[gmx-users] extra factor in PMF

Thomas Schlesier schlesi at uni-mainz.de
Fri Apr 20 11:36:14 CEST 2012


AFAIK this factor isn't included in g_wham. Think i tested this some 
years ago with an older GROMACS version.
If one looks into the 'gmx_wham.c' and searches for 'ln' one finds it 
only in the comments for the gaussian random numbers.
Greetings
Thomas


>
> Hi all
>
> According to some references
>
> 1) J. Chem. Phys, 128, 044106, (2008)
>
> 2) Comparison of Methods to Compute the Potential of Mean Force
>
> Daniel Trzesniak, Anna-Pitschna E. Kunz, Wilfred F. van Gunsteren Prof.
> Article first published online: 28 NOV 2006
> DOI: 10.1002/cphc.200600527
>
>
>
> There is an extra factor that has to be taken into account when
> calculating the PMF. This extra factor (2kTln(r)) results from the
> transformation from the 3N Cartesian coordinates to 3N internal
> coordinates. In the case of a reaction coordinate defined by a distance
> r between two species the Jacobian (r*2sin(theta)) of the transformation
> for 3D to  1D leads to an extra term in the PMF.
> Therefore the true PMF should be
>
> PMF(true) = PMF(g_wham) + 2kTln(r)
>
>
> This extra factor is only significant at lower values of the reaction
> coordinate. My question is; Does g_wham take this into account ?
>
> Cheers
>
> Gavin
>
>



More information about the gromacs.org_gmx-users mailing list