[gmx-users] dna and cnt get distorted in md simulation

Justin A. Lemkul jalemkul at vt.edu
Thu Apr 21 21:01:46 CEST 2011 


majid hasan wrote:
> Okay, so here is the file that I used for both energy minimization (with 
> integrator = l-bfgs), and MD (integrator = md). Everything other than 
> the value of integrator was same for both energy minimization and MD.
>  http://phas.ubc.ca/~majid/md.mdp
> 
> and here is what I see by running .trr files obtained from EM, and MD.
> 
> On running EM.trr file:
> In the beginning of simulation, DNA is very stretched i.e atoms of DNA 
> are widely separated from each other, and CNT crumples. Then, gradually 
> DNA shrinks and converges onto CNT. 
> 

OK, so the DNA is experiencing the charge repulsion I described earlier.  The 
CNT sounds like it does not have proper bonded interactions assigned.

> On running MD.trr file:
> CNT suddenly moves toward and DNA and one part is mixed with DNA and 
> whole structure is crumpled (similar to the final state of EM.trr 
> simulation). Then this crumpled structure wiggles around until the end 
> of simulation.
> 

If EM is showing such weird results, then you can't really trust anything you 
see in the MD afterwards.

> 
> Yes, I ran the whole process in vacuum. I am going to do this simulation 
> by changing cutoff to shift, and see which one works better, and then I 
> will do it with a solvent.
> 

There is no point in making this assessment in vacuo and then transferring it to 
solution.  Plain cutoffs should not be used.  Their artifacts are well-known and 
modern simulations should not use them.  Shifted functions have discontinuities 
at their longest cutoff and neglect electrostatic interactions beyond this 
cutoff.  Your best option in solution (especially for highly-charged molecules 
like DNA) is PME.

