<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7654.12">
<TITLE>RE: [gmx-users] Fatal Equilibration Errors</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<BR>
<P><FONT SIZE=2> It is extremely surprising that you have not found solute-solvent hydrogen bonds in aqueous solution. What is the O-C-C-O conformation? If it is trans, then what do the OH hydrogens do?<BR>
If O-C-C-O is gauche, then existence of two internal hydrogen bonds is possible, but it is<BR>
far from the quantum mechanically most stable form in the gas phase. I think that at least one solute-solvent hydrogen bond should have been seen. There are papers in the literature which do show solute-solvent hydrogen bonds in water. The exciting question is: there is one<BR>
H-O-C-C trans moiety in the gas phase, and the other H is H-bonded to this oxygen, when the O-C-C-O skeleton is gauche. Then the question of solvation is, whether this intramolecular hydrogen bond is maintained or disrupts. In aqueous solution it may, whereas in chloroform or CCl4 both OH may be internally H-bonded. Experiments found 12% O-C-C-O trans and 88% O-C-C-O gauche conformers in aqueous solution. Modeling should find what are the positions of the hydroxy hydrogens.<BR>
If you really do not find solute-solvent hydrogen bonds, your ethylene-glycol charges may be<BR>
questionable. What charge values did you use? A possibility is that the glycol OH charges cannot compete with those of the water model (SPC, TIPxP or what) in forming intermolecular hydrogen bonds.<BR>
<BR>
Peter Nagy<BR>
The University of Toledo,<BR>
Toledo, OH 43606<BR>
<BR>
-----Original Message-----<BR>
From: gmx-users-bounces@gromacs.org on behalf of Nancy<BR>
Sent: Sat 8/8/2009 3:33 PM<BR>
To: Discussion list for GROMACS users<BR>
Subject: [gmx-users] Fatal Equilibration Errors<BR>
<BR>
Hello,<BR>
<BR>
I have successfully minimised and equilibrated ethylene glycol in a water<BR>
box. I have noticed that there seem to be no hydrogen bonds between the<BR>
solute and solvent, but there are hydrogen bonds forming and breaking<BR>
between solvent molecules. Is this a normal behavior during minimsation and<BR>
equilibration?<BR>
<BR>
I am now trying to run MD on glycerol. I start with a "glycerol.mol2" file<BR>
which contains the structure of glycerol. I am using the following commands<BR>
to setup and run minimisation and equilibration on glycerol:<BR>
<BR>
$ .../topolbuild1_2_1/src/topolbuild -n glycerol -dir<BR>
.../topolbuild1_2_1/dat/gromacs -ff gmx53a6<BR>
<BR>
I modified the "defaults" part of the "ffglycerol.itp" (generated by<BR>
topolbuild) file to read:<BR>
<BR>
===========================<BR>
[ defaults ]<BR>
;nbfunc comb-rule gen-pairs fudgeLJ fudgeQQ<BR>
1 1 no 1.0 1.0<BR>
===========================<BR>
<BR>
$ editconf -f glycerol.gro -o glycerol_box.gro -box 2.65 2.65 2.65<BR>
$ genbox -cp glycerol_box.gro -cs spc216.gro -o glycerol_solv.gro -p<BR>
glycerol.top -box 2.65 2.65 2.65<BR>
$ grompp -f em.mdp -c glycerol_solv.gro -p glycerol.top -o em.tpr<BR>
<BR>
my "em.mdp" file is as follows:<BR>
<BR>
===========================<BR>
constraints = none<BR>
integrator = steep<BR>
nsteps = 10000<BR>
<BR>
emtol = 10.0<BR>
emstep = 0.01<BR>
<BR>
nstlist = 2<BR>
coulombtype = PME<BR>
nstcomm = 2<BR>
<BR>
ns_type = grid<BR>
rlist = 1.0<BR>
rcoulomb = 1.0<BR>
rvdw = 1.3<BR>
nstxout = 2<BR>
<BR>
pbc = xyz<BR>
pme_order = 4<BR>
===========================<BR>
<BR>
$ mdrun -v -deffnm em<BR>
<BR>
the last few lines of output are:<BR>
<BR>
===============================================================<BR>
Step= 6990, Dmax= 4.7e-04 nm, Epot= -3.34590e+04 Fmax= 9.13695e+01, atom= 4<BR>
Step= 6992, Dmax= 2.8e-04 nm, Epot= -3.34590e+04 Fmax= 1.40130e+02, atom= 5<BR>
Step= 6993, Dmax= 3.4e-04 nm, Epot= -3.34591e+04 Fmax= 8.57813e+01, atom= 4<BR>
