[gmx-users] g_sas

[Mark Abraham]

Cheong Wee Loong, Daniel wrote:
> Thanks for the explanation.  It is much clearer now.  Although I
> still don't quite understand why we can't just use Protein A as the
> calculation group AND output group to find the SASA of protein A.

If A is complexed to B in solvent you might want the true SASA of 
protein A (since B is not solvent, A's surface with B is not part of the 
SASA), or maybe the SASA of A while pretending that B was solvent, or 
indeed something else.

>  I
> know it states that the calculation group should be all non-solvent
> atoms, but I guess I am just trying to understand why.  As in why do
> we need both the calculation and output groups in the first place?
> Why can't we just use the group of atoms that we are want to find the
> SASA for?

You can - just specify the same group twice. Your suggestion would save 
a negligible amount of effort for such simple usage, and leave no 
available solution for more complex usage.

Justin's example was a good one. The use of two different index groups 
allows the user total control over the form of the calculation. This is 
how software should work. As a programmer, you can't anticipate all 
possible usage patterns, and so making available a general and flexible 
user interface is usually best when most of your users will have 
suitable expertise to use it. At the same time, you shouldn't make life 
difficult for simple usage.

Mark