[gmx-users] g_energy -fee -f2
feenstra at chem.vu.nl
Fri Sep 19 09:24:00 CEST 2003
Xavier Periole wrote:
> One changes the topology according to the perturbation and generate a
> new file.tpr and rerun. That what I would do. But I guess that if you change
> the number of atoms (I don't know about the type though) the trajectory
> will no longer correspond to the topology defined in the tpr and the program
> will complain.
> At that point I think somebody with the experience of FEP calculation
> within gromacs would be of better use than my comments.
You're right, FEP doesn't work if you change the number of atoms. There
are possibilities involving changing atoms to a 'non-interacting' type
(aka 'dummy' atoms), but that involves corrections to the entropy that
I'm not familiar with.
If you don't need to add or remove atoms, it's fairly simple, as I've
already briefly explained in my other mail. Make two topologies, that
e.g. differ in charges, atom types, bond lengths, angles, dihedrals.
Run your simulation with the one (non-perturbed) .tpr, then do a re-run
with the non-perturbed and a rerun with the perturbed .tpr. Feed those
energy files to g_energy. You do need to re-run the non-perturbed also
due to slight differences in 'original' and 're-run' energies that are
large enough to show up in the free energy estimate...
For additional accuracy (and an estimate of the error), you could also
run the reverse: simulate the perturbed system and then re-run both and
calculate the backward FEP.
| | |
| _ _ ___,| K. Anton Feenstra |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|( | )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands |
| \_/ \_/ | | | Room P275 Tel: +31 20 44 47608 Fax: +31 20 44 47610 |
| | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/ |
| | "What If None Of Your Dreams Come True ?" (E. |
| | Costello) |
More information about the gromacs.org_gmx-users