[gmx-users] dmso topology

[Markus Tusch]

Hi everybody,

I am currently doing MD simulations with Gromacs-4.0.2 (ffG45a3) using 
dmso as solvent.
While checking the topology dmso.itp, which is given in 
/gromacs-4.0.2/share/gromacs/top, I have noticed some inconsistencies 
compared to the default parameters for dmso given in ffG45a3bon.itp:

[ atoms ]     ;   nr    type          ...................
   1     SDMSO   .............
   2     CDMSO   .........
   3     ODMSO   .....
   4     CDMSO   ...

_
###dmso.itp:###_

[ bonds ]
;  ai    aj funct          b0       kb
   1     2     1     0.195    376560    ; kb from Methionine
   1     3     1     0.153    502080    ; kb from C=O bond  
   1     4     1     0.195    376560


_###according to ffG45a3bon.itp:###

_[ bonds ]
;  ai    aj funct          b0       kb
   1     2     1     0.195    4.9500e+06    ; gb_39
   1     3     1     0.153    8.0400e+06    ; gb_38
   1     4     1     0.195    4.9500e+06    ; gb_39


Obviously the corresponding values of kb differ by one order of 
magnitude (105 vs. 106). Is it possible that the dmso.itp is sort of 
out-dated? Does anyone hav any experience with this and can recommend a 
parameter set?

ffG45a3bon also provides some additional parameters for "virtual" bonds 
in the dmso molecule which are completely neglected in the existing 
dmso.itp (gb_44, gb_45). Since the tetrahedral surroundings of the 
sulfur atom are already fixed by means of the dihedral angle gi_2, are 
these additional bonds in your opinion really necessary? In some 
simulations they seem to cause some trouble with the lincs algorithm...

Well, it would be very kind if anyone could give me a little help with 
the problems depicted above.

Thanks in advance!

Kind regards
Markus