[gmx-users] dihedral parameters in oplsaa

[Mark Abraham]

friendli wrote:
> Hi gmx users,
> 
> In order to understand the oplsaa all atom force field, I use pdb2gmx to 
> generate topology file of a pdb which contains only one AA( the glycine).
> 
> After the EM simulation, I use gmxdump to look for the details of the 
> parameters.
> 
> I find,
> in .top file, there are 15 proper dihedrals under directive [dihedrals] 
> and the funtion type is 3 (i.e. RB dihedrals).
> However, in the output generated by gmxdump, there are explicit 
> parameters for only 9 of these dihedrals.  The 9 dihedrals corresponds 
> to 9 of the 15 in .top correctly, but the other 6 are missing.
> 
> In the six missed dihedrals,  two are for N-CA-C-O, four are for  
> HA-CA-C-O.
> 
> Can anybody give a hint on this? Is there any default value not output 
> by gmxdump? All other parameters are fully printed.
> 
> 
>     H   H  O
>      |    |    | H- N - C - C = O
>      |
>     H

Well done for using your initiative to see how things work :-)

As David said, each unique function is only stored once. Later in the 
gmxdump there's another field that has an entry for each atom number 
involved in an actual dihedral, and an index into this table of 9 
dihedral functions to actually look up the parameters. Also, do be aware 
that gmxdump numbers beginning from 0, but your .top begins from 1.

Mark