[gmx-users] Minimum periodic distance

[Mark Abraham]

On 28/05/2011 2:16 PM, Kavyashree M wrote:
> Dear users,
>
> I ran a simulation of a protein in water with rlist = rcoloumb = 1.4nm,
> rvdw = 1.0nm, with PME and the distance between in the periodic
> was set to a minimum of 2.0 i.,e distance between protein and cell
> edge was set to 1.0nm, Even then, when I ran g_mindist, there was
> a message stating that  -
>
> "shortest periodic distance is 1.39714 (nm) at time 21824 (ps),
> between atoms 2062 and 3688"
> where 2062 is a protein atom and 21824 is a water molecule.
>
> I wanted to know why this happens in spite of taking care of it
> earlier.

Apparently you didn't achieve what you thought you achieved. Without 
knowing what you did, nobody can say why, unless your protein unfolded. 
Run g_mindist on your input structure to grompp.

> what are the consequences, what is the solution?

Maybe you didn't set up as large a box as you thought (read 
documentation carefully for diameter-vs-radius effects), maybe your 
simulation conditions are not stable for your protein, maybe you've run 
a simulation that is flawed because the periodic images could see each 
other.

Mark