[gmx-users] high repulsion in packed system

[Justin A. Lemkul]

Moeed wrote:
> 
> Hello everyone,
> 
> I am trying to energy minimize a system of 8 polyethylene molecules with 
> 60 units. In literature density of PE is reported 0.9 gr/cm3. Based on 
> this I need to have each single chain in 3.3 nm^3. If I put this chain 
> in a box of 15*0.5*0.5 =3.75 nm^3 I am close to the density I want.
> 
> genconf -f PE60-3.75nm3.gro -nbox 1 4 2 -o PE60-3.75nm3-nbox142.gro
> 

You've already been given several tips, but I'll weigh in with one more, related 
to the construction of your system.  Using this setup, you will wind up with box 
dimensions of 15 x 2 x 1 nm, the most problematic of which is the z-dimension. 
Unless you're using a cutoff of < 0.5 nm (which you certainly shouldn't!), 
grompp should have raised an error.  Based on the values of your cutoffs, I'd 
suspect that either your box dimensions are (for some reason) not what I think 
they are, or your bypassing the error message somehow.

If these are not your box vectors, I'd be curious to know what dimensions your 
system actually occupies.

> Now I have 8 molecules but these long chains are very packed and I get very large repulsive potential after EM.I increased nsepts to 4000 but still I have problem ( I tried different integrator algorithms as well). Later I want to work with chains much longer than 60 units and system becomes even more packed. 
> 
> 
> Any suggestions is appreciated.
> 
> 
> Potential Energy  =  1.22210418131690e+08
> Maximum force     =  1.23807588295471e+07 on atom 364
> Norm of force     =  1.82513032084293e+06
> 

With an Fmax this high, anything you do will certainly fail.  I've already given 
you very thorough advice for how to solve such situations, and I've even 
immortalized these tips on the wiki :)

If a system of 8 molecules is unstable, start with a system containing just one...

-Justin

> 
> ******************************************************em.mdp
> 
> 
> constraints         =  all-bonds
> integrator          =  steep      ; use cg for conjugate gradient
> dt                  =  0.002            ; ps ! not necessary for EM, only needed for dynamics integrator like MD
> 
> nsteps              =  40000            ; 200   in minimization runs, this is just the max no. of iterations
> nstlist             =  10               ; frequency to update neighbor list. updates every 10 steps
> ns_type             =  grid
> 
> 
> rlist               =  1.0              ; cut-off distance for short-range neighbor list
> rcoulomb            =  1.0              ; distance for coulomb cut-off
> rvdw                =  1.0
> 
> ;       Energy minimizing stuff
> 
> 
> emtol               =  1000.0           ; minimization converges when max force is smaller than this value, in units of KJ/mol.nm
> emstep              =  0.01             ; initial step-size
> 
> 

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================