[gmx-developers] Force Field validation

[Justin Lemkul]

On 6/10/14, 8:53 AM, Alexey Shvetsov wrote:
> В письме от 10 июня 2014 08:09:05 пользователь Justin Lemkul написал:
>> On 6/10/14, 2:35 AM, Alexey Shvetsov wrote:
>>> Hi all!
>>>
>>> Are there some common GROMACS workflow to validate forcefields and/or
>>> forcefields additions like glycam?
>>>
>>> I have converted GLYCAM forcefield to gromacs format. However it has large
>>> number of building blocks (~2.5k). GROMACS should have case sensetive
>>> building block names to deal with it
>>
>> The first thing I'd do is scan across the different molecules and do
>> single-point energy evaluations in Gromacs and a known implementation,
>> presumably Amber.  That should give a good indication that the parameters
>> have been ported over correctly.
>>
>> -Justin
>
> Thats actulay what i wanna do. However it seems like very hard to check all
> building blocks and its combinations =\
>

Sure, that's a few million possible combinations, but are all of those possible 
linkages unique?  I imagine some parameters have to be shared among the 
different building blocks and connections.  When we validated the CHARMM36 
ports, we didn't check every combination.  That's just not reasonable.  We 
selected representative residues from different classes, then cut back based on 
parameters that were shared - if the bonds/angles/dihedrals worked in one 
combination, there's no reason to test the same exact parameters again.

-Justin

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

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA 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
http://mackerell.umaryland.edu/~jalemkul

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