[gmx-users] Regarding soft core parameter: Relative solvation free energy calculation

[bipin singh]

Hello All,

I am trying to calculate relative solvation free energy for p-Cresol and
p-Chlorophenol using Gromacs 4.6.3.

I am using equispaced lamda values viz. 0.1, 0.05, 0.1,...., 1.0

During one transformation (lambda 0.0 to 1.0) I am switching off the vdW
terms (Changed to DUM atoms) for the -CH3 and -Cl group in p-Cresol amd
p-Chlorophenol respectively.

Thus I am using a constant sc-alpha=0.5 at each lambda values (0.0 to 1.0)
for both the molecules.

But I am getting a large fluctuation in dh/dl values only at initial lambda
values i.e. from 0.0 to 0.5, and only in the case of p-Cresol (-CH3 is
being changed to DUM).

If I change the sc-alpha=0 for initial lamda ( i.e. from 0.0 to 0.5), I get
consistent values of dh/dl (i.e. less deviation during simulation).


Can anyone suggest the reason for this behaviour and how to select
appropriate values of sc-alpha based on transformation. And can we use
different sets of sc-alpha values for two molecules to compute their
relative solvation free energies.


############################################
I have used the following mdp settings:

sc-power                 = 1
sc-alpha                 = 0.5
sc-r-power               = 6
sc-coul                  = no


and


sc-power                 = 1
sc-alpha                 = 0.0
sc-r-power               = 6
sc-coul                  = no

###############################################





*--------------------Thanks and Regards,Bipin Singh*