[Fwd: Re: [gmx-users] A problem with a "detaching Calpha/s"]
Arik Cohen
acohen at biochem.duke.edu
Tue Jan 5 22:30:45 CET 2010
Thanks a lot the ultrafast response !
Arik
Justin A. Lemkul wrote:
>
>
> Arik Cohen wrote:
>> Dear Gromacs users,
>>
>> In continuation to the problem below, it seems that self interaction
>> is not the problem since a -d 1.5nm around the protein should have
>> been more than enough. Not noticing earlier, I see strange files
>> being created with the name "step1150b_n1.pdb"
>> (a bug like structural report of that step ?). In those files and in
>> the pdb created after applying trjconv, the protein is drifted from
>> the cell's center to one of the edge and a fragment appears at the
>> other side.
>> Saying that, trjconv does not solves the problem.
>>
>> I would be most thankful for any comment/suggestion as to how to
>> prevent this drift which causes in turn to the "fragment" problem.
>>
>
> Check your log file for LINCS warnings, etc. These step_*.pdb files
> are written when the system becomes unstable and is on the verge of
> collapse. Your system is probably blowing up, hence the weird
> coordinates.
>
> -Justin
>
>> Thank
>>
>> Arik
>>
>> Mark James Abraham wrote:
>>>
>>> On 01/01/10, *"Justin A. Lemkul" *<jalemkul at vt.edu> wrote:
>>>>
>>>>
>>>> Arik Cohen wrote:
>>>> >No, sorry for the confusion. The images are only from a simulation
>>>> of one protein(tmRBP_Unliganded, PDB: 2FN9). The problem seen with
>>>> BPTI was a bit different in a way that only 1 C-alpha was
>>>> "detached" (cell size ?).
>>>> >
>>>>
>>>> If there is supposed to be only a single protein in the images
>>>> you've shown (and after having a look at the 2FN9 structure from
>>>> the PDB), it seems pretty clear to me that part of your protein has
>>>> simply crossed a periodic boundary and trjconv -pbc mol (or some
>>>> such command) should fix it.
>>> Indeed. Further, note that the reappearance of a subset of the
>>> seemingly detached atoms back with the main group is totally normal.
>>> With domain decomposition in GROMACS 4, such "broken" molecules can
>>> be written to the trajectory. Judicious use of trjconv fixes the
>>> appearance of this non-problem.
>>>
>>> Mark
>>>
>>>>
>>>> >Arik
>>>> >
>>>> >Justin A. Lemkul wrote:
>>>> >>
>>>> >>
>>>> >>Arik Cohen wrote:
>>>> >>>Which two proteins ? I have at least in beginning only one
>>>> protein which some how is divided into two along the calculation.
>>>> >>>Any way I'll try both increasing the cell and fix it with trjconv.
>>>> >>>
>>>> >>
>>>> >>Quoting your original message:
>>>> >>
>>>> >>"While running a simple MD simulation with both a small protein
>>>> such as BPTI and a larger one such as tmRBP_Unliganded_2FN9.pdb..."
>>>> >>
>>>> >>I assumed that what I was seeing in the images was a set of two
>>>> proteins. My concern was that you defined a box relative to the
>>>> larger protein, then inserted the smaller one (BPTI?), leaving
>>>> insufficient space in the box to satisfy the minimum image
>>>> convention. If that's not what we're looking at, then that'd be
>>>> useful to know :)
>>>> >>
>>>> >>If you have a single protein, "divided into two" then the problem
>>>> is almost certainly a simple periodicity artifact. Bonds do not
>>>> break in a normal MD calculation (in fact they can't using the
>>>> standard MM approximations).
>>>> >>
>>>> >>-Justin
>>>> >>
>>>> >>>Thanks a lot
>>>> >>>
>>>> >>>Arik
>>>> >>>
>>>> >>>Justin A. Lemkul wrote:
>>>> >>>>
>>>> >>>>
>>>> >>>>Arik Cohen wrote:
>>>> >>>>>I'm using dodecahedron -d 0.7
>>>> >>>>>
>>>> >>>>>
>>>> >>>>
>>>> >>>>Was that distance specified with respect to both of the protein
>>>> molecules in the unit cell? You can check for spurious PBC
>>>> interactions with g_mindist -pi. Anyway, I'd be curious to see how
>>>> you do with trjconv.
