[gmx-users] pdb2gmx angles and dihedrals

[Elton Carvalho]

Dear Gromacs users

I am currently preparing an RTP file to describe a polymer for use with MARTINI.

MARTINI does not include default bonded interactions, so all bonds,
angles and dihedrals must be explicitly defined in the topology file
or, in my case, in the RTP file.

The problem is that pdb2gmx generates at least one dihedral for each
bond and an angle for each pair of bonds sharing an atom, even it they
are not explicitly defined in the RTP. In that case, pdb2gmx will
include the triplet (or quadruplet) for the angle (or dihedral), but
with empty parameters, so that grompp looks for general parameters in
the forcefield ITP file.

With coarse-graned forcefields, it does not necessarily make sense to
have all angles and dihedrals, since, for example, rings are modelled
as triangles of beads (so that the angles are redundant with the bond
lengths) and trans alkyl chains are modelled as colinear beads (such
that a dihedral is undefined).

Is there a way to prevent pdb2gmx from creating these angles and
dihedrals, unless explicity defined in the RTP?


Also, the manual states the following regarding RTP files:

"The first field in the file is the [ bondedtypes ] field, which is
followed by four numbers, indicating
the interaction type for bonds, angles, dihedrals, and improper dihedrals. "

And includes this as an example:

[ bondedtypes ] ; mandatory
; bonds angles dihedrals impropers
1       1       1       2  ; mandatory


But the aminoacids.rtp file included in the opls-aa forcefield starts with:

[ bondedtypes ]
; Col 1: Type of bond
; Col 2: Type of angles
; Col 3: Type of proper dihedrals
; Col 4: Type of improper dihedrals
; Col 5: Generate all dihedrals if 1, only heavy atoms of 0.
; Col 6: Number of excluded neighbors for nonbonded interactions
; Col 7: Generate 1,4 interactions between pairs of hydrogens if 1
; Col 8: Remove propers over the same bond as an improper if it is 1
; bonds  angles  dihedrals  impropers all_dihedrals nrexcl HH14 RemoveDih
    1       1          3          1        1         3      1     0

Which is inconsistent with the manual. In particular, Col8 set to 1
seems to remove dihedrals over bonds with any dihedral, not only those
defined under the  [ impropers ] directive.

Could anyone clarify the role of these extra columns? I think this
also hints for a review in this section of the manual.


Thanks!
--
Elton Carvalho
Faculty of mathematics and natural sciences
University of Groningen