[gmx-users] charge during the free energy calculation

[Binbin Liu]

Dear all,

I have never done free energy calculation before, so bare me with my question 
may sounds silly. I try to calculate the free energy difference of binding an 
ion to the wild-type and the mutant type of membrane protein. 

The problem is
1) The  binding process is performed by decreasing/increasing the charge on 
the binding ion. This will cause the net charge of the system not to be 
neutral. Persumbely  this will affect the coulomb potential energy of the 
system. Is it very seriously?  How could this problem be dealt with ?

2)Similar problem about the charge too. The charge on HIS is changed to mimic 
the mutant-type membrane protein. How could the net charge be dealt with?

So besides the above charge issue, I believe Gromacs can perform both TI and 
SG by switching on free energy calculation option in .mdp file. Also g_lie 
makes an estimation to the free energy. Because my simulation system is big 
(4,000 atoms of protein, 56,000 atoms of sol, and 9,500 atoms of membrane 
plus ions), it will be quite time-consuming (~4 days for one 300ps 
simulation)  if TI methodology is performed. For example, 300ps equilibration 
plus 300 ps lambda simulation for about 15-25 intermediate lambda values. Is 
that really worthy doing? How could I know the correctness of my result 
without known experimental calculation?
  
Thanks in advance.

Binbin