[gmx-users] Clarity on TI free energy terms

[Hannes Loeffler]

On Thu, 2 Jun 2016 08:27:15 +0000
"Nash, Anthony" <a.nash at ucl.ac.uk> wrote:

> Dear Hannes,
> 
> Thanks for all the help yesterday, it helped. I, hopefully, have just
> the one final question.
> 
> I am still a little confused how Gromacs deals with the interactions
> (vdW & Coul) with the environment when a soft-core potential has been
> used to switch between molecules (typeA and typeB in the topology
> file). I understand how lambda can be used to 1) phase out the charge
> then 2) phase out the vdW interactions for a molecule that is
> disappearing (or reverse for appearing) to capture hydration
> energetics e.g., Justin¹s methane in water FE tutorial. 
> 
> However, in the case of a D2K (aspartic-acid to lysine)
> dual-topology

How is your setup dual-topolgy?  http://dx.doi.org/10.1002/jcc.23804
indicates to me that the pmx approach seems to go for as much single
topology as possible.  Your case seems to fall into this particular
category.


> how are the vdW and charges brought back into play for
> typeB, when the interactions of typeA have been coupled to lambda as
> per your example below? 

In a second setup as explained earlier, assuming you want to separate
vdW from Coulomb.  See attachment for an unrelated example system.