>>>> >>>>
>>>> >>>>-Justin
>>>> >>>>
>>>> >>>>>
>>>> >>>>>Justin A. Lemkul wrote:
>>>> >>>>>>
>>>> >>>>>>
>>>> >>>>>>Arik Cohen wrote:
>>>> >>>>>>>Hi,
>>>> >>>>>>>
>>>> >>>>>>>I have not tried yet to fix it with trjconv which I will .
>>>> Attached is a picture with 4 snapshots taken from the simulation.
>>>> The C-alphas in question are emphasized with red color.
>>>> >>>>>>>
>>>> >>>>>>
>>>> >>>>>>Is your unit cell sufficiently large? It looks like the
>>>> C-alphas indicated are simply crossing the periodic boundary on the
>>>> "left" of the frame and interacting with the protein molecule in
>>>> the "right" of the frame, which would indicate to me that the unit
>>>> cell is too small and you're seeing spurious PBC interactions
>>>> (i.e., violation of the minimum image convention).
>>>> >>>>>>
>>>> >>>>>>-Justin
>>>> >>>>>>
>>>> >>>>>>>Thanks
>>>> >>>>>>>
>>>> >>>>>>>Arik
>>>> >>>>>>>
>>>> >>>>>>>Justin A. Lemkul wrote:
>>>> >>>>>>>>
>>>> >>>>>>>>
>>>> >>>>>>>>Arik Cohen wrote:
>>>> >>>>>>>>>Hi,
>>>> >>>>>>>>>
>>>> >>>>>>>>>Sorry to bother you again ,but its not only a periodic
>>>> effect since only *some of the atoms* in the "Detached" group are
>>>> vanishing from this group and reappearing in the main protein
>>>> group. The rest of the atoms are either always in the detached or
>>>> the main group.
>>>> >>>>>>>>>In addition, the "detached" group includes three segments
>>>> of the protein(8 residues(126-131), 8 residues(157-164) and 4
>>>> residues186-189).
>>>> >>>>>>>>>
>>>> >>>>>>>>
>>>> >>>>>>>>From your description, this sounds exactly like a
>>>> periodicity problem - some of the atoms are crossing the periodic
>>>> boundary and are appearing in strange locations. Have you even
>>>> tried trjconv to fix it? That would be useful information, as I
>>>> see that Mark long ago also suggested the same sort of fix.
>>>> >>>>>>>>
>>>> >>>>>>>>It is hard for me to envision what you are seeing. It
>>>> would be enormously helpful if you could post images (screenshots,
>>>> etc) of the problematic structures to get a more expedient resolution.
>>>> >>>>>>>>
>>>> >>>>>>>>-Justin
>>>> >>>>>>>>
>>>> >>>>>>>>>Thanks a lot
>>>> >>>>>>>>>
>>>> >>>>>>>>>Arik
>>>> >>>>>>>>>
>>>> >>>>>>>>>Justin A. Lemkul wrote:
>>>> >>>>>>>>>>
>>>> >>>>>>>>>>
>>>> >>>>>>>>>>Arik Cohen wrote:
>>>> >>>>>>>>>>>Hi,
>>>> >>>>>>>>>>>
>>>> >>>>>>>>>>>With regards to your question I do see some periodicity
>>>> in which for a section of time in the trajectory some of the
>>>> Calphas in the "detached group" are vanishing from it and reappear
>>>> in the main protein.
>>>> >>>>>>>>>>>In addition,
>>>> >>>>>>>>>>>I would appreciate as before any suggestion you might
>>>> have in the matter.
>>>> >>>>>>>>>>>
>>>> >>>>>>>>>>
>>>> >>>>>>>>>>If this is just a periodicity artifact, fix it with trjconv.
>>>> >>>>>>>>>>
>>>> >>>>>>>>>>-Justin
>>>> >>>>>>>>>>
>>>> >>>>>>>>>>>Thanks
>>>> >>>>>>>>>>>
>>>> >>>>>>>>>>>Arik
>>>> >>>>>>>>>>>
>>>> >>>>>>>>>>>Mark Abraham wrote:
>>>> >>>>>>>>>>>>Arik Cohen wrote:
>>>> >>>>>>>>>>>>>Hi,
>>>> >>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>Thanks for answering so quickly !. Apparently whole
>>>> residues have detached from the protein.
>>>> >>>>>>>>>>>>
>>>> >>>>>>>>>>>>So... like I asked last time, are you seeing a
>>>> periodicity artefact? "Detached" covers a whole gamut of
>>>> possibilities.
