Improper dihedral coefficients in oplsaa.xml

[HUANGZR86]

Dear developer,

I am using the default oplsaa.xml force field with customized atom types on my molecule "CCC[S+]1CCCC1". However, all the improper dihedrals cannot be parameterized.

In this case, will there be any problem with my MD simulation if these improper parameters are missing? Or how can I find and add those improper coefficients?

Can you give me some advice to deal with it?

Best wishes, Ziru

[HUANGZR86]

More specifically, is there any symbol that stands for all atom classes in foyer, e.g. class="*", similar to the atom class 0 in the Tinker OPLSAA.prm ?

[daico007]

Usually, we only expect improper parameters for ring structure, so for the molecule you have above, that may be necessary. Unfortunately, I don't think our current oplsaa.xml has any impropers in it (but you can add those in to your local xml file). And to present a wildcard (that can take any atom classes), you just leave it as an empty string, for example:

<Proper class1="CA" class2="CA" class3="N" class4="" c0="8.7864" c1="0.0" c2="-8.7864" c3="0.0" c4="0.0" c5="0.0"/>

Here, class4 is a wildcard.

[HUANGZR86]

Thank you very much for the reply. I am now adding the OPLSAA improper parameters from Gromacs ffbonded.itp to my xml file as:

` <Improper class1="C" class2="" class3="" class4="O*" c0="87.86400" c1="0.00000" c2="-87.86400" c3="0.00000" c4="0.00000" c5="0.00000"/> <Improper class1="C_2" class2="" class3="" class4="O*" c0="87.86400" c1="0.00000" c2="-87.86400" c3="0.00000" c4="0.00000" c5="0.00000"/> <Improper class1="C_3" class2="" class3="" class4="O*" c0="87.86400" c1="0.00000" c2="-87.86400" c3="0.00000" c4="0.00000" c5="0.00000"/>

`

I think the improper dihedral with ids (0, 1, 8, 9) and types ['opls_135', 'opls_136', 'opls_140', 'opls_140'], which is CT-CT-HC-HC, should be able to be parameterized. However, the error message of missing dihedral still appears. Is it because I am using the RB torsion so foyer processes those improper as propers?

[HUANGZR86]

<Improper class1="CT" class2="" class3="" class4="" c0="9.20480" c1="0.00000" c2="-9.20480" c3="0.00000" c4="0.00000" c5="0.00000"/>

[bc118]

RB torsions are regular dihedrals, not impropers, so you need to add those parameters. This is expected because it appears you are adding atom types to the XML file. Also, yes, the * is the wildcard symbol in the XML files.

[HUANGZR86]

Thank you for the reply. However, I tried the wildcards "", " ", and "*", but none of them can work.

<Improper class1="*" class2="*" class3="*" class4="*" c0="9.20480" c1="0.00000" c2="-9.20480" c3="0.00000" c4="0.00000" c5="0.00000"/> <Improper class1=" " class2=" " class3=" " class4=" " c0="9.20480" c1="0.00000" c2="-9.20480" c3="0.00000" c4="0.00000" c5="0.00000"/> <Improper class1="" class2="" class3="" class4="" c0="9.20480" c1="0.00000" c2="-9.20480" c3="0.00000" c4="0.00000" c5="0.00000"/>

[bc118]

I understand.
From my understanding, but I can not be sure for every writer and engine...., dihedral and improper wildcards are only accepted with wildcards in these locations.

• dihedral = "" - CT - CT - "" ---> ("star_wildcard" - CT - CT - "star_wildcard" )
• improper = CT - "" - "" - CT ---> (CT - "star_wildcard" - "star_wildcard" - CT)

If this does not work, it may be that the specific file writer (for GROMACS or LAMMPS, ...) is not supporting the "*" symbol as a wildcard. In this case, you can just add in all the combinations and it will work. like: CT-CT-NT-CT CT-CT-CT-NT etc