[gmx-users] Post-analysis of residue residue interactions

[CROUZY Serge 119222]

Yes that's a good idea
I didn't know about different energy groups 
I'll try that
Thanks a lot Mark 


-----Message d'origine-----
De : gromacs.org_gmx-users-bounces at maillist.sys.kth.se [mailto:gromacs.org_gmx-users-bounces at maillist.sys.kth.se] De la part de Mark Abraham
Envoyé : vendredi 3 février 2017 14:19
À : gmx-users at gromacs.org
Objet : Re: [gmx-users] Post-analysis of residue residue interactions

Hi,

You can have more than one energy-group pair-of-interest per rerun. That will compute more pairwise combinations that are not of interest, requiring more sophisticated use of gmx energy, but this will run a lot faster than the I/O you're saving. e.g.

energygrps = A ResInter1 ResInter2 ... ResInterN for some chunk size of N distinct residues of B.

IIRC there's a cap of 64 energy groups because of how they were used in the implementation of the old group scheme.

Mark

On Thu, Feb 2, 2017 at 2:41 PM CROUZY Serge 119222 <serge.crouzy at cea.fr>
wrote:

>
> Dear gromacs users
>
> We've run umbrella sampling simulations (with pull) of 
> proteinA/ProteinB dimer turning into two separate monomers We have 25 
> simulation windows each of 10 ns MD (1000 frames) and proteins 
> containing 132 residues We now want to calculate interaction energy 
> (VDW Elec) between each residu of B (1 by 1) and all residues of A (as 
> a whole) to understand which residues most contribute to the PMF
>
>  The way we do that is described below :
> for (( f=0; f<=25; f++ ))                                (all frames)
> for (( res =1; res<=132; res++ ))                 (all residues)
> make_ndx  ResInter                                 define residue number
> in B : Resinter
> gmx_mpi grompp                                 build tpr  define new
> energy groups energygrps  =  A ResInter
> mpirun -np 8 gmx_mpi mdrun -rerun umbrella$f.xtc                   rerun
> dynamics on the 1000 frames
> g_energy                                                    average energy
> between Resinter and total A protein
> done
> done
>
> that's 25*132*1000 = 33 10⁵ energy calculations and particularly time 
> consuming 132*25 = 3300 times rereading the trajectories
>
> I was wondering is there a more efficient way to do this ? For 
> instance reading each trajectory only once instead of 132 times That's 
> basically exchanging loop order
> for (( f=0; f<=25; f++ ))   (all frames)
>    mpirun -np 8 gmx_mpi mdrun -rerun umbrella$f.xtc
>    for (( res =1; res<=132; res++ ))
>      g_energy
>    done
> done
> but this does not seem possible because the interacting groups have to 
> be defined before calling md rerun
>
> Any idea would be greatly appreciated
>
> Thanx
>
> Serge Crouzy
>
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before 
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or 
> send a mail to gmx-users-request at gromacs.org.
--
Gromacs Users mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.