Hi,
On Tue, Mar 23, 2021 at 7:21 PM Ming Tang <m21.tang.qut.edu.au> wrote:
> MD with the above conditions initially produced same distance (indicating that the atom index is correct in cpptraj and in IGR) in dist_vs_t and test.dat until thousands of steps of simulation when the distance in dist_vs_t and test.dat are different.
To understand this, you need to understand how imaging works in Amber.
During the MD simulation, imaging (when iwrap=1) applies *only* to the
output trajectory, i.e. molecules are not actually "wrapped" when they
cross the unit cell boundary. So things like restraint distance do
*not* use the minimum image convention. This is why it is important to
use 'autoimage' on input coordinates prior to restarting a simulation
of multiple molecules with restraints - if one of the molecules was
imaged, the restraint distance will not account for that.
However, by default the 'distance' action in cpptraj *does* use the
minimum image convention. So I'm guessing what happens is in the
simulation, one of the molecules crosses a unit cell boundary. To test
this, calculate the distance again with cpptraj, but this time use the
'unwrap' action followed by 'distance' with the 'noimage' keyword,
e.g.
unwrap
distance d1 :178.N,C8,C9,C10 :179.N,C8,C9,C10 out test.dat noimage
I'm guessing the two distances will then match.
The long-term solution here is to use 'iwrap=0' (i.e. no imaging) when
you have NetCDF trajectory/restarts enabled. I think we need to start
making this the default behavior in Amber to avoid imaging issues.
-Dan
>
> Is there anything wrong with my control file?
>
> Thanks,
> Ming
> ________________________________
> From: Carlos Simmerling <carlos.simmerling.gmail.com>
> Sent: Tuesday, 23 March 2021 8:35 PM
> To: AMBER Mailing List <amber.ambermd.org>
> Subject: Re: [AMBER] Apply Lennar-Jones potential between small molecules
>
> Hi Ming,
> can you include your mdin file and also the cpptraj script? then people
> will be able to see what you're doing.
>
> On Tue, Mar 23, 2021 at 12:31 AM Ming Tang <m21.tang.qut.edu.au> wrote:
>
> > Hey Carlos,
> >
> > Thanks for your suggestion. The system shows the restraint is working as
> > two compounds initially closer than r2 angstroms were pulled away from each
> > other until they are more than r2 angstroms away due to the repulsive
> > force. But after the simulation goes on for a while, some compounds do not
> > repel each other even if there are close to each other. I extracted the
> > distance between two restrained groups from the output file generated by
> > DUMPAVE and calculated the distance via cpptraj distance command.
> > Initially, their values are same because the two molecules are in the same
> > box, but after a while (the length of the time depends on the selected
> > groups and the simulation), their values are different, e.g., the distance
> > calculated from distance command is smaller than r2, whereas the distance
> > from DUMPAVE is much larger than r2. Therefore, the two close groups are
> > not restrained. I have been outputting and calculating the distance per 1
> > step, and after restart the simulation multiple times, I came across the
> > above same issue. I am totally not sure what is wrong here.
> >
> > Thanks,
> > Ming
> >
> >
> > ________________________________
> > From: Carlos Simmerling <carlos.simmerling.gmail.com>
> > Sent: Friday, 19 March 2021 1:05 AM
> > To: AMBER Mailing List <amber.ambermd.org>
> > Subject: Re: [AMBER] Apply Lennar-Jones potential between small molecules
> >
> > I would not think about what happens after autoimage- that's not what is
> > going on in your simulation and the restraints will only be applied to what
> > happens during MD.
> > focus on the input structure that you give to sander or pmemd, and whether
> > you use imaging during MD or not. also on how you set up the pairs and
> > distances for the restraints.
> > can you simplify the problem to a single restraint that you think isn't
> > working, and explain what your restraint is, the initial coordinates for
> > that pair, and why you think it isn't working during MD?
> >
> >
> >
> > On Thu, Mar 18, 2021 at 10:53 AM Ming Tang <m21.tang.qut.edu.au> wrote:
> >
> > > Yes. I tried both iwrap=0 and 1, and the distance is same and not real
> > > (not equal to the distance after the cpptraj autoimage process)
> > > ________________________________
> > > From: Carlos Simmerling <carlos.simmerling.gmail.com>
> > > Sent: Thursday, 18 March 2021 8:35 PM
> > > To: AMBER Mailing List <amber.ambermd.org>
> > > Subject: Re: [AMBER] Apply Lennar-Jones potential between small molecules
> > >
> > > what do you mean by "before autoimage"? autoimage is a cpptraj command,
> > but
> > > it sounds like you are talking about what happens in the MD run.
> > > do you have iwrap=1 in your simulation?
