atom type (N1, C1, C2, C3…), 8 columns whilst the latter doesn’t (N, C, CA, C3…) and it has only 4 columns. Does anyone know of a script like itptortp.py or perl for streamlining this step?

2. No new atom types are introduced apart for one covalent bond between the oxygen of my residue and the first atom of my inhibitor. Step 4 of the manual says "If you require any new bonded types, add them to ffbonded.itp". Would this be my case then?

3. Add the residue to residuetypes.dat with the appropriate specification (Protein, DNA, Ion, etc). Where is this file located? I have searched for the name of this file in the local installation directory but it is not there. Any clues?

[jalemkul]

@areyesromero

Question n1: I have uploaded the .mol2 format coordinate to get the topologies without energy minimization. Would a minimization (e.g. with MMFF94s) of the non-natural amino acid before generating the topology be appropriate?

In a rigorous parametrization, one first obtains a QM-optimized geometry and compares the MM-optimized structure with it to identify any incorrect bonded terms, and the QM geometry becomes the reference for all subsequent calculations (e.g. dipole moment, water interactions).

@areyesromero

No new atom types are introduced apart for one covalent bond between the oxygen of my residue and the first atom of my inhibitor. Step 4 of the manual says "If you require any new bonded types, add them to ffbonded.itp ". Would this be my case then?

If you have a new type of bonded interaction, yes. Just because you don’t use any new atom types does not mean you aren’t connecting them in some new way.

@areyesromero

Add the residue to residuetypes.dat with the appropriate specification (Protein, DNA, Ion, etc). Where is this file located? I have searched for the name of this file in the local installation directory but it is not there. Any clues?

It is in $GMXLIB, which is /usr/local/gromacs/share/gromacs/top by default.