Hi,
I was manually adding a larger box as that was a suggested solution in the
list to get rid of the minimisation errors.
I ended up using cpptraj as suggested and got rid of the minimisation
stages. During this process, I realised that the errors were occurring
because I had imaged the original trajectory from which I was obtaining the
restart files. Using the original unimaged trajectory sorted things out.
Thanks to everyone for their comments. They have been very helpful!
On Tue, Aug 26, 2014 at 8:09 PM, Jason Swails <jason.swails.gmail.com>
wrote:
> On Tue, 2014-08-26 at 18:58 +0100, George Green wrote:
> > Just to clarify:
> >
> > 1) I ran a simulation of a ligand entering a protein binding pocket.
> >
> > 2) I needed to strip out frames ~0.4A apart as input starting structures
> > for different umbrella sampling windows.
>
> I'm confused what exactly you mean here. We usually use "strip" to mean
> remove atoms from a system (like stripping waters with the "strip"
> command in cpptraj or ParmEd). I think you mean that you need to
> extract certain frames (but not strip out any atoms), yes?
>
> > 3) Initially I was attempting to use cpptraj to try and put the box in.
> > However, I realised that Jason has a helpful script called
> > get_restarts.py that outputs restart files from the initial trajectory. I
> > opted to use this and manually modified the box information on the last
> > line to match that of the parameter file used in step1.
>
> I'm confused as to how the box was lost in the first place? Just a
> little FYI, I wrote get_restarts.py specifically because cpptraj ignored
> velocities if they happened to be present in the trajectory file (if you
> set ntwv = -1, for instance). But cpptraj does keep track of velocities
> now (I'm pretty sure, at least). So the only advantage to using
> get_restarts.py is that it's arguably simpler to use (much like a brick
> is easier to use than a smartphone).
>
> Or is the box not lost, but you are changing the box to match what the
> prmtop file has in it? And if so, why? tleap creates periodic boxes
> that are much too large, so I can't really imagine a situation where you
> would want to replace a box inside a coordinate file with a box from the
> prmtop (not even sander or pmemd use the box info in the prmtop
> anymore).
>
> >
> > 4) However, now I have a problem during minimisation which tends to stall
> > and quit for some of the restart files that form the basis of the various
> > US windows. I am currently trying to employ a more gentle two