[gmx-users] PMF, Enthalpy, Entropy

[David van der Spoel]

On 14/03/17 13:23, Justin Lemkul wrote:
>
>
> On 3/13/17 2:11 PM, Alex wrote:
>> Hello all,
>>
>> I have the calculated the PMF for the adsorption of a peptide into a
>> solid
>> surface using umbrella sampling, a 60 ns of NVT-production for each
>> window
>> was considered so that the PMF is well converged.
>> Now, following a nice paper by Prof. van der Spoel "*DOI:
>> *10.1021/ct400404q"
>> and in the same fashion as (Fig.4 of the paper) I also would like to
>> calculate the enthalpy and entropy along the reaction coordinate of
>> the PMF.
>>
>> To do so, for each window I calculated the enthalpy via "gmx energy
>> case.edr ... " and selected "potential" as in the NVT ensemble, enthalpy
>> and potential are identical. And then using the Gibbs free energy
>> equationI
>> calculated the "- [T*\Delat S]", so that at the end I have all of three
>> quantity as a function of the reaction coordinate.
>>
>> I do not know why the PMF and Enthalpy (potential) are not in the same(or
>> close) order of magnitude.  The trusted PMF is around "60 kcal/mol=251.0
>> kj/mol" while the Enthalpy is something like "-1557.92e+03 kj/mol", I was
>> wondering if the unit of enthalpy (potential) here is really meaningful?
>> Does the enthalpy's unit is "kj/mol" or "j/mol", as a division to "1000"
>> could give more sense to it?
>>
>> I know the convergence of the enthalpy is hard respect to the convergence
>> of the PMF, but, here the convergence is not the matter, rather the
>> unit of
>> enthalpy is the matter.
>>
>> Here is the potential average of enthalpy (potential) for one of the
>> window
>> via gmx energy:
>>
>> Energy                      Average   Err.Est.       RMSD  Tot-Drift
>> ------------------------------------------------------------------------------------
>>
>> Potential                -1.5579e+06         14    726.909    -14.203
>> (kJ/mol)
>>
>> Which of the "Err,Est" or "Tot-Drift" is the error bar for enthalpy
>> (potential)?
>>
>> And finally, what does the "normalization to zero at infinite distance"
>> means?
>>
>
> Because if you're just calculating the enthalpy of a system, that
> includes (presumably) a ton of solvent-solvent interaction energy that
> doesn't affect the binding free energy.  So the enthalpy of the window
> in which the binding species are at the maximum distance is set to zero,
> e.g. the reference of the unbound state.  Then all subsequent enthalpy
> values are offset by the same amount so you have DeltaH and not just H.
>
> If I've messed that up, maybe David will correct me; I read that paper
> but it's been a while so I may be wrong on teh details.
>

This is correct and what we have done too. The main problem is that it 
is harder to have the enthalpy converge than the free energy, that is it 
requires more sampling.

> -Justin
>
>


-- 
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