[gmx-users] Isovalue for g_spatial representation

chris.neale at utoronto.ca chris.neale at utoronto.ca
Sun Oct 31 01:12:24 CEST 2010

No you can not do that. You can only say that it is drawn at N times  
the average number density of solvent molecules. To determine N, find  
out the average grid.cube value in a region of pure solvent, let that  
value be A, and N=10/A.

Clearly you're trying to do a figure caption, and I'm sorry that the  
tool does not do this for you. There are technical issues that make it  
harder than you might think to automate although it certainly could be  

You can estimate A by sliding the bar in VMD and A is next to where  
the bulk water disappears. This will give you an estimate that you can  
later use to ensure that your quantitative value is correct. To get  
the quantitative value, you'll probably need to run awk over the  
grid.cube file and average over a predefined section. Careful though,  
don't go all the way to the boundaries of the cube, the corners there  
are likely low density due to rotation from trjconv.


-- original message --

Dear Chris, thank you for information, they were very enlightening.
Just one more question. Suppose I use an isovalue, say 10, I can
say that the drawn surface is 10 times the average number density of solvent
molecules? I am using -div option.


More information about the gromacs.org_gmx-users mailing list