[gmx-users] question about using new potential in groamcs

Mark Abraham Mark.Abraham at anu.edu.au
Tue Sep 15 17:58:02 CEST 2009


青 叶 wrote:
> Hello everybody:
>        I have read GROMACS' manual and found that the non-bond interaction in GROAMCS is describled as Coulomb term add A/r^12 -B/r^6 which is named Lennard-Jones term, but recently I have read some papers which use a effective-two-body potential to describle the non-bond interaction in the GROAMCS MD simulation, and this potential works much better in dealing with transition metal ion-water system, in this potential, non-bond interaction between ion and O atom in water molecular is describled as follow:
>       V(r)=A/r^4+B/r^6+C/r^8+D/r^12+E*exp(-F*r)+coulomb term; 
> where the A,B,C,D,E and F are parameters which can be obtained by fitting the potential energy surface calculated from Ab initio method. r is the distance between the ion and O atom in water molecular, the interaction between the ion and H atom in water molecular has the similar form.
>        So my question is how to modify the GROAMCS in order to simulate dynamic course by using this new potential instead of the default L-J interaction. I have tried to find the solution in internet, but so far I still have no idea on this problem, can anybody help me on this problem, I will be very, very grateful for your help.

GROMACS allows the user to supply a tabulated function in r for either 
bonded or non-bonded interactions. See the manual for details and 
$GMXLIB/gromacs/top/*xvg for examples here. Assuming you can cast the 
ion-O interaction and the ion-H interaction in the same functional form 
your problem might be simple. At worst, you might need to look up 
different tables for ion-O and ion-H interactions, and this would 
probably need a code modification.

Mark



More information about the gromacs.org_gmx-users mailing list