[gmx-users] Fast Question on g_msd

[Ivan Gladich]

Dear all
I am running a simulation of water slab, i.e. a water system with two 
air/vacuum interface, using a 5 site water model.
I am doing a simple test calculating the water diffusivity by mean 
square displacement in different regions of my water slab.
In order to do, I created an index that contains all the atoms of the 
water molecules that belong to a certain region using g_select

g_select -select " all and same residue as (y>2.5 and y<12.5)" -on 
select_index.ndx

At the end, my select_index.ndx contains all the atoms of all the water 
molecules that are between 2.5 and 12.5 nm
[all]
1 2 3 4 5 11 12 13 14 15 .....

After that, I calculate the diffusivity of all the waters that belong to 
this region.
g_msd -n select_index.ndx -s topol.tpr

I have a silly question.
1) Does g_msd calculate the means square displacement of all the atom 
that are select_index.ndx (and then it takes the average of all of them)?

OR

2)does g_msd read the topol.tpr, recognize that the atoms in 
select_index.ndx belong to specific water molecules and so, it calculate 
the msd based on the displacement of
the center of mass of each water molecule (and then it takes the average)?

Sorry, I tried to find something in the archive but I was able (or 
lucky) to find something.
Thanks for any possible help
Ivan

-----