[gmx-users] Re: Some additions to the free energy question

[Anton Feenstra]

Markus O Kaukonen wrote:

> Dear All,
> 
> Is it really a good idea to calculate free energy
> differences of the two 'inhibitors' docking from
> vacuum to protein (see the original question below, manual fig 3.9b)
> when these two inhibitors differ each other by 'only' one proton by
> 'mdrun -rerun' method, that is calculating energies of one trajectory
> using the topology of the other one.
> 
> In the case when the difference of I and I' is one proton it is not
> just the proton but many structural rearrangements of aminoacids
> (in E and I/I') and re-arrangement of the water network (in E).
> The energy gain in the rearrangement can easily be 100kJ/mol.
> 
> So does one still think that the 'mdrun -rerun' is a good idea here
> or should one consider the "slow growth" method described in the
> manual in sec 3.12. Or is the difference between IE and I'E
> too big in this case and neither of methods does work?

IMHO, this is a quite straightforward example where FEP (perturbation)
is appliccable. See recent papers of Van Gunsteren on this as well.

The issue here is not how extensive the rearrangements in the enzyme
can be due to the I->I' change, but how much overlap there is in
energetically relevent (in terms of contributing to the free energy)
conformations. The basic assumption made here is in fact that sampling
in one state (say, EI) will produce sufficiently relevant conformations
for the other state (EI') as well, so that only the EI' energies need
be re-evaluated (or, actually only the I'-I energy differences).

I've just been on the MGMS conference in Manchester this week, and seen
it applied by several groups/people. One can even use it in a sort of
QM/MM combined approach, where sampling is done in regular (MM) MD
simulation, and afterwards energies are re-evaluated for a small part
of the system by QM (e.g. cavity & ligand) on snapshots (e.g. every ps)
from the trajectory to get an accurate, QM based, free energy!


-- 
Groetjes,

Anton
  _____________ _______________________________________________________
|             |                                                       |
|  _   _  ___,| K. Anton Feenstra                                     |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|(   |   )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands   |
| \_/ \_/ | | | Tel: +31 20 44 47608 - Fax: +31 20 44 47610           |
|             | Feenstra at few.vu.nl - www.few.vu.nl/~feenstra/         |
|             | "If You See Me Getting High, Knock Me Down" (RHCP)    |
|_____________|_______________________________________________________|