Dear Servaas,<div><br></div><div>Tested again in 'vacuum' and I saw no problems. Here goes what I did:</div><div><br></div><div><div><br></div><div>#----------------------------------</div><div>cat << EOF >| <a href="http://leap.in">leap.in</a></div>
<div>verbosity 1</div><div>source leaprc.ff99SB</div><div>ad = sequence { DA5 DA DA3 }</div><div>saveamberparm ad da_amber.top da_amber.crd</div><div>savepdb ad DA.pdb</div><div>quit</div><div>EOF</div><div>tleap -f <a href="http://leap.in">leap.in</a> >| leap.out</div>
<div><br></div><div>acpypi -x da_amber.crd -p da_amber.top -d # acpypi generates em.mdp and md.mdp</div><div><br></div><div>cat << EOF >| md.mdp</div><div>cpp = /usr/bin/cpp</div><div>define = ;-DFLEXIBLE</div>
<div>integrator = md</div><div>nsteps = 250000</div><div>constraints = none</div><div>emtol = 1000.0</div><div>emstep = 0.01</div><div>comm_mode = angular</div>
<div>ns_type = simple</div><div>nstlist = 0</div><div>rlist = 0</div><div>rcoulomb = 0</div><div>rvdw = 0</div><div>Tcoupl = no</div>
<div>Pcoupl = no</div><div>gen_vel = no</div><div>nstxout = 100</div><div>pbc = no</div><div>EOF</div><div><br></div><div>editconf -bt cubic -d 1.0 -f da_amber_GMX.gro -o da_amber_GMX.gro</div>
<div><br></div><div>#Single precision</div><div>grompp -f em.mdp -c da_amber_GMX.gro -p da_amber_GMX.top -o em.tpr</div><div>mdrun -v -deffnm em</div><div><br></div><div>grompp -f md.mdp -c em.gro -p da_amber_GMX.top -o md.tpr</div>
<div>mdrun -v -deffnm md</div><div><br></div><div>vmd md.gro md.trr</div><div>#----------------------------------</div><div><br></div><div>As you may suspect from the beginning it may be something in your mdp file. Case the example above works, I would suggest you to try the mdp for solvent box I sent before in a long simulation.</div>
<div><br></div><div>Good luck.</div><div><br></div>Regards,</div><div><br></div><div>Alan</div><div><br></div><div><div class="gmail_quote">On Wed, Dec 2, 2009 at 11:10, Alan <span dir="ltr"><<a href="mailto:alanwilter@gmail.com">alanwilter@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Dear Servaas,<br>
<br>
In tleap did you really did:<br>
<br>
TLEAP<br>
tleap -f leaprc.ff99SB<br>
ad = sequence { DA5 DA DA3 }<br>
saveamberparm da da_amber.top da_amber.crd<br>
<br>
<br>
If so, it's wrong, it should be:<br>
<br>
saveamberparm ad da_amber.top da_amber.crd<br>
^^^<br>
and not 'da'<br>
<br>
Besides, I tried to reproduce what you did using what I think would be<br>
fine and... everything went fine! Energies after minimisation in<br>
single and double were almost identical and trajectories diverted<br>
normally.<br>
<br>
Please check what I did.<br>
<br>
# begin commands<br>
<br>
cat << EOF >| em.mdp<br>
define = -DFLEXIBLE<br>
integrator = cg ; steep<br>
nsteps = 200<br>
constraints = none<br>
emtol = 1000.0<br>
nstcgsteep = 10 ; do a steep every 10 steps of cg<br>
emstep = 0.01 ; used with steep<br>
nstcomm = 1<br>
coulombtype = PME<br>
ns_type = grid<br>
rlist = 1.0<br>
rcoulomb = 1.0<br>
rvdw = 1.4<br>
Tcoupl = no<br>
Pcoupl = no<br>
gen_vel = no<br>
nstxout = 0 ; write coords every # step<br>
optimize_fft = yes<br>
EOF<br>
<br>
<br>
cat << EOF >| md.mdp<br>
integrator = md<br>
nsteps = 1000<br>
dt = 0.002<br>
constraints = all-bonds<br>
nstcomm = 1<br>
ns_type = grid<br>
rlist = 1.2<br>
rcoulomb = 1.1<br>
rvdw = 1.0<br>
vdwtype = shift<br>
rvdw-switch = 0.9<br>
coulombtype = PME-Switch<br>
Tcoupl = v-rescale<br>
tau_t = 0.1 0.1<br>
tc-grps = protein non-protein<br>
ref_t = 300 300<br>
Pcoupl = parrinello-rahman<br>
Pcoupltype = isotropic<br>
tau_p = 0.5<br>
compressibility = 4.5e-5<br>
ref_p = 1.0<br>
gen_vel = yes<br>
nstxout = 2 ; write coords every # step<br>
lincs-iter = 2<br>
DispCorr = EnerPres<br>
optimize_fft = yes<br>
EOF<br>
<br>
<br>
cat << EOF >| <a href="http://leap.in" target="_blank">leap.in</a><br>
verbosity 1<br>
source leaprc.ff99SB<br>
ad = sequence { DA5 DA DA3 }<br>
solvatebox ad TIP3PBOX 10.0<br>
addions ad Na+ 5<br>
addions ad Cl- 3<br>
saveamberparm ad da_amber.top da_amber.crd<br>
savepdb ad DA.pdb<br>
quit<br>
EOF<br>
tleap -f <a href="http://leap.in" target="_blank">leap.in</a> >| leap.out<br>
<br>
acpypi -x da_amber.crd -p da_amber.top -d<br>
<br>
#Single precision<br>
grompp -f em.mdp -c da_amber_GMX.gro -p da_amber_GMX.top -o em.tpr<br>
mdrun -v -deffnm em<br>
