[gmx-users] Convert CNS .top into a GROMACS .itp for a small ligand?

[Justin Lemkul]

On 8/19/13 6:45 AM, massimo sandal wrote:
> Hi Justin,
>
>
>>> I have a small ligand that apparently PRODRG gets wrong -it completely
>>> alters the structure, borks angles etc. making it not what we want. So I
>>> used xplo2d to obtain a suitable set of files that I used for docking
>>> (with
>>> HADDOCK).
>>>
>>>
>> Interesting.  The hallmark of the PRODRG development process was that
>> bonded geometries were very well maintained.  I have never seen such a
>> problem; the issues were always with the charges.
>>
>>
> Well, apparently it fails when it has to deal with some systems, even when
> turning energy minimization off. The ligand I'm talking about is
> costunolide ( http://en.wikipedia.org/wiki/Costunolide ). Strychnine is
> another ligand where it introduces artefacts -in the case of strychnine,
> turning energy minimization off is enough to bypass artefacts, but somehow
> not for costunolide.
>
> (Let's leave alone the fact that PRODRG ligands are not very well
> parametrized for MD, I'm aware of that, but we know it's generally good
> enough for our purposes).
>
>
>>
>>   Now, it seems to me that xplo2d does not output a Gromos ff -compatible
>>> .itp file as PRODRG does. However it produces a CNS .top file, which seems
>>> to contain most of the same information ( charges, dihedrals, masses etc.)
>>>
>>> Is there a way to convert between the two?
>>>
>>>
>> Yes, but it will involve using multiple files.  The CNS topology should be
>> reasonably straightforward to parse with, e.g. Perl, but you will have to
>> cross-reference bonded parameters from ffbonded.itp by atom type and
>> somehow assign charge groups (although that can be done manually after the
>> fact, if needed).
>>
>>
> Okay, so there is no utility that does it already, it seems? If it's matter
> of scripting, it's okay, I can do that, but I hoped there was someone who
> already had my problem. :)
> Can you elaborate a bit more / point to relevant documentation for .itp
> files?
>

An .itp is just a topology, so it's identical to a .top except that it does not 
have system-level directives and does not #include a parent force field.  Manual 
sections 5.7.1 and 5.7.2 describe the relevant contents.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441

==================================================