[gmx-users] How to avoid generating wrong dihedrals?(Tsjerk)

[Mark Abraham]

wuxiao wrote:
> Dear Tsjerk and other gmx users and developers,
>   I use OPLS-AA force field to model a dendrimer PAMAM. And the addition 
> of hydrogens is found correct. The rtp files were written for the core, 
> middle and end repeating units, respectively.  The both repeating units 
> cause the problem. When I delete all those wrong dihedrals in the top 
> file, the grompp and mdrun can run successfully. I check the top file. 
> It seems that other dihedrals are all correct. What to happen to this? 
> Could you give me further information? Any reply would be thanked very much.

I think it likely you have a malformed [ dihedrals ] section in your
.rtp file. pdb2gmx will generate dihedrals for each rotatable bond (see
manual section 5.6.1), so the only way you can get such a dihedral is to
have defined all three of those bonds, or specifically defined that
dihedral. Perhaps these bonds are being created erroneously by your
.hdb? You will have to look in detail at what you are doing with these.

Mark