<br>
#Polak-Ribiere Conjugate Gradients converged to Fmax < 1000 in 22 steps<br>
#Potential Energy = -6.2280516e+04<br>
#Maximum force = 7.5868494e+02 on atom 98<br>
#Norm of force = 1.0447179e+02<br>
<br>
grompp -f md.mdp -c em.gro -p da_amber_GMX.top -o md.tpr<br>
mdrun -v -deffnm md<br>
<br>
#Double precision<br>
grompp_d -f em.mdp -c da_amber_GMX.gro -p da_amber_GMX.top -o em.tpr<br>
mdrun_d -v -deffnm em<br>
<br>
#Polak-Ribiere Conjugate Gradients converged to Fmax < 1000 in 22 steps<br>
#Potential Energy = -6.22813514022256e+04<br>
#Maximum force = 7.58238100790309e+02 on atom 98<br>
#Norm of force = 1.04358667410458e+02<br>
<br>
grompp_d -f md.mdp -c em.gro -p da_amber_GMX.top -o md.tpr<br>
mdrun_d -v -deffnm md<br>
<br>
# end commands<br>
<br>
Regards,<br>
<font color="#888888"><br>
Alan<br>
</font><div><div></div><div class="h5"><br>
On Tue, Dec 1, 2009 at 13:56, Alan <<a href="mailto:alanwilter@gmail.com">alanwilter@gmail.com</a>> wrote:<br>
> Dear Servaas,<br>
><br>
> I've been following your thread. I am the developer of acpypi and<br>
> thanks for giving a try.<br>
><br>
> So, as you may already know, you are trying acpypi as <a href="http://amb2gmx.pl" target="_blank">amb2gmx.pl</a> so<br>
> far, but you also seemed to have read acpypi wikis and realise that<br>
> acpypi can help you to generate the whole topology for a ligand.<br>
><br>
> However, AFAIU you have only regular NA and not modified ones neither<br>
> ligands, right? But then why are you using RED?<br>
><br>
> I understand your approach about using tleap to create your whole<br>
> system and then convert it to GMX. It should work at first but it is<br>
> clearly not as you reported.<br>
><br>
> So, here goes some of my recommendations:<br>
><br>
> 1) GMX is vacuum is unrealistic and prone for errors. There's no GB<br>
> implementation as far as I know.<br>
><br>
> 2) Have you try to use pdb2gmx to generate your files from your pdb<br>
> directly to GMX?<br>
><br>
> 3) When you say that gmx double precision works, is your system in<br>
> vacuum or with solvent?<br>
><br>
> 4) if using tleap, create your system with solvent and ions and then<br>
> use acpypi to convert to gmx.<br>
><br>
> The use of amb2gmx or acpypi is to give you a system to be run<br>
> immediately in gromacs doing just a grompp and mdrun. Using editconf<br>
> will change the parameters of your box and it may have serious<br>
> implications besides that in amber we don't have dodecahedron, so if<br>
> doing what you're doing then you're not replicating the conditions you<br>
> have in amber with those in gmx (although it puzzles me that gmx<br>
> double works, with the commands you gave in gmx?).<br>
><br>
> I would ask you to give more details and even a detailed step by step<br>
> of commands of what you're doing including tleap.<br>
><br>
> Regards,<br>
> Alan<br>
><br>
><br>
><br>
> On Tue, Dec 1, 2009 at 11:00, <<a href="mailto:gmx-users-request@gromacs.org">gmx-users-request@gromacs.org</a>> wrote:<br>
>><br>
>> Thanks for your suggestion, I tried without success and I also tried<br>
>> shake. But this is also rather fighting the symptoms than the cause...<br>
>> And amber simulations in vacuum do work fine... My personal guess was<br>
>> that another parameter in my mdp file was not compatible with the amber<br>
>> force field, but I could not figure out which one. I also tried<br>
>> different settings, e.g. the one I found on the acpypi wiki.<br>
>><br>
><br>
> --<br>
> Alan Wilter Sousa da Silva, D.Sc.<br>
> PDBe group, PiMS project <a href="http://www.pims-lims.org/" target="_blank">http://www.pims-lims.org/</a><br>
> EMBL - EBI, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK<br>
> +44 (0)1223 492 583 (office)<br>
><br>
<br>
<br>
<br>
--<br>
Alan Wilter Sousa da Silva, D.Sc.<br>
PDBe group, PiMS project <a href="http://www.pims-lims.org/" target="_blank">http://www.pims-lims.org/</a><br>
EMBL - EBI, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK<br>
+44 (0)1223 492 583 (office)<br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Alan Wilter Sousa da Silva, D.Sc.<br>PDBe group, PiMS project <a href="http://www.pims-lims.org/">http://www.pims-lims.org/</a><br>EMBL - EBI, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK<br>
+44 (0)1223 492 583 (office)<br>
</div>