Questions about some LJ parameters in charmm36 ff

User discussions
1

GROMACS version:
GROMACS modification: Yes/No

hello.
I noticed something strange about the entries within the file “charmm36_ljpme-jul2022.ff/ffnonbonde.itp”, available on the MacKerell lab website

These are the ones from this contents down. (probably lines 541~597)
; from file toppar_c36_jul22/stream/misc/toppar_ions_won.str

The units for epsilon given in Table1, which is the reference for toppar_ions_won.str, are kJ/mol.

Why are the epsilon values in this file written as if they are kcal/mol converted to kJ/mol?

AG1P as an example,
When the epsilon value of AG1P provided in the reference is 0.2870 kJ/mol.
0.2870*4.184 = 1.200808.

; type atnum mass charge ptype sigma epsilon
AG1P 47 107.870000 0.000 A 0.216363662688 1.2008080 ; Silver

It’s the same as the file contents. But I think it should be 0.2870, not 1.200808.
I’m not sure which one is right. Maybe I’m just not thinking straight.

Thanks in advance for answers.

2

In that reference epsilon seems to be in kcal/mol whereas all energies are in kJ/mol.

3

I misread that, thank you.

4

I’m really sorry, but I can’t seem to find an explanation in the reference that the units for ε are kcal mol-1, could you point me to where? I assumed it was kJ mol-1 based on what was written at the bottom of Table1.

aCompiled data are from ref 10, where the hydration model depicts the ion as a sphere with the radius r that is surrounded by the hydration shell with the thickness Δr. ΔGhyd, ΔGcalcd, and ε are in kJ mol-1 and r, Δr, and Rmin are in pm.
bε is evaluated in eq 2 based on r and Δr.

5

I don’t see a unit for epsilon in table 1 (which is bad). But in the text epsilon has units kcal/mol.

6

Isn’t the unit of epsilon in kcal mol-1 only in the force field parameter of CHARMM (cgenff)…?
In gromacs, isn’t the unit of epsilon in kJ mol-1? Is charmm36_ljpme-jul2022.ff/ffnonbonde.itp specially using kcal/mol…?

The reference describes the following, it is the same as the value in Table 1.
The formula based on the hydration shell model yields ε values of 0.2007 kJ mol-1 for the sodium ion and 0.3600 kJ mol-1 for the potassium ion.

7

Yes, CHARMM uses kcal/mol, GROMACS kJ/mol.

The paper doesn’t mention the unit, but checking once more I expect epsilon should have unit kJ/mol.

Maybe @jalemkul can say more?

1 Like
8

The paper incorrectly states that ε is in kJ/mol. It is actually kcal/mol. The parameters in GROMACS are translated directly from the CHARMM toppar file provided by the author and I checked the energies (I had the same confusion some time ago when using those LJ parameters).

9

Thanks.

I don’t think the paper states the units of epsilon that are given in table 1. So not incorrect, but lacking (and confusing).

10

Okay. Thank you both.