Best practice for using newer AMBER force fields (ff19SB) in GROMACS?

User discussions
,
1

GROMACS version: 2024.4

Hi everyone,

What is currently the best way to use a newer AMBER force field such as ff19SB in GROMACS?

I’m aware of the approach where you convert the AMBER parameter files to GROMACS format using ParmEd. However, when I use ParmEd-converted ff19SB in my protein–ligand + water simulations, I observe some strange behavior:

  • a vacuum “cylinder” or cavity forms in the center of the simulation box, and

  • the MM/PBSA results show no correlation with equivalent simulations run in NAMD using the same force field.

I know that CHARMM-GUI can generate GROMACS-ready systems with AMBER force fields, including ff19SB, but I’m looking for a solution I can run locally, integrate into a pipeline, and avoid relying on a web interface.

Has anyone successfully and reliably used ff19SB in GROMACS with fully local tools? Any recommended workflows, conversion scripts, or known pitfalls with ParmEd I should be aware of?

Thanks in advance!

2

Vedran Miletić @vedran is currently working on the official ff19SB port for Gromacs, and the reasons things are not as straightforward as ParmEding the topology are explained e.g. in his post below:

I believe they should be advanced enough to have an experimental version running, but he’d have to tell you if it’s shareable at this point.

1 Like
3

Hello. As @milosz.wieczor suggested, the endeavor is non-trivial, but we are progressing. His work on Gromologist was very helpful to get started.

As the last step, GROMACS will get the converted force field files. We are still going through the validation (and making changes) internally on a larger set of molecules. Therefore, the current version of the force field is not yet public, but we hope it will be ready for release sometime soon.

1 Like