Targeted molecular dynamics

[Qinghua]

Hello everyone,

Is it possible to do targeted molecular dynamics simulations using gromcacs interfaced with plumed?
I know that amber and namd can do this kind simulations, but we did the other simulations for the same
project with gromacs and plumed. Thus it is better to insist on the same programs. Thanks very much!


All the best,
Qinghua

[Giovanni Bussi]

Yes,

use RMSD collective variable coupled with a MOVINGRESTRAINT. You can either increase KAPPA or decrease AT with time (different flavours of targeted md)

Giovanni

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/dc913292-c295-4b99-83ce-671ed5e833e0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[qingh...@gmail.com]

Hello Prof. Bussi,

I tried to run TMD with Gromacs 4.6.7 interfaced with plumed2.1, but I failed to get it running.

My aim is to do TMD simulation to drive a loop from the close state to open state. The backbone atoms of the loop at open state were chosen as the
reference coordinates.  And here is my plumed file:

# #
     UNITS LENGTH=A TIME=ps ENERGY=kcal/mol
#
        chainA: GROUP ATOMS=1-8258
        WHOLEMOLECULES ENTITY0=chainA
#
     rmsd: RMSD REFERENCE=reference.pdb TYPE=OPTIMAL
   restraint: ...
            MOVINGRESTRAINT
           ARG=rmsd
    AT0=0.9   STEP0=0            KAPPA0=0
    AT1=0.0   STEP1=250000       KAPPA1=100   
 ...
#
 PRINT STRIDE=10 ARG=* FILE=colar.dat

AT0 was set as 0.9 angstrom as I calculated the RMSD of the backbone atoms of the loop between close and open states, it is around 0.9 angstrom. But the my simulation crashed several seconds after the simulation started. And the first line I got in the colar.dat file is:

#! FIELDS time rmsd restraint.bias restraint.force2 restraint.rmsd_cntr restraint.rmsd_work restraint.rmsd_kappa
 0.000000 -nan -nan -nan 0.000000 -nan 0.000000

It seems that there is something wrong about the simulation, but I could not figure it out. The simulation ran successfully if I comment out the plumed input flag.
Could you please help me with fixing the problem here? Thanks very much!

All the best,
Qinghua

On Monday, January 16, 2017 at 11:37:04 AM UTC+1, Giovanni Bussi wrote:

Yes,

use RMSD collective variable coupled with a MOVINGRESTRAINT. You can either increase KAPPA or decrease AT with time (different flavours of targeted md)

Giovanni

On Mon, Jan 16, 2017 at 11:35 AM, Qinghua <qingh...@gmail.com> wrote:

Hello everyone,

Is it possible to do targeted molecular dynamics simulations using gromcacs interfaced with plumed?
I know that amber and namd can do this kind simulations, but we did the other simulations for the same
project with gromacs and plumed. Thus it is better to insist on the same programs. Thanks very much!


All the best,
Qinghua

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.

[Giovanni Bussi]

Try without MOVINGRESTRAINT to monitor the RMSD computed by plumed.

Giovanni

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/10b3c608-eefd-4a06-856b-87305f9e8b5f%40googlegroups.com.

[Massimiliano Bonomi]

Hi!

Have you correctly identified the atoms you want to use for RMSD alignment and calculation in the pdb file?
Those should be marked in the beta and occupancy columns of the pdb, as indicated in the manual.

Max

> To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/10b3c608-eefd-4a06-856b-87305f9e8b5f%40googlegroups.com.

[qingh...@gmail.com]

Dear Max,

Thanks very much for your suggestion, it helped me fix the problem.
At the beginning, I only make the atom number in the reference pdb file match the the ones in the whole system.
I didn't check the beta and occupancy columns.

For the RMSD calculation, both the beta and occupancy columns are set as 1.00 normally, right?


All the best,
Qinghua

[qingh...@gmail.com]

Dear Prof. Bussi,

Thanks very much for your fast respond. I tried to run without MOVINGRESTRAINT, but I got nothing.

Max's reply helped me fix the problem!


All the best,
Qinghua

[Massimiliano Bonomi]

Well, not necessarily both of them.
The atoms you want to align the system on might be different from those you want to use to calculate the deviation.
It depends on the kind of problem you want to address.
For example, if you are studying ligand binding to a protein, you might want to align
the system on the protein positions, and then calculate the RMSD only on the ligand atoms.

Max

> To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/dbea7ce5-d209-4a07-9a54-a91b7f9bf5b5%40googlegroups.com.

[qingh...@gmail.com]

Dear Max,

Thanks very much for your explanation.

Actually, this is also one question I want to ask! In my case, it is about a loop opening from the close state to the open state.
Then, I guess that I need to align the whole protein, then calculate the RMSD only on the loop, right?

For your example, I still didn't get how to set the beta and occupancy differently for the protein and ligand for alignment and RMSD calculation, respectively.
Could you please be more detailed? I fixed the problem by changing all the occupancy to 1.00.


Another question, as TMD is one special case of steered MD, is it reasonable to derive the PMF based on multiple TMD run?


All the best,
Qinghua

[Massimiliano Bonomi]

> On Apr 3, 2017, at 17:34, qingh...@gmail.com wrote:
>
> Dear Max,
>

> Thanks very much for your explanation.
>
> Actually, this is also one question I want to ask! In my case, it is about a loop opening from the close state to the open state.
> Then, I guess that I need to align the whole protein, then calculate the RMSD only on the loop, right?

There is not a secret recipe to carry out this calculation. If you have the structure of the open and close state,
I would measure with VMD or other tools, the RMSD of the open state from the close state using two definitions:

1) aligning on the proteins, calculating on the loop atoms
2) aligning and calculating on the loops atoms.

It might very well happen that with the second definition, you would not see a huge difference in RMSD between the open
and close state. In this case, I would use the first definition. But I am sure there are many examples in the literature you can
learn a lot from!

>
> For your example, I still didn't get how to set the beta and occupancy differently for the protein and ligand for alignment and RMSD calculation, respectively.
> Could you please be more detailed? I fixed the problem by changing all the occupancy to 1.00.

You can find detailed explanations in our manual:

https://plumed.github.io/doc-v2.3/user-doc/html/_r_m_s_d.html

>
>
> Another question, as TMD is one special case of steered MD, is it reasonable to derive the PMF based on multiple TMD run?

Not an expert of TMD, but I suppose yes, you can.

Max

> To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/32700e42-2725-489e-86bd-dcf70791b550%40googlegroups.com.

[qingh...@gmail.com]

Dear Max,

Thanks very much for your detailed explanation and suggestions. I will go through the manual to figure it out.


All the best,
Qinghua