[gmx-users] Energy minimization has stopped, but the forces have not converged

[Justin Lemkul]

On 6/25/14, 10:12 AM, Iris Nira Smith wrote:
> Hello gromacs users,
>
> I have searched the gmx-users archive and have successfully found multiple posts that provide advice on how to troubleshoot my issue. Unfortunately, after trying multiple ways to troubleshoot as suggested by each of these posts, I am still having the same issue.
>
> I successfully mutated a residue in my .pdb file using VMD mutator and am now in the minimization phase using steepest descent step-wise energy minimization. The error I keep getting is as follows:
>
> Steepest Descents:
>     Tolerance (Fmax)   =  1.00000e+04
>     Number of steps    =         2000
> Step=    0, Dmax= 1.0e-02 nm, Epot=  5.18534e+14 Fmax= 3.39138e+04, atom= 3228^M
> Step=    1, Dmax= 1.0e-02 nm, Epot=  5.18534e+14 Fmax= 1.18748e+04, atom= 10021^M
> Step=    2, Dmax= 5.0e-03 nm, Epot=  5.18534e+14 Fmax= 1.73483e+04, atom= 3228^M
> Step=    3, Dmax= 2.5e-03 nm, Epot=  5.18534e+14 Fmax= 2.44124e+04, atom= 3228^M
> Step=    4, Dmax= 1.2e-03 nm, Epot=  5.18534e+14 Fmax= 2.88081e+04, atom= 3228^M
> Step=    5, Dmax= 6.2e-04 nm, Epot=  5.18534e+14 Fmax= 3.12647e+04, atom= 3228^M
> Step=    6, Dmax= 3.1e-04 nm, Epot=  5.18534e+14 Fmax= 3.25644e+04, atom= 3228^M
> Step=    7, Dmax= 1.6e-04 nm, Epot=  5.18534e+14 Fmax= 3.32326e+04, atom= 3228^M
> Step=    8, Dmax= 7.8e-05 nm, Epot=  5.18534e+14 Fmax= 3.35718e+04, atom= 3228^M
> Step=    9, Dmax= 3.9e-05 nm, Epot=  5.18534e+14 Fmax= 3.37432e+04, atom= 3228^M
> Step=   10, Dmax= 2.0e-05 nm, Epot=  5.18534e+14 Fmax= 3.38282e+04, atom= 3228^M
> Step=   11, Dmax= 9.8e-06 nm, Epot=  5.18534e+14 Fmax= 3.38707e+04, atom= 3228^M
> Step=   12, Dmax= 4.9e-06 nm, Epot=  5.18534e+14 Fmax= 3.38920e+04, atom= 3228^M
> Step=   13, Dmax= 2.4e-06 nm, Epot=  5.18534e+14 Fmax= 3.39031e+04, atom= 3228^M
> Step=   14, Dmax= 1.2e-06 nm, Epot=  5.18534e+14 Fmax= 3.39092e+04, atom= 3228^M
>
> Energy minimization has stopped, but the forces havenot converged to the
> requested precision Fmax < 10000 (whichmay not be possible for your system).
> It stoppedbecause the algorithm tried to make a new step whose sizewas too
> small, or there was no change in the energy sincelast step. Either way, we
> regard the minimization asconverged to within the available machine
> precision,given your starting configuration and EM parameters.
>
> Double precision normally gives you higher accuracy, butthis is often not
> needed for preparing to run moleculardynamics.
> You might need to increase your constraint accuracy, or turn
> off constraints altogether (set constraints = none in mdp file)
>
> writing lowest energy coordinates.
>
> Steepest Descents converged to machine precision in 15 steps,
> but did not reach the requested Fmax < 10000.
> Potential Energy  =  5.1853375e+14
> Maximum force     =  3.3913793e+04 on atom 3228
> Norm of force     =  6.5862207e+02
>
> I have analyzed my .gro file for steric clashes with atom 3228 (HE2 atom on a tyrosine located 20 amino acids from c-terminal tail) and the only thing that could potentially clash with it is a hydrogen atom on a nearby water which is 1.2 Angstroms away. I first moved x-coordinate of atom 3228 by 0.1nm (or 1.0 Angstrom), but the same error occurred. I then removed that entire water molecule that was near atom 3228, re-ran the energy minimization and the same error occurred. I changed the emtol from 1,000 to 10,000 even to 100,000 and the same error occurred. I then analyzed the energy using g_energy, but the following error occurred:
>

Look at the energy terms in the .log file.  Something should stick out like a 
sore thumb here.

> -------------------------------------------------------
> Program g_energy, VERSION 4.6.3
> Source code file: /var/tmp/packages/gromacs-4.6.3/src/gmxlib/enxio.c, line: 828
>
> Fatal error:
> Energy file g36r_em2.edr not recognized, maybe different CPU?
> For more information and tips for troubleshooting, please check the GROMACS
> website at http://www.gromacs.org/Documentation/Errors
> -------------------------------------------------------
>
> I finally started over and created a new configuration (.gro) file as well as a topology (.top) file, but the same error occurred. Note: I have 13 other mutant model systems that I have run minimization-equilibration successfully but this particular mutant model keeps giving me the same repeated error. The potential energy of this mutant system as well as the force on atom 3228 is rather high. Can you offer any advice or alternate suggestions to troubleshooting this error? Should I run a few steps of CG first and then SD?
>

The fact that the energy is essentially infinite and never changes means it's an 
unresolvable problem.  There is likely something intrinsically wrong with the 
coordinates or the topology.  You're not missing any atoms or causing pdb2gmx to 
do anything strange, are you?

-Justin

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

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

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