[gmx-users] Energy minimisation query

[Justin A. Lemkul]

nikhil damle wrote:
> Hi,
> I am trying to simulate a dimeric structure and I have already simulated 
> its monomeric form. I use exactly same parameters which i use for 
> monomer during dimer minimisation in solvent.
> 
> System Details:
> # Atoms = 4488 (# Res = 582)
> octahedron box with d = 0.9
> Solvent (spce water model) density = 1050.6 g/L (During monomer MD it 
> was 1048.86 g/L)
> 
> System subjected to energy minimisation,
> Only warning arising during grompp:
> -------------------------------------------------------
> WARNING 1 [file dimer.top, line 46]:
>   86974 non-matching atom names
>   atom names from dimer.top will be used
>   atom names from neutral.pdb will be ignored
> -------------------------------------------------------
> This was neglected using -maxwarn option and contd...
> Error during minimisation run (mdrun):
> -------------------------------------------------------
> Steepest Descents:
>    Tolerance (Fmax)   =  1.00000e+03
>    Number of steps    =        50000
> Step=   14, Dmax= 1.2e-06 nm, Epot=  7.04150e+23 Fmax=         inf, atom= 37
> Stepsize too small, or no change in energy.
> Converged to machine precision,
> but not to the requested precision Fmax < 1000
> 
> Double precision normally gives you higher accuracy.
> 
> writing lowest energy coordinates.
> 
> Steepest Descents converged to machine precision in 15 steps,
> but did not reach the requested Fmax < 1000.
> Potential Energy  =  6.9065950e+23
> Maximum force     =            inf on atom 37
> Norm of force     =            inf
> -----------------------------------------------------
> 
> 1) I thought that the box size is too small so that dimer is not able to 
> move at all and hence the energy is too high. But I am getting same 
> error even if box size is increased to 2.0 nm
> 2) I checked my initial structure and even the neutralised structure 
> post solvation only to find that it is not broken. (My system has a 
> missing loop which I have modeled using modeler9v8 and then subjected it 
> to MD in solvent.)
> 
> What could possibly go wrong ? Could it be really the step size or 
> precision although I do not think so !
> 

Nearly 87,000 atoms are being assigned parameters that do not belong to them. 
The order of the [molecules] directive and the coordinate file must match. 
Otherwise, grompp (and then mdrun) will think that water is protein, and protein 
residues are other residues, etc.  I would be a fan of making -maxwarn a hidden 
option; it really should not be used unless one knows the exact consequences of 
doing so.

-Justin

> Regards,
> Nikhil
> 

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

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

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