>>>> >>>>>>>>>>>>
>>>> >>>>>>>>>>>>>Another strange thing that happens in pyMol and VMD is
>>>> that when I select an atom or a residue in the detached group the
>>>> selection appears twice: one in the detached group and one in the
>>>> main part.
>>>> >>>>>>>>>>>>
>>>> >>>>>>>>>>>>If you've got atoms duplicated, then it sounds like
>>>> something's going wrong with how they're interpreting the structure
>>>> file, or how you're manipulating it afterwards. Either way, it's
>>>> not a problem for the GROMACS mailing list unless you can
>>>> demonstrate the atoms are duplicated in the structure file (which
>>>> they aren't!).
>>>> >>>>>>>>>>>>
>>>> >>>>>>>>>>>>Mark
>>>> >>>>>>>>>>>>
>>>> >>>>>>>>>>>>>Arik
>>>> >>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>Mark Abraham wrote:
>>>> >>>>>>>>>>>>>>Arik Cohen wrote:
>>>> >>>>>>>>>>>>>>>Dear GROMACS users,
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>While running a simple MD simulation with both a
>>>> small protein such as BPTI and a larger one such as
>>>> tmRBP_Unliganded_2FN9.pdb, I'm encountering an odd situation in
>>>> which one (in the case of BPTI) or several Calphas (in the later
>>>> case) are "detaching them selfs" from the main group.
>>>> >>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>"main group" of what? Do the atoms bound to them move
>>>> also? Are you seeing a periodicity artefact?
>>>> >>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>Mark
>>>> >>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>The problem appeared only after adding salt to the
>>>> simulation(at least in the case of BPTI).
>>>> >>>>>>>>>>>>>>>I would appreciate any suggestions and comments on
>>>> the matter.
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>Thanks
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>Arik
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>The run files are:
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>*em.mdp:*
>>>> >>>>>>>>>>>>>>> title = tmRBP_Unliganded_2FN9
>>>> Minimization
>>>> >>>>>>>>>>>>>>>integrator = steep ; (steep)using
>>>> steepest descent
>>>> >>>>>>>>>>>>>>>nsteps = 50000
>>>> >>>>>>>>>>>>>>>nstlist = 1
>>>> >>>>>>>>>>>>>>>rlist = 1.0
>>>> >>>>>>>>>>>>>>>coulombtype = PME
>>>> >>>>>>>>>>>>>>>rcoulomb = 1.0
>>>> >>>>>>>>>>>>>>>vdw-type = cut-off
>>>> >>>>>>>>>>>>>>>rvdw = 1.0
>>>> >>>>>>>>>>>>>>>nstenergy = 10
>>>> >>>>>>>>>>>>>>>emtol = 5.0 ; tolerance kJ/(Mol -1
>>>> nm-1) instead of 10.0
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>*pr.mdp
>>>> >>>>>>>>>>>>>>>*
>>>> >>>>>>>>>>>>>>>title = tmRBP_Unliganded_2FN9 PR
>>>> >>>>>>>>>>>>>>>integrator = md
>>>> >>>>>>>>>>>>>>>nsteps = 50000
>>>> >>>>>>>>>>>>>>>dt = 0.002 ;(in ps) doing a 100ps
>>>> traj.
>>>> >>>>>>>>>>>>>>>constraints = all-bonds
>>>> >>>>>>>>>>>>>>>nstlist = 10 ; neighbour list updates
>>>> every number of steps
>>>> >>>>>>>>>>>>>>>rlist = 1.0
>>>> >>>>>>>>>>>>>>>coulombtype = PME
>>>> >>>>>>>>>>>>>>>rcoulomb = 1.0
>>>> >>>>>>>>>>>>>>>vdw-type = cut-off
>>>> >>>>>>>>>>>>>>>rvdw = 1.0
>>>> >>>>>>>>>>>>>>>tcoupl = Berendsen
>>>> >>>>>>>>>>>>>>>tc-grps = Protein non-protein
>>>> >>>>>>>>>>>>>>>tau-t = 0.1 0.1
>>>> >>>>>>>>>>>>>>>ref-t = 298 298
>>>> >>>>>>>>>>>>>>>Pcoupl = Berendsen
>>>> >>>>>>>>>>>>>>>tau-p = 1.0
>>>> >>>>>>>>>>>>>>>compressibility = 5e-5 5e-5 5e-5 0 0 0
