[gmx-users] Free energy in many conformation of molecular

[Mark Abraham]

mlyu wrote:
> Hello,
>=20
>   I want to find a best (or stable) conformation for one molecular in=20
> water.
>=20
> First, I built many conformations for my molecular. Then, I run MD to g=
et
>=20
> local minima. I want to know the free energy difference for all conform=
ation
>=20
> in molecular. If one conformation had lowest free energy difference, I=20
> think
>=20
> it is better than others.
>=20
>   Now I want to know the free energy difference between conformation N =
and
>=20
> conformation A.(N =3D A, B, C =A1K, etc). Compare the results.
>=20
>   I don=A1=A6t know how to do this work.

Your approach is upside-down. To evaluate free-energy differences with
explicit solvation you need to evaluate/estimate the density of states
over energy space. You can't do this with MD for some preconceived
configurations because you won't re-sample the same configurations in
accessible simulation times. See the paper below for a step in the right
direction here.

F A Hamprecht, C Peter, X Daura, W Thiel and W F van Gunsteren (2001).
A strategy for analysis of (molecular) equilibrium simulations:
Configuration space density estimation, clustering, and visualization.
In: Journal of Chemical Physics, 114(5):2079-2089.

Mark