Inconsistencies in converting TIP3P bond and angle constants from OPLS-AA to GROMACS

User discussions
,
1

GROMACS version:2018
GROMACS modification: No

Dear all,

I am writing the top file for a system made of a molecule in water solution. I am converting OPLS-AA parameters in GROMACS.

For both bond and angle interactions I’d like to use an harmonic potential (type 1)
(1/2) k (d-d0)^2, where d0 is the equilibrium distance (nm) or the equilibrium angle (degrees), k is the bond or angle constant (respectively in kJ/mol nm^2 and kJ/mol rad^2). I easily converted equilibrium distances and angles from OPLS-AA to GROMACS.

Concerning the bond constant, K (OPLS-AA) includes the1/2 term of the harmonic potential and it’s expressed in kcal/mol Å^2. Therefore, in order to get k, the bond constant in correct GROMACS units, I did

K_bond (OPLS-AA) * 2 * 100 * 4.184
(this was also suggested in a previous conversation in this forum The different bond parameters obtained from LigParGen and from oplsaa.ff in Gromacs)

Concerning the angle constant, K (OPLS-AA) includes the1/2 term of the harmonic potential and it’s expressed in kcal/mol rad^2. Therefore, it was sufficient to do

K_bond (OPLS-AA) * 2 * 4.184

to get the angles constant in correct GROMACS units.

These conversion rules work well for all my atoms, since I found full consistency with parameters obtained through LigParGen. Unfortunately, only parameters for TIP3P water (the original model by Jorgensen, 1983) differ. In fact, I got

[ bondtypes ]
; i j func b0 (nm) kb (kJ/mol nm^2)
opls_12 opls_13 1 0.0957 376560.000 ;OW-HW

and
[ angletypes ]
; ai aj ak funct th0 (degrees) cth(kJ/mol rad^2)
opls_13 opls_12 opls_13 1 104.520 460.240 ;HW-OW-HW

while the tip3.itp available in GROMACS (OPLS-AAM_for_Gromacs/tip3p.itp at master · leelasd/OPLS-AAM_for_Gromacs · GitHub) reads

[ bonds ]
; i j funct length force.c.
1 2 1 0.09572 502416.0
1 3 1 0.09572 502416.0

[ angles ]
; i j k funct angle force.c.
2 1 3 1 104.52 628.02

I’m struggeling with this problem for a while, but I cannot figure out what I am missing. I would be very grateful if anyone would help me.

Thank you very much in advance for your clarifications.

Best regards,
Emma Rossi

2

Hi!

I’m no expert on water models, but the original Jorgensen paper uses rigid water, so it does not contain any spring constants for bonds or angles. AFAIK, flexible TIP3P is quite exotic. Usually, SETTLE is used to fix the geometry, making spring constants for bonds and angles unnecessary.

1 Like