[gmx-users] a99SB-disp forcefield

[Joe Greener]

Dear Gromacs users,

I am looking to use the a99SB-disp forcefield from Robustelli et al. 
PNAS 2018 (https://www.pnas.org/content/115/21/E4758) in Gromacs 2019.3.

The details of the forcefield are in the paper supplementary material. I 
have carried out the following steps:
- Downloaded the Amber99SB*-ILDN-Q forcefield from 
http://www.gromacs.org/Downloads/User_contributions/Force_fields as 
ff99sb-star-ildn-q.tgz.
- Downloaded the TIP4P/2005 water model from the same site as ff03w.tgz 
and extracted the water files.
- Modified the water parameters to match a99SB-disp and created a 
pre-equilibrated water box.
- Modified the partial charges on certain residues as required.

Two steps remain that I could do with some advice on:

- The torsion parameters in the paper are given as phi_0, k_1, k_2, ..., 
k_6. Sometimes k_2 onwards are missing, and values generally range -1.0 
to 1.0. I have looked at the list of dihedral types in the Gromacs 
manual and can't find which type these correspond to, or how to convert 
the values to a valid dihedral format. I guess this format is the one 
used by Desmond.
- The a99SB-disp forcefield has a non-bonded LJ override between all 
backbone carbonyl oxygens and amide hydrogens. It seems I could have an 
entry something like this in the forcefield files:

[ nonbond_params ]
O     H    1    sigma_val   epsilon_val

Would this be sufficient to add this constraint?

Has anyone else tried to implement the a99SB-disp forcefield in Gromacs? 
If not I'm happy to make it available if I get it working.

Best,
Joe

Joe Greener
Research Associate, UCL
http://jgreener64.github.io/