[gmx-users] Bug in pdb2gmx: terminal COOH in GROMOS force fields

[Thomas Schlesier]

Hi all,
for the COOH terminal group one of the angles (the one including two 
carbons and the oxygen which carries the hydrogen) does not get 
overwritten by aminoacids.c.tdb.

Ok, somewhat slower:
Used ASPH as a test molecule and chose a -COOH capping group.

 From aminoacids.rtp:

[ ASPH ]
  [ atoms ]
...
     C     C       0.450     3
     O     O      -0.450     3
...
  [ angles ]
;  ai    aj    ak   gromos type
...
    CA     C     O     ga_30
...

 From aminoacids.c.tdb:

[ COOH ]
[ replace ]
C       C       C       12.011  0.33
O       O       OA      15.9994 -0.288
OXT     OT      O       15.9994 -0.45
...
[ angles ]
O       C       OT      ga_33
C       O       HO      ga_12
CA      C       O       ga_19
CA      C       OT      ga_30

This mean that from [ASPH] the atom named O with the type O should 
exchange by an atom named O with the typ OA.
Additionally the angle for CA-C-O should change form ga_30 to ga_19

Result from pdb2gmx:

;       It was generated using program:
;       pdb2gmx - VERSION 4.6.5
;
;       Command line was:
;       pdb2gmx -f asp.pdb -o asp.gro -p asp.top -asp -ter
...
; Include forcefield parameters
#include "gromos53a5.ff/forcefield.itp"
...
[ atoms ]
...
    10          C      1   ASPH      C      4      0.658     12.011   ;
    11          O      1   ASPH     OT      4      -0.45    15.9994   ;
    12         OA      1   ASPH      O      4     -0.611    15.9994   ;
...
[ angles ]
...
     4    10    11     2    ga_30
     4    10    12     2    ga_30
...

First line from the posted [angle] is the right angle with the atom 
named OT.
In the second line we have still the angle from [ASPH] which doesn't get 
replaced by the angle from [COOH].

I tested this with GMX 4.6.5 and GMX 4.0.7 -> Think every version from 
GMX 4.X.Y series is affected.
Also tested it with the gromos53a5, gromos53a6, gromos43a1 force fields 
-> Think every gromos force field might be affected.

Greetings
Thomas