[gmx-users] Re: TI and heavy H

[reich at mpikg.mpg.de]

Thanks David,

giving me some hints about what questions I might want to find answers for
definetely sets me on the right track!

Cheers

Luther

>> I agree that there will be a d(Ekin)/dlambda term. It will actually
>> occur
>> independent of whether I shift mass from the oxygen to the dummy atom or
>> not as the oxygen will not have the nitrogen's mass in any case.
>> Still a dummy atom will affect phase space. I suppose I am not sure
>> about
>> how exactly it will do that (I don't think it is included in the
>> evaluation of the potential and kinetic energy of a molecule). This is
>> why
>> I don't understand whether to transfer some mass to the dummy atom or
>> not.
>> Do you have a clue about the effects of a dummy atom on momentum space?
>
> So far, all of the free energy calculations I've done have involved
> thermodynamic cycles where I disappear a ligand from a protein on one
> side and from water on the other side, so all of these effects cancel
> out. So I can't speak from experience. However, if I were doing
> transformations which involved computing the free energy of changing
> something with one number of atoms into something with a different
> number of atoms, I'd be concerned about these issues. I don't have any
> references, but I'm sure there's something in the literature on this.
> Some particular concerns might be (a) do the bonded terms left over
> for the dummy atom need to be corrected for? (b) Does the dummy atom
> perturb phase space in some other way? I suspect once you're able to
> answer these questions for the case where you don't have heavy
> hydrogens and you aren't shifting mass around, it will become more
> clear how you should handle your case.
>
> Sorry I can't be more helpful.
>
> David