[gmx-users] A strange phenomenon in the new version opls-aa force field

Erik Lindahl lindahl at csb.stanford.edu
Fri Oct 10 09:03:01 CEST 2003

Hi PeiQuan,

That might indeed be an effect of the order of the impropers - I think 
it only happens at the N-terminus.

The main reason why I only uploaded it to the contrib section is that I 
have changed my development version of pdb2gmx not to remove any 
dihedrals over bonds with impropers (which pdb2gmx used to do 
automatically), but I haven't tested it with the "vanilla" version of 

This should be fixed in an upcoming release. For now you could do one 

1. Change the order of impropers in the force field back to the setup 
in the distribution.
2. Hack pdb2gmx not to remove dihedrals when there is an improper over 
the same bond
3. Add the dihedral manually to topol.top

However, I only think it affects one dihedral at the N-terminus, so the 
effect isn't large...



In my version of

On Thursday, October 9, 2003, at 05:28 PM, PeiQuan Chen wrote:

> Dear gmx-users:
>   I have noted that a strange phenomenon in the new version opls-aa 
> force field, which Erik Lindahl updated
> on the August 06, 2003. The changes of new version parameters are 
> minor, but involve using 4.184 instead of
> 4.1868 for converting between kcal and kJ, and changing the atom order 
> for some improper torsions.
>   I have noted the changes of the atom order for some improper 
> torsions can affect the topol files generated
> by the pdb2gmx program. Comparing the new version and old one of the 
> force field files, I have noted that
> in the ffoplsaa.rtp file, Erik have correct the atom order for the 
> improper torsions. But when I use the
> pdb2gmx to generate topol files, the change on the improper torsions 
> can affect the topol of normal torsions
> section. The new version files generated less torsion numbers than the 
> old one. I think that torsion section
> in the old version are correct and the old one is not.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1922 bytes
Desc: not available
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20031010/25a98945/attachment.bin>

More information about the gromacs.org_gmx-users mailing list