Re: FEP on charged residue mutation

From: Sebastian Stolzenberg (
Date: Sat Nov 22 2008 - 11:34:10 CST

Hi Jerome,

thank you for your quick answer,
> What we've done in the past is to use a counterion that "vanishes" at
> the same time as the charged residue is mutated - this happens in both
> "legs" of the thermodynamic cycle, so that contributions from ion
> solvation cancel out. As you said, it might also be a good idea to
> turn the ion into a water molecule instead.
both "legs"... so you mean 4 runs?:
prot(Arg)+env.+1*H2O=> prot(Ala)+env.+1*H2O
|| ||
\/ \/
prot(Arg)+env.+1*SOD => prot(Ala)+env.+1*SOD

The size of my system forces me to use merely 1 run:

prot(Arg)+env. =>prot(Ala)+env.+SOD

I am just interested in a final structure, not any exact free energy values.
Can NAMD-FEP handle this in the same fashion as described for the other
examples in the tutorial?

Thank you again,

> Either way, I recommend applying a position restraint on the ion (and
> on the water molecule if applicable) to keep it somewhere in bulk
> solution from diffusing and interacting with other solutes in a way
> that would be difficult to fully sample. For an analysis of how ions
> can spell trouble (however unlikely this sounds, as there is no 'i' in
> 'trouble') see there:
> In addition, the positional restrain will keep the water and the ion
> on top of each other, which is what you want to accelerate convergence
> of the free energy for that part of the mutation.
> Best,
> Jerome
> On Fri, Nov 21, 2008 at 5:22 PM, Sebastian Stolzenberg
> <> wrote:
>> Dear All,
>> I would like to do FEP on an R2A mutation. The final state will have a total
>> net charge of -1e, if I do not counter-ionize with a sodium ion that I call
>> R2ASOD.
>> I see three options:
>> 1.) replace R2ASOD with a water molecule *prior* to FEP
>> 2.) replace R2ASOD with a water molecule *after* to FEP
>> 3.) *gradually* replace R2ASOD with a water molecule *during* FEP
>> Option 3.) seems most accurate to me, but I would need to subtract the
>> solvation energy R2ASOD to get the free energy change for the R2A mutation.
>> From the implementation point of view, is option 3.) really the best choice?
>> Are there any other possible options that I did not consider yet?
>> Thank you,
>> Best,
>> Sebastian

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:48:40 CST