[gmx-users] OPLS-AA nonbonded force field format

[David van der Spoel]

On 01/02/17 02:50, Justin Lemkul wrote:
>
>
> On 1/31/17 8:16 PM, Shi Li wrote:
>> Dear GMX users,
>>
>> I am editing an OPLS-AA non-bonded itp file for my simulation system. I
>> noticed in the ffoplsaa in gromacs folder, the non-bonded force field has
>> following format:
>>
>> [ atomtypes ] ; full atom descriptions are available in ffoplsaa.atp ;
>> name
>> bond_type    mass    charge   ptype          sigma      epsilon
>> opls_001   C
>> 6      12.01100     0.500       A    3.75000e-01  4.39320e-01 ; SIG …
>>
>> However, the non-bonded format I find from the manual is like this:
>>
>> [ atomtypes ] ;name at.num mass charge ptype V(c6) W(c12) O 8 15.99940
>> 0.000
>> A 0.22617E-02 0.74158E-06 OM 8 15.99940 0.000 A 0.22617E-02 0.74158E-06
>>
>> The difference is in the manual, it doesn’t contain the first column,
>> which I
>> think is the atom name. When I built my own non bonded itp file, I try
>> make
>> my file looks like the first one. For example, if I have an atom name
>> S1, and
>> belongs to type “S”, my first line of the file looks like this: [
>> atomtypes
>> ] ; full atom descriptions are available in ffoplsaa.atp ; name
>> bond_type
>> mass    charge   ptype          sigma      epsilon ;   S1    S      16
>> 32.0600     0.266702   A  0.3550    1.046700
>>
>> However, this give me the error as followings: Fatal error: Atomtype S
>> not
>> found
>>
>> This could be fixed if I remove the S1 column and make my file format
>> likes
>> the one in the manual. So, my question is, what is the correct non-bonded
>> forcefield format for OPLS? If I don't specify the name of the atoms,
>> I think
>> they are specified in the topology file, under the [ atoms ], is that
>> correct? In this case, how should the first format work in the grompp?
>>
>
> Atom *names* never appear in the ff*bonded.itp files.  These are all
> atom *types* because names are irrelevant from the perspective of the
> force field. OPLS is a bit funny because it uses bonded and nonbonded
> types.  There are far fewer bonded types and parameters are shared among
> multiple nonbonded types.

OPLS does this because in many cases the only difference between 
atomtypes is the partial charge. Partial charge is a property of the 
atom type in other words, unlike most other force fields.


>
> So column 1 is the nonbonded type, and column 2 is the bonded type (you
> will note that there's nothing that says "opls_" in ffbonded.itp).
>
>> Another question about this non-bonded file is, I think I just need to
>> specify the atom types, instead of every single atom in my molecule,
>> is this
>> correct? As I tried to include all the atoms in the nonbonded file,
>> but it
>> will give me the warning of override the atoms. But the format in opls
>> seems
>> to have all the atoms listed. In this case, how should I edit the
>> non-bonded
>> file correctly?
>>
>
> Parameters need to be uniquely defined.  You don't need to have an entry
> for every atom in the structure, because likely some types are the same.
>
> -Justin
>


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se