[gmx-users] Re: pdb2gmx disulfide bond formation - missing bond, angle, & dihedral types with all GROMOS FFs (Tsjerk Wassenaar)

[Elizabeth Ploetz]

Hi Tsjerk,

Thank you for your advice!  grompp does not complain (with all three of 
the previously mentioned versions of GROMACS), but the processed topology, 
processed.top, also does not have the terms filled in, as I show below for the
bonds section:

[ bonds ]
;  ai    aj funct            c0            c1            c2            c3
...skipping...
   56    57     2    0.1830  5.6200e+06
   57   585     2
   58    59     2    0.1230  1.6600e+07
...
(This is also true for the angles and dihedrals.)

But, I ran some minimization and 100 ps of equilibration (with no position 
restraints) and viewed the output using Gromacs's ngmx trajectory viewer so 
that bonds would be drawn based upon the topology and not the proximity of atoms 
(what e.g., Rasmol does).  There does appear to be three disulfide bonds present.
I then used g_dist to calculate the distance between the two sulfur atoms 
in each of the disulfides during that short equilibration.  They all came out on 
average to be 0.2040 nm, which is exactly what the equilibrium bond length would
be if it were using the correct bond type. 

So, while it seems strange to me that they do not appear in the topology 
(preprocessed or processed), I am reassured that GROMACS really is using 
the disulfide information with the GROMOS FFs.  

I appreciate your help.

Cheers,

Elizabeth


Hi Elizabeth,

These missing terms are filled in automatically by grompp from the
bondtypes, angletypes and dihedraltypes definitions in the *bon.itp.
Unless grompp complains about missing terms, you'll be fine. You can
check whether everything is okay by writing out a processed topology
(grompp -pp) and looking up the bonds.angles/dihedrals there.

Cheers,

Tsjerk