[gmx-developers] OPLS-AA validation

[chris.neale at utoronto.ca]

Sounds great, and very important.

While version control of FFs is very important for methods sections,  
typos and other mistakes in ffXXX.itp files are also key. So if  
gromacs ever does get a set of files version controlled, then it would  
be useful to implement something like a checksum over a particular set  
of ffXXX.itp files. This would combine the idea of controlling the  
intended version and also the gromacs specific file.

Ideally this would be by direct comparison, but for coding convenience  
could also be by way of expecting a binary identical .tpr given a  
specified version of gromacs and an inclusive .gro file by way of  
something like a pangram (The quick brown fox jumps over the lazy  
dog). Or perhaps these proprietary Schrodinger programs you speak of  
are good at doing this? We use only free software.

BTW, I recall an error-fix in opls dihedrals (involving Leu, Arg, Trp,  
or some combination IIRC) for the switch from gmx 3.3 to 3.3.1 -- I  
can no longer find revision notes though for anything prior to gmx 4.  
This particular revision is not all that important, except as a way to  
highlight that this is an ongoing issue.

Thanks,
Chris.

Chris.

Quoting "Shirts, Michael (mrs5pt)" <mrs5pt at eservices.virginia.edu>:

> Random tidbit-  About a year ago, I validated the Gromacs OPLS-AA
> implementation versus OPLS-AA 2005 from Schrodinger, both in Macromodoel and
> Prime. Currently, this is the closest thing to versioning in OPLS-AA, since
> Jorgensen has generally avoided version numbers.
>
> There were a number of small differences in torsions and angles that were
> identified, which had mostly to do with atom typing, plus a few rounding
> errors.   I'm wondering if it would be appropriate to use these changes as
> the default OPLS-AA implementation?  I've got a lot of different data on the
> exact differences, but I'll wait to post all of it until I see if people are
> actually interested!
>
> Best,
> ~~~~~~~~~~~~
> Michael Shirts
> Assistant Professor
> Department of Chemical Engineering
> University of Virginia
> michael.shirts at virginia.edu
> (434)-243-1821
>
> --
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-developers-request at gromacs.org.
>