[gmx-users] Re: Failed Energy Minimization

[Jennifer Rendell]

Justin, Your email makes me wonder two things.

1. Are you doing a series of energy minimizations? Do you start with an
    emtol of say 50000, then depending on the results of that, reduce it to
    25000 and so on until you reach the recommended value of 100?

It is unclear when you say
> If I try running with ns_type = simple, the em runs until it
> hits around 35
whether you mean energy minimization is at 35 steps in one cycle at a 
particular tolerance, or is 35 the tolerance at which it fails or I don't 
know...

2. Have you tried an energy minimization cycle with only your peptide9s(s)
and no water at all?

I hope this is helpful, Jennifer

-------------------------------------------------------------------------------
Jennifer Rendell 
Department of Biochemistry
Memorial University of Newfoundland
St. John's, Newfoundland
CANADA	A1B 3X9
-------------------------------------------------------------------------------

>
> Message: 2
> Date: Thu, 19 Apr 2007 18:01:05 -0500
> From: "Justin M. Shorb" <shorb at wisc.edu>
> Subject: [gmx-users] Failed Energy Minimization
> To: gmx-users at gromacs.org
> Message-ID: <3d47a42570a3ecd325852e707a2377a5 at wisc.edu>
> Content-Type: text/plain; charset=US-ASCII; format=flowed
>
> Greetings:
>
> I have been having trouble running any energy minimization with any
> sort of system. I have small polypeptides in water, and then large
> proteins in water with the same error message:
>
> Steepest Descents:
>    Tolerance (Fmax)   =  2.00000e+02
>    Number of steps    =        10000
> Step=    0, Dmax= 2.0e-02 nm, Epot= -2.84070e+04 Fmax=         inf,
> atom= 17
> -------------------------------------------------------
> Program mdrun_d, VERSION 3.3.1
> Source code file: nsgrid.c, line: 226
>
> Range checking error:
> Explanation: During neighborsearching, we assign each particle to a grid
> based on its coordinates. If your system contains collisions or
> parameter
> errors that give particles very high velocities you might end up with
> some
> coordinates being +-Infinity or NaN (not-a-number). Obviously, we cannot
> put these on a grid, so this is usually where we detect those errors.
> Make sure your system is properly energy-minimized and that the
> potential
> energy seems reasonable before trying again.
>
> Variable ci has value -2147483648. It should have been within [ 0 ..
> 125 ]
> Please report this to the mailing list (gmx-users at gromacs.org)
> -------------------------------------------------------
>
> No matter what run parameters I get, I always get ci = -2147483648.
> Even different systems. (This, I realize is simply numerically
> infinity). If I try running with ns_type = simple, the em runs until it
> hits around 35 and then says it didn't converge. Since I am running
> grompp_d -v, I get that there are certain atoms that have infinite
> potential. But, depending on the system, the atom type (solvent or
> molecule) changes.
>
> I have noted that previous posts have been answered with an admonition:
> "Find out which atoms are infinite force, and then fix it." I suppose I
> could go through, find the atoms and remove those waters that overlap,
> but shouldn't this be taken into account by the genbox_d program? I
> also have tried to increase the vdw radii in the vdwradii.dat file to
> have it delete more waters. But, even after having 60 fewer waters
> around a 88 kD protein, the system still fails to energy minimize.
>
> Given that this seems to be a common problem in solvating a system, are
> there better options (options for genbox??) than running the following
> shell commands? The em.mdp file is also shown below.
>
> Thanks for your help,
> Justin
>
>
> #           Set up the files
> # First, the pdb is generated into a gro file, and other files
> pdb2gmx_d  -f ${INPUTPDB} -o ${MOL}.gro -i posre.itp -p ${MOL}_spc.top
> -ter < p2g.in
>
> # Now modify the .pdb to use a periodic box
> editconf_d -bt cubic -f ${MOL}.gro -o ${MOL}.gro -c -d ${PBCx}
>
> #Solvate the box with spc, but the water type can be changed in next
> line
> genbox_d -cp ${MOL}.gro -cs spc216.gro -o ${MOL}-sol_b4em.gro -p
> ${MOL}_spc.top
>
> # Now we specify which topology to use for our water (by default, SPC)
> sed -e ${WATER} ${MOL}_spc.top >> ${MOL}-sol.top
>
> # Now everything is solvated, so we switch molecular names (for
> simplicity)
> MOL="${MOL}-sol"
>
> #           Perform Energy Minimization (necessary when solvating)
> # GROmacs Pre-Processor
> grompp_d -v -f em.mdp -c ${MOL}_b4em.gro -p ${MOL}.top -o ${MOL}_em
> mdrun_d -v -s ${MOL}_em -o ${MOL}_em -c ${MOL}_b4eq
>
> #           Perform equilibration of the system
> grompp_d  -f eq.mdp -c ${MOL}_b4eq.gro -p ${MOL}.top -o ${MOL}_eq
> mdrun_d -s ${MOL}_eq -o ${MOL}_eq -c ${MOL}_b4md
>
> #           Perform MD
> grompp_d  -f md.mdp -c ${MOL}_b4md.gro -p ${MOL}-v.top -o ${MOL}_md
> mdrun_d -s ${MOL}_md -o ${MOL}_md -c ${MOL}_final
>
>
> CUT=1.20
>
> #-------       Energy Minimization: em.mdp       -------#
> cat > em.mdp <<EOF
> title               =  ${MOL}
> cpp                 =  /usr/bin/cpp
> define              =  -DFLEX
> constraints         =  none
> integrator          =  steep
> nstlog              =  100
> ns_type             =  grid
> dt                  =  0.001    ; ps !
> nsteps              =  10000
> nstlist             =  1
> coulombtype         =  PME
> rlist               =  ${CUT}
> rcoulomb            =  ${CUT}
> rvdw                =  ${CUT}
> pme_order           =  6
> fourierspacing      =  0.05
> ewald_rtol          =  1e-5
> ;
> ;       Energy minimizing stuff
> ;
> emstep              =  0.02
> emtol               =  200.0
>
> EOF
>
> _________________________________________________________________
> Justin M. Shorb			Phone: (608) 262-0483
> Skinner Group				shorb at wisc.edu
> University of Wisconsin-Madison,  Department of Chemistry
> 1101 University Ave., Madison, WI, 53706
>
>