[gmx-users] Calculations of area per lipid using g_sas

[Hwankyu Lee]

Dear GMX users,

I'm interested in calculating non-projected area/lipid of huge bilayer 
with undulations. I think that if the bigger solvent probes were used, 
the non-projected area per lipid may be reasonably calculated from the 
solvent accessible surface area because solvent probes cannot be 
inserted between lipid headgroups, which will avoid overestimation of 
area per lipid.  So, I'm testing this with small bilayer (128 lipids) 
using g_sas, and have some questions.

1) If the bigger solvent probes (such as -solsize 1.7 or so) are used, 
SASA of lipid will become smaller, but should be always bigger than the 
XY-area of this small bilayer system.  But, when I increased the value 
of -solsize, SASA becomes even smaller than the XY-area at some point. 
In this case, how can SASA be smaller than XY-area of the bilayer?

2) The manual says that -ndots is related to accuracy, but when I tried 
different values, I don't see trends of increase or decrease.  Could 
you tell me what it is and what the reasonable value will be?

3) I saw SASA fluctuate quite a lot. Is there any way to decrease fluctuation?

Thanks for your help in advance.

best,
Hwankyu.