[gmx-developers] Enthalpy decomposition and solvent-solvent interaction energies (extracted from g_energy)

[David van der Spoel]

On 24/09/16 21:13, Tânia Firmino Guerra Guerreiro Cova wrote:
> Dear All,
>
> I’m trying to decompose the enthalpic profile of a set of noncovalent
> complexes, containing beta-cyclodextrin and small drugs, substituted
> with hydrophobic, hydrophilic and charged groups.
> The enthalpy change equals the potential energy difference with respect
> to the unbonded state between host and guest.
> The enthalpic profile of each system can be further decomposed into
> eight terms,
>
> DeltaH = DeltaH_host + DeltaH_guest + DeltaH_hostguest + DeltaH_hostsol+
> DeltaH_guestsol + DeltaH_hosthost + DeltaH_guestguest + DeltaH_solsol
>
> where the first two terms contain bonded interactions and the remaining
> terms are inter- and intramolecular nonbonded interactions.
>
> My question is related to the correct procedure for the simulation of
> the completely separate state between host and guest, and the correct
> calculation of the solvent-solvent interaction term (DeltaH_solsol).
> In a first approach, it was performed a simulation in which the atoms of
> host and guest were restrained at a fixed reference position (completely
> separated) in the same simulation box. As expected, there was a large
> energy penalty for applying position restraints, and the solvent-solvent
> interaction term was positive. It should be expected that the latter
> term becomes negative, upon complexation.
> In a second approach, two different simulations corresponding to the
> solvation of each single molecule (simulation 1 - host + solvent,
> simulation 2 - guest + solvent) were performed. The number of water
> molecules was the same in both simulations. From these simulations, the
> energies of interaction related to host-host, host-solvent (from
> simulation 1), guest-guest and guest-solvent (from simulation 2) were
> extracted. These include LJ-SR, Coul-SR and the decomposed terms of the
> reciprocal sum for electrostatic calculations using PME. However, two
> solvent-solvent contributions (LJ-SR_SOL-SOL, Coul.SR_ SOL-SOL) can also
> be extracted from each simulation. Can these terms simply be added? It
> seems to me unreasonable, given the doubling of water molecules ...
> What is the correct procedure for the simulation used as reference and
> corresponding to the completely separate state between host and guest?
>
> I have used a similar procedure reported in J. Chem. Theory Comput.,
> 2013, 9 (10), pp 4542–4551 (DOI: 10.1021/ct400404q), but it is not clear
> to me what to do in the separate state.
Did you also read the supporting information? It describes the procedure 
in some detail.
>
> Can you please help me to understand which procedure or kind of
> corrections should I use?
>
> Thank you in advance for every kind of help.
>
>
> Firmino
>
>
>
>
>
>
>


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se