[gmx-users] Add gromacs forcefield w/ virtual site

[chris.neale at utoronto.ca]

Sorry Marcelo, that sounds like a job for an author ;) If nobody else  
chimes in, then I suggest that you do some testing and come back to  
the list with specific problems. As an analogy, you might look at how  
the virtual site is handled in tip4p in the absence of settle.

Chris.

-- original message --

Thank you Chris for your answer:

1) The molecule has no net charge because the virtual site in the center
of mass is a point charge twice the charge in the O atom.
2) Until now I've created 5 files but I don't know if I am doing the
right thing:

     *   forcefield.itp

#define _FF_OXY

[ defaults ]
; nbfunc        comb-rule         gen-pairs        fudgeLJ     fudgeQQ
1                    2                        no
1.0            1.0

#include "ffnonbonded.itp"
#include "ffbonded.itp"

     *    ffbonded.itp

[ bondtypes ]
; i    j       func
    O    O         5  (I've used function 5 because the forcefield has a
fixed length, is it right?) ;

     *    ffnonbonded.itp

[ atomtypes ]
; name  at.num        mass              charge   ptype
sigma          epsilon
       O                8       15.99940        -0.123       A
3.01300e-01    4.0780822e-01
     COM            0       0.000000         0.246       V
0.00000e+00    0.0000000e+00

     *   atomtypes.atp

O     15.99940  ; Oxygen Atom
COM   0.0       ; Virtual site (COM charge)

     *   residues.rtp

; Oxygen
[ OXY ]
   [ atoms ]
       O         O              -0.123    1
       COM   COM        0.246    1
   [ bonds ]
      O    O

3) Is it necessary more or less files? And in which file do I
specificate de virtual site?
4) The virtual site in the COM is [ virtual_sites2 ] with a = 0.5 or [
virtual_sitesn ]?