[gmx-users] LIE interaction energy calculations

[Bert de Groot]

Hi,

if your sampling is good enough, you can obtain a good estimate of 
the relative free energies by "counting", and forget about the MD 
energies. Eliminate the bias caused by the choice of starting
conformations, and directly calculate the free energies from the 
configurational densities (G=-kT ln (rho)) evaluated along properly 
chosen coordinates (either on a grid or using a nearest neighbors
approach). We did this for the Daura et al reversible peptide folding
simulations and got a useful estimate of the folding/unfolding free
energy landscape:

Bert L. de Groot, Xavier Daura, Alan E. Mark and Helmut Grubmüller; 
Essential dynamics of reversible peptide folding: Memory-free 
conformational dynamics governed by internal hydrogen bonds; 
J. Mol. Biol., 390 (1): 299-313 (2001) 

--

Bert


Anton Feenstra wrote:
> 
> Hi All,
> 
> I'm in the middle of performing LIE calculations, or rather LIE analysis
> of my simulations. The problem I have is that traditionally, LIE is
> performed on/with short simulations of ~100ps, with a substrate in a
> more or less confined binding pocked. Optionally, results from several
> simulations in different binding modes (orientations) can be combined.
> 
> I am facing substrates without confined binding, i.e. they have much
> room to move around in the active site and binding region, but do stick
> preferentially to some (several) locations. In that light, I did several
> very long (>5 x 10ns) simulations to get the best possible sampling.
> 
> My problem now is how best to combine these results. For the short
> simulations of more or less well-defined binding modes, the energies are
> averaged per simulation, and then averaged with their respective Boltzmann
> weights, so that high energy (and thus less likely) states weigh less.
> Doing that for my long simulations produces nonsense results, so I was
> thinking of taking already the boltzmann-weighing for the average over
> the simulations. The problem I have with that, is that there is of course
> already a Boltzmann factor involved due to the statistical sampling
> during the simulation, so it would seem that I boltzmann-weigh it twice.
> 
> Has any of you thought about this and come up with some ideas or
> suggestions...? Any input is much appreciated!
> 
> --
> 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 chem.vu.nl - www.chem.vu.nl/~feenstra/       |
> |             | "If You See Me Getting High, Knock Me Down"           |
> |             | (Red Hot Chili Peppers)                               |
> |_____________|_______________________________________________________|
> 
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.


____________________________________________________________________________
Dr. Bert de Groot

Max Planck Institute for Biophysical Chemistry
Theoretical molecular biophysics group
Am Fassberg 11 
37077 Goettingen, Germany

tel: +49-551-2011306, fax: +49-551-2011089

email: bgroot at gwdg.de
http://www.mpibpc.gwdg.de/abteilungen/071/bgroot
____________________________________________________________________________