Re: question about setting up FEP calculations

From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Tue Oct 15 2013 - 12:00:58 CDT

No, that short run is not correct, because it does not describe any physically meaningful state.

Now I suspect one thing: that the two end-point groups are clashing, and that causes instability. The only reason why that could happen is if you don't have the proper nonbonded exclusions, which could happen if you are running the memory-optimized version of NAMD.

To make sure, you could try running Alchemify on your PSF before running:
http://www.edam.uhp-nancy.fr/Alchemify/

Best,
Jerome

----- Original Message -----
> thanks again Jerome. Sorry my question was kind of hidden .
> Just to be clear,
> If I submit FEP calculations straight from the dual topology file, it
> wasn't running. So when I submitted the short run (~ 3ns) and then did FEP
> calculations it worked.
> So, as mentioned before, for this short run, the system has both hydrogen
> (disappearing atom) and methyl group (appearing) bonded to the carbon atom
> at the same time. I wasn't sure if this part (short run with all
> atoms(appearing & disappearing ) before FEP ) is correct or not?
>
> thanks,
> Shyno
>
>
> On Tue, Oct 15, 2013 at 12:29 PM, Jérôme Hénin <jerome.henin_at_ibpc.fr> wrote:
>
> > Sorry, there was a misunderstanding. Then what you were doing seems
> > correct, and I don't know why it didn't run properly.
> >
> > Jerome
> >
> >
> > ----- Original Message -----
> > > thanks again for your reply. Sorry for asking more questions. Please bear
> > > with me as I explain few details:
> > > This is exactly what I did. The dual topology was made in the beginning
> > (i
> > > used non equilibrated systems to make it). Then added counterions and
> > > water. After this the dual topology system was minimized and submitted
> > for
> > > a short run before FEP calculations. As mentioned before, for this short
> > > run, the system has both hydrogen (disappearing atom) and methyl group
> > > (appearing) bonded to the carbon atom at the same time during the 3 ns
> > run.
> > >
> > > Or you meant to say the dual topology has to be made using the
> > equilibrated
> > > systems?
> > >
> > > thanks again,
> > > Shyno
> > >
> > >
> > >
> > > On Tue, Oct 15, 2013 at 11:44 AM, Jérôme Hénin <jerome.henin_at_ibpc.fr>
> > wrote:
> > >
> > > > In general, it is simpler to setup the complete "dual topology" system
> > > > from the beginning, then add solvent and other components.
> > > >
> > > > Jerome
> > > >
> > > >
> > > > ----- Original Message -----
> > > > > Hello Jerome,
> > > > > Thank you very much for your reply.
> > > > > Ok, so I should use the equilibrated system for lambda=0 (this is
> > where I
> > > > > want to start). To this I can add the appearing atoms (lambda=1) also
> > > > after
> > > > > the equilibration run and then create the .fep file and so on.
> > > > > Initially, I was trying this approach, one difficulty I faced was
> > how to
> > > > > number the atoms after 99999.
> > > > > For example, when I add the appearing atoms to the pdb file
> > (lambda=0 )
> > > > > after equilibration, the seriel numbers change. This problem arises
> > from
> > > > > the water molecules. Maybe another option would be just to select
> > > > > everything except water (for lambda=0 after equilibration) and add
> > > > > appearing atoms and then solvate the system? Is this approach
> > correct?
> > > > >
> > > > > thanks again for your help,
> > > > > Shyno
> > > > >
> > > > >
> > > > >
> > > > > On Mon, Oct 14, 2013 at 3:32 PM, Jérôme Hénin <jerome.henin_at_ibpc.fr>
> > > > wrote:
> > > > >
> > > > > > Hi Shyno,
> > > > > >
> > > > > > There seems to be something wrong with your approach. If you do an
> > > > > > equilibration run prior to the FEP, you should do that at a
> > > > well-defined
> > > > > > lambda state, by enabling the alchemical options and setting
> > lambada to
> > > > > > either 0 or 1, depending on where you want to start from.
> > > > > >
> > > > > > Cheers,
> > > > > > Jerome
> > > > > >
> > > > > > ----- Original Message -----
> > > > > > > Dear all,
> > > > > > > I am doing FEP calculations. For my system, I am mutating more
> > than 1
> > > > > > > residue at the same time (all are same type of mutations). Since
> > > > this is
> > > > > > my
> > > > > > > first time doing free energy calculations, I am trying to set up
> > the
> > > > > > system
> > > > > > > as mentioned in the tutorial:
> > > > > > > "A tutorial to set up alchemical free energy perturbation
> > > > calculations in
> > > > > > > NAMD"
> > > > > > > As the first step, I am writing the .fep input file which
> > contains
> > > > all
> > > > > > toms
> > > > > > >
> > > > > > > (appearing, disappearing and non-changing atoms). Initially, I
> > was
> > > > doing
> > > > > > > FEP immediately after minimization. However it wasn't running.
> > So I
> > > > did
> > > > > > a
> > > > > > > short run (about 3 ns) after minimization. Then submitted FEP
> > runs,
> > > > this
> > > > > > > worked. I would like to confirm this approach is correct.
> > > > > > >
> > > > > > > When I did the short run (3 ns), the system has both
> > disappearing and
> > > > > > > appearing atoms. For eg. the hydrogen attached to a carbon atom
> > will
> > > > > > > disappear and a methyl group will appear. In my system I have
> > both
> > > > > > hydrogen
> > > > > > > and methyl atom bonded to the carbon atom at the same time during
> > > > the 3
> > > > > > ns
> > > > > > > run. Is this approach correct?
> > > > > > > Any help will be appreciated,
> > > > > > > thanks,
> > > > > > > Shyno
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Shyno Mathew
> > > > > > > PhD student
> > > > > > > Department of Chemical Engineering
> > > > > > > Columbia University
> > > > > > >
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Shyno Mathew
> > > > > PhD student
> > > > > Department of Chemical Engineering
> > > > > Columbia University
> > > > >
> > > >
> > >
> > >
> > >
> > > --
> > > Shyno Mathew
> > > PhD student
> > > Department of Chemical Engineering
> > > Columbia University
> > >
> >
>
>
>
> --
> Shyno Mathew
> PhD student
> Department of Chemical Engineering
> Columbia University
>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:46 CST