[gmx-users] Questions about GB parameters

[Mark Abraham]

On 15/06/2011 4:24 AM, Justin A. Lemkul wrote:
>
> Hi All,
>
> I wanted to dig up an old discussion that hit the list a long time ago 
> because I'm now encountering some problems understanding the GB 
> settings myself.  The discussion in question is here:
>
> http://lists.gromacs.org/pipermail/gmx-users/2010-August/053373.html
>
> I wanted to post a couple of questions based on Per's response.
>
> 1. Based on that post, it seems to me that the value in the gbr column 
> should have a dielectric offset added to it during the GB 
> calculations.  In the code, though (genborn.c, around line 484 in the 
> latest release-4-5-patches), it looks like the dielectric offset is 
> subtracted, not added.  I guess the code is reversing the process, 
> going from GB radius back to vdW radius by subtracting the dielectric 
> offset?  It seems, then, that the parameters in gbsa.itp should 
> specify GB radii, not vdW radii, though the manual says the gbr column 
> is "atomic van der Waals radii, which are used in computing the Born 
> radii."  Is the opposite actually true?

It does look to me as though something is mis-sense here. The quantity 
given in gb_dielectric_offset is always subtracted from a value that I 
think is found in the gbr column of [implicit_genborn_parameters].

Mark

> 2. I am still unclear on the source of the HCT scaling factors.  From 
> the reference cited in the manual, it would seem that scaling factors 
> are interaction-dependent, at least when H atoms are concerned.  I 
> also cannot find any indication of where these values came from.  None 
> of the values of Table 2 in the HCT reference match the contents of 
> the "hct" column.  Again I wonder if I'm missing something obvious :)
>
> Thanks for any insight!
>
> -Justin
>