solvent excluded volume of large molecule with periodic boundary condition
As I suspect that the current gmx sasa command (source code would be nsc.c) does not give correct solvent excluded volume in a particular condition,
I would like to describe (and confirm) the issue and request the revision, if possible.
In the attached pdf file, I described the issue, but
the bottom line, I believe, is that, when solvent excluded volume of large molecules or those close to unit cell boundary is calculated with periodic boundary condition, there is no closed surface for the molecule volume envelope, and so the volume calculation based on Gauss theorem would not work. So if we add dots (or surface) to the simulation cell unit boundary that is occluded by atoms (with extra probe radius), we would get correct volume. (This correction should not be done for SASA calculation.)