>>>> >>>>>>>>>>>>>>>ref-p = 1.0
>>>> >>>>>>>>>>>>>>>nstenergy = 100
>>>> >>>>>>>>>>>>>>>define = -DPOSRES ; include
>>>> posre.itp(position restraint) file
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>*run.md
>>>> >>>>>>>>>>>>>>>*title = tmRBP_Unliganded_2FN9
>>>> >>>>>>>>>>>>>>>integrator = md
>>>> >>>>>>>>>>>>>>>nsteps = 300000
>>>> >>>>>>>>>>>>>>>dt = 0.001
>>>> >>>>>>>>>>>>>>>constraints = all-bonds
>>>> >>>>>>>>>>>>>>>nstlist = 10
>>>> >>>>>>>>>>>>>>>rlist = 1.0
>>>> >>>>>>>>>>>>>>>coulombtype = PME
>>>> >>>>>>>>>>>>>>>rcoulomb = 1.0
>>>> >>>>>>>>>>>>>>>vdw-type = cut-off
>>>> >>>>>>>>>>>>>>>rvdw = 1.0
>>>> >>>>>>>>>>>>>>>tcoupl = V-rescale ;V-rescale
>>>> >>>>>>>>>>>>>>>tc-grps = Protein non-protein
>>>> >>>>>>>>>>>>>>>tau-t = 0.8 0.8
>>>> >>>>>>>>>>>>>>>ref-t = 298 298
>>>> >>>>>>>>>>>>>>>nstxout = 1000
>>>> >>>>>>>>>>>>>>>nstvout = 1000
>>>> >>>>>>>>>>>>>>>nstxtcout = 1000
>>>> >>>>>>>>>>>>>>>nstenergy = 1000
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>The runs commands are(integrated inside a C++ code):
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>SysCommand1 = "echo 6 | pdb2gmx -f " + FileName + "
>>>> -water tip3p";
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>> system("editconf -f conf.gro -bt dodecahedron -d
>>>> 0.7 -o box.gro");
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>system("genbox -cp box.gro -cs spc216.gro -p
>>>> topol.top -o solvated.gro");
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>minimization:
>>>> >>>>>>>>>>>>>>>--------
>>>> >>>>>>>>>>>>>>> if(Mode == "NoSalt")
>>>> >>>>>>>>>>>>>>> {
>>>> >>>>>>>>>>>>>>> system("grompp -f MDP/em.mdp -p topol.top -c
>>>> solvated.gro -o em.tpr");
>>>> >>>>>>>>>>>>>>> //system("mpirun -np 4 mdrun -v -deffnm em");
>>>> >>>>>>>>>>>>>>> }
>>>> >>>>>>>>>>>>>>> if(Mode == "WithSalt")
>>>> >>>>>>>>>>>>>>> {
>>>> >>>>>>>>>>>>>>> system("grompp -f MDP/em.mdp -p topol.top -c
>>>> solvated.gro -o em.tpr");
>>>> >>>>>>>>>>>>>>> system("mpirun -np 4 mdrun -v -deffnm em");
>>>> >>>>>>>>>>>>>>> }
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>
>>>> >>>>>>>>>>>>>>>Salting:
>>>> >>>>>>>>>>>>>>>--------
>>>> >>>>>>>>>>>>>>> system("echo 12 | genion -s em.tpr -conc 0.1
>>>> -neutral -o solvated.gro");
>>>> >>>>>>>>>>>>>>> pr:
>>>> >>>>>>>>>>>>>>>----
>>>> >>>>>>>>>>>>>>>system("grompp -f MDP/prmd.mdp -p topol.top -c
>>>> em.gro -o pr.tpr");
>>>> >>>>>>>>>>>>>>> /* The actual run*/
>>>> >>>>>>>>>>>>>>> system("mpirun -np 4 mdrun -v -deffnm pr");
>>>> >>>>>>>>>>>>>
>>>> >>>>>>>>>>
>>>> >>>>>>>>
>>>> >>>>>>>
>>>> >>>>>>>------------------------------------------------------------------------
>>>>
>>>> >>>>>>>
>>>> >>>>>>
>>>> >>>>>
>>>> >>>>
>>>> >>>
>>>> >>
>>>> >
>>>>
>>>> --
>>>> ========================================
>>>>
>>>> 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
>>>>
>>>> ========================================
>>>> --
>>>> gmx-users mailing list gmx-users at gromacs.org
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>> Please search the archive at http://www.gromacs.org/search before
>>>> posting!
>>>> Please don't post (un)subscribe requests to the list. Use the www
>>>> interface or send it to gmx-users-request at gromacs.org.
>>>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>
More information about the gromacs.org_gmx-users
mailing list