Step= 6994, Dmax= 4.0e-04 nm, Epot= -3.34591e+04 Fmax= 2.41887e+02, atom= 5<BR>
Step= 6995, Dmax= 4.8e-04 nm, Epot= -3.34592e+04 Fmax= 9.36159e+01, atom= 4<BR>
Step= 6999, Dmax= 7.3e-05 nm, Epot= -3.34592e+04 Fmax= 4.95606e+01, atom= 4<BR>
Step= 7006, Dmax= 1.4e-06 nm, Epot= -3.34592e+04 Fmax= 4.89221e+01, atom= 4<BR>
Stepsize too small, or no change in energy.<BR>
Converged to machine precision,<BR>
but not to the requested precision Fmax < 10<BR>
<BR>
Double precision normally gives you higher accuracy.<BR>
You might need to increase your constraint accuracy, or turn<BR>
off constraints alltogether (set constraints = none in mdp file)<BR>
<BR>
writing lowest energy coordinates.<BR>
<BR>
Steepest Descents converged to machine precision in 7007 steps,<BR>
but did not reach the requested Fmax < 10.<BR>
Potential Energy = -3.3459230e+04<BR>
Maximum force = 4.9560604e+01 on atom 4<BR>
Norm of force = 3.4399371e+00<BR>
===============================================================<BR>
<BR>
As I believe these forces to be acceptable, I proceed to equilibration:<BR>
<BR>
$ grompp -f nvt.mdp -c em.gro -p glycerol.top -o nvt.tpr<BR>
<BR>
where my "nvt.mdp" file is:<BR>
<BR>
===========================<BR>
title = Glycerol NVT equilibration<BR>
define = -DPOSRES<BR>
<BR>
integrator = md<BR>
nsteps = 5000<BR>
dt = 0.002<BR>
<BR>
nstxout = 10<BR>
nstvout = 10<BR>
nstenergy = 10<BR>
nstlog = 10<BR>
<BR>
continuation = no<BR>
constraint_algorithm = lincs<BR>
constraints = all-angles<BR>
lincs_iter = 1<BR>
lincs_order = 4<BR>
<BR>
ns_type = grid<BR>
nstlist = 5<BR>
rlist = 1.0<BR>
rcoulomb = 1.0<BR>
rvdw = 1.3<BR>
<BR>
coulombtype = PME<BR>
pme_order = 4<BR>
fourierspacing = 0.16<BR>
<BR>
tcoupl = V-rescale<BR>
tc-grps = GOL SOL<BR>
tau_t = 0.1 0.1<BR>
ref_t = 300 300<BR>
<BR>
pcoupl = no<BR>
<BR>
pbc = xyz<BR>
<BR>
DispCorr = EnerPres<BR>
<BR>
gen_vel = yes<BR>
gen_temp = 300<BR>
gen_seed = -1<BR>
===========================<BR>
<BR>
where "GOL" refers to GlycerOL.<BR>
<BR>
$ mdrun -v -deffnm nvt<BR>
<BR>
When I execute the above command, I get numerous errors of the following<BR>
type:<BR>
<BR>
===============================================================<BR>
Step 0, time 0 (ps) LINCS WARNING<BR>
relative constraint deviation after LINCS:<BR>
rms 0.326337, max 0.500490 (between atoms 1 and 3)<BR>
bonds that rotated more than 30 degrees:<BR>
atom 1 atom 2 angle previous, current, constraint length<BR>
starting mdrun 'glycerol.pdb in water'<BR>
5000 steps, 10.0 ps.<BR>
<BR>
Step 0, time 0 (ps) LINCS WARNING<BR>
relative constraint deviation after LINCS:<BR>
rms 0.663850, max 1.227203 (between atoms 1 and 7)<BR>
bonds that rotated more than 30 degrees:<BR>
atom 1 atom 2 angle previous, current, constraint length<BR>
1 3 172.3 0.2091 0.1644 0.2004<BR>
7 9 164.1 0.2082 0.1625 0.2004<BR>
2 3 124.5 0.1022 0.0787 0.1000<BR>
1 2 148.3 0.1477 0.0882 0.1435<BR>
5 6 30.4 0.0992 0.1265 0.1000<BR>
8 9 134.1 0.1014 0.0660 0.1000<BR>
7 8 138.2 0.1476 0.0991 0.1435<BR>
<BR>
Warning: 1-4 interaction between 2 and 7 at distance 6.002 which is larger<BR>
than the 1-4 table size 2.300 nm<BR>
These are ignored for the rest of the simulation<BR>
This usually means your system is exploding,<BR>
if not, you should increase table-extension in your mdp file<BR>
or with user tables increase the table size<BR>
<BR>
t = 0.004 ps: Water molecule starting at atom 1234 can not be settled.<BR>
Check for bad contacts and/or reduce the timestep.<BR>
Wrote pdb files with previous and current coordinates<BR>
<BR>
There were 12 inconsistent shifts. Check your topology<BR>
Segmentation fault<BR>
===============================================================<BR>
<BR>
I don't know what is happening wrong. Please advise.<BR>
<BR>
Thank you.<BR>
<BR>
Nancy<BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>