-Justin

> Thanks,
> Majid
> ------------------------------------------------------------------------
> *From:* Justin A. Lemkul <jalemkul at vt.edu>
> *To:* Gromacs Users' List <gmx-users at gromacs.org>
> *Sent:* Thu, April 21, 2011 10:25:35 AM
> *Subject:* Re: [gmx-users] dna and cnt get distorted in md simulation
> 
> 
> 
> majid hasan wrote:
>  > first .mdp file is the original one, and modified .mdp is the one 
> where I made modifications, and I have tried both, and both lead to 
> stable structures for individual molecules, and distorted structures for 
> combined system.
>  >
> 
> The reason I asked is that the two are very different - one is for MD 
> and the other is for EM, and in some cases, many of the options are 
> irrelevant for either process so it is somewhat hard to deduce what 
> you're actually trying to accomplish with each, especially given the 
> differences.  It is best to only post the actual .mdp files you're using 
> and a description of the output corresponding to each.
> 
>  > In the first .mdp file, free energy calculations are turned off, but 
> even with this file, I get huge distortions in the shape of molecules.
> 
> So, the "first" .mdp file is the one that actually specifies an MD run, 
> not EM?  Or does "first" correspond to the order of the workflow?  It 
> might be best to focus on one process at a time, rather than trying to 
> troubleshoot both EM and MD with some arbitrary designators.
> 
>  > CNT, DNA atoms do form bonds in the simulation, but they lose their 
> shapes.
> 
> Bonds don't break or form during classical MD.  Any bonds "forming" or 
> "breaking" are simply a visualization artifact since you're not reading 
> a topology into the visualization software.
> 
>  From your description, it sounds like these simulations are being 
> conducted in vacuo?  I tend to suspect that is the source of the 
> problems.  Plain cutoffs for electrostatics lead to nasty artifacts and 
> the presence of a highly charged molecule (DNA) that has no shielding 
> between these charges is quite likely to become very distorted due to 
> its own intrinsic repulsion.
> 
> -Justin
> 
>  > Thank You,
>  > Majid
>  >
>  > ------------------------------------------------------------------------
>  > *From:* Justin A. Lemkul <jalemkul at vt.edu <mailto:jalemkul at vt.edu>>
>  > *To:* Discussion list for GROMACS users <gmx-users at gromacs.org 
> <mailto:gmx-users at gromacs.org>>
>  > *Sent:* Thu, April 21, 2011 4:02:57 AM
>  > *Subject:* Re: [gmx-users] dna and cnt get distorted in md simulation
>  >
>  >
>  >
>  > majid hasan wrote:
>  >  > Dear All,
>  >  >
>  >  > I minimized the energy of my CNT-DNA system with l-bfgs 
> integrator, and then ran the mdrun with integrator = md. I am using a 
> small ssDNA consisting of two residues only (66 atoms), and a small CNT 
> of about 80 atoms.
>  >  > My commands are:
>  >  > For energy minimization,
>  >  >
>  >  > grompp -f lbfgs.mdp -po mdoutlb.mdp -c gc.gro -p gc.top -o 
> gclb.tpr -maxwarn 20
>  >  > mdrun -s gclb.tpr -o gclb.trr -c gcoutlb.gro -e gclb.edr -g 
> mdlb.log -pd
>  >  >
>  >  > and then I ran molecular dyamics,
>  >  >
>  >  > grompp -f md.mdp -po mdoutmd.mdp -c gc.gro -p gc.top -o gcmd.tpr 
> -maxwarn 20
>  >  > mdrun -s gcmd.tpr -o gcmd.trr -c gcoutmd.gro -e gcmd.edr -g 
> mdmd.log -pd
>  >  >
>  >  > For individual DNA, and CNT alone, both produce reasonable 
> results, where molecule stays together and jiggles around. However on 
> combining the two molecules, both energy minimization and mdrun lead to 
> distorted structures: bond lengths don't remain fixed neither for CNT 
> nor for DNA, and atoms of both molecules get intertwined and produce a 
> mess. I tried two .mdp files.
>  >  > I got the  first .mdp from a colleague who used it for a simple 
> simulation of CNT only (without solvent, and any other molecule) . I 
> made few changes in this file after going through manual e.g enabled 
> free energy calculations, added "nsttcouple = -1" value, changed valued 
> of "comm_mode" from Angular to Linear, changed "ns_type" value from grid 
> to simple, changed "rcoulomb" and "rvdw" values from 0.9 to 1,
>  >  >  Both .mdp files are placed at following addresses:
>> >  http://phas.ubc.ca/~majid/md.mdp  (first .mdp file)
>  >
>  > Is this the file that has had the above modifications made to it for 
> MD?  If so, please post the actual file, not the unmodified one.
>  >
>> >  http://phas.ubc.ca/~majid/lbfgs.mdp (modified .mdp file)
>  >  >
>  >  > It seems to me that problem might be related to non-bonded 
> interaction because this is the significant difference between one and 
> two molecule system. Any help would be much appreciated.
>  >
>  > Why are you employing the free energy code?  It seems to me this 
> could be the source of your problems.  Each molecule alone is fine, but 
> then by decoupling van der Waals and Coulombic interactions between 
> them, you could be getting instability.
>  >
>  > Turn off the free energy options and see if you get stable trajectories.
>  >
>  > -Justin
>  >
>  >  > Thanks,
>  >  > Majid
>  >  >
>  >
>  > -- ========================================
>  >
>  > Justin A. Lemkul
>  > Ph.D. Candidate
>  > ICTAS Doctoral Scholar
>  > MILES-IGERT Trainee
>  > Department of Biochemistry
>  > Virginia Tech
>  > Blacksburg, VA
>  > jalemkul[at]vt.edu <http://vt.edu> <http://vt.edu> | (540) 231-9080
>>  http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>  >
>  > ========================================
>  > -- gmx-users mailing list    gmx-users at gromacs.org 
> <mailto:gmx-users at gromacs.org> <mailto:gmx-users at gromacs.org 
> <mailto:gmx-users at gromacs.org>>
>>  http://lists.gromacs.org/mailman/listinfo/gmx-users
>>  Please search the archive at 
> http://www.gromacs.org/Support/Mailing_Lists/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 
> <mailto:gmx-users-request at gromacs.org> 
> <mailto:gmx-users-request at gromacs.org 
> <mailto:gmx-users-request at gromacs.org>>.
>>  Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> 
> -- ========================================
> 
> 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 
> <mailto:gmx-users at gromacs.org>
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at 
> http://www.gromacs.org/Support/Mailing_Lists/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 
> <mailto:gmx-users-request at gromacs.org>.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

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

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

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

More information about the gromacs.org_gmx-users mailing list