> > >
> > >
> > > On Thu, Mar 18, 2021 at 12:15 AM Ming Tang <m21.tang.qut.edu.au> wrote:
> > >
> > > > Dear Carlos and David,
> > > >
> > > > Thank you very much for your guidance. Yes, I found that a positive rk2
> > > > does the right thing and the problem lies in how the system calculates
> > > the
> > > > distance between two IGRs. I found out that the system calculates the
> > > > distance based on the coordinate before autoimage, and therefore the
> > > small
> > > > molecules do not get close in the trajectory before the autoimage
> > > process.
> > > > I guess this comes from the periodic boundary condition setting.
> > > >
> > > > Is there a way to fix this?
> > > >
> > > > Many thanks,
> > > > Ming
> > > > ________________________________
> > > > From: Carlos Simmerling <carlos.simmerling.gmail.com>
> > > > Sent: Tuesday, 16 March 2021 11:41 PM
> > > > To: AMBER Mailing List <amber.ambermd.org>
> > > > Subject: Re: [AMBER] Apply Lennar-Jones potential between small
> > molecules
> > > >
> > > > On Tue, Mar 16, 2021 at 8:56 AM David A Case <david.case.rutgers.edu>
> > > > wrote:
> > > >
> > > > > On Tue, Mar 16, 2021, Carlos Simmerling wrote:
> > > > >
> > > > > >yes negative rk gives an inverted parabola, so it is repulsive if it
> > > is
> > > > > >outside the flat region.
> > > > >
> > > > > I'm a little confused here, and would recommend careful testing if
> > you
> > > > are
> > > > > setting rk2 to a negative number. I've not done that but it *sounds*
> > > > > dangerous. Probably a good idea to numerically plot the function vs.
> > > > > distance, and make sure you are getting the behavior you want.
> > > > >
> > > > >
> > > > Dave is right, this can be really dangerous since it goes to infinitely
> > > > negative energies. I've used it and technically it does work, but it's
> > > > probably much safer to use other options. I should have been more
> > careful
> > > > in giving caution along with my reply!
> > > >
> > > > re-reading the original question, a negative rk2 will make them
> > > ATTRACTIVE
> > > > below r2, and it doesn't sound like this is what you want. *For
> > repulsive
> > > > interaction below r2 you want a positive rk2.*
> > > > _______________________________________________
> > > > AMBER mailing list
> > > > AMBER.ambermd.org
> > > >
> > > >
> > >
> > https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!XwuuxffEWbPzL5ZpjA3qQaJ8iH75Wnx1PeCXQgwlVDUABFVmZE0Vxkz9p7xRCDPvM3A-$
> > <
> > https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!XwuuxffEWbPzL5ZpjA3qQaJ8iH75Wnx1PeCXQgwlVDUABFVmZE0Vxkz9p7xRCDPvM3A-$
> > >
> > > > _______________________________________________
> > > > AMBER mailing list
> > > > AMBER.ambermd.org
> > > >
> > >
> > https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!T4XmYozjpg_nUWYf5aiS384bigysFF4gMFp237urlpol8HaAUeK-PsSgtrES04TavSv4$
> > <
> > https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!T4XmYozjpg_nUWYf5aiS384bigysFF4gMFp237urlpol8HaAUeK-PsSgtrES04TavSv4$
> > >
> > > >
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > >
> > >
> > https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!T4XmYozjpg_nUWYf5aiS384bigysFF4gMFp237urlpol8HaAUeK-PsSgtrES04TavSv4$
> > <
> > https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!T4XmYozjpg_nUWYf5aiS384bigysFF4gMFp237urlpol8HaAUeK-PsSgtrES04TavSv4$
> > >
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > >
> > https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!S02kZLVTFmG6R7T7TLZ2U9LxpyC5VthOgS37EEyLzgYvW8T1seQGVgclaCcK862BNrv4$
> > <
> > https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!S02kZLVTFmG6R7T7TLZ2U9LxpyC5VthOgS37EEyLzgYvW8T1seQGVgclaCcK862BNrv4$
> > >
> > >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> >
> > https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!S02kZLVTFmG6R7T7TLZ2U9LxpyC5VthOgS37EEyLzgYvW8T1seQGVgclaCcK862BNrv4$
> > <
> > https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!S02kZLVTFmG6R7T7TLZ2U9LxpyC5VthOgS37EEyLzgYvW8T1seQGVgclaCcK862BNrv4$
> > >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!RMlX5NwA2QIUVndnhz5ehQcpHbC88r6UZ_LFQVZL7vTN-ZkvFdfgk50B_p6nsSbY_S_S$
> >
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!NVzLfOphnbDXSw!RMlX5NwA2QIUVndnhz5ehQcpHbC88r6UZ_LFQVZL7vTN-ZkvFdfgk50B_p6nsSbY_S_S$
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Mar 24 2021 - 07:00:04 PDT
