[gmx-developers] Non Equilibrium alchemy implementation in gromacs

[David van der Spoel]

Den 2018-12-15 kl. 18:24, skrev Piero Procacci:
>> The main novelty of the paper in my opinion is the usage of
>> statistical efficiency to compare different methods. It could well be
>> that more efficient methods than MBAR can be devised but without
>> efficiency measure, how are we to judge which method is more
>> efficient?
> 
> You are definetely right. However, efficiency is meaningless without
> reproducibility. For the molecular set you have used (rigid and small
> molecules) reproducibility is not a problem whether one uses FEP or NE
> techniques.  However do you feel that your conclusions on
> computational efficiency for benzene or ethanol hydration energies
> apply unscathed to monsters such as the MCL1 inhibitor S63845 (
> https://www.apexbt.com/s63845.html )?
> Piero
I do not want to make predictions but I see no reason to expect that NE 
methods would be more efficient than equilibrium methods, but if someone 
  proves me wrong I won't shed a tear and indeed it would be good for 
the community.

In my, possibly subjective, opinion, NE methods have their main raison 
d'être in systems (mainly experimental) where equilibrium measurements 
are inherently impossible, e.g. kinetics in living cells.

Cheers,
-- 
David van der Spoel, Ph.D., Professor of Biology
Head of Department, Cell & Molecular Biology, Uppsala University.
Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
http://www.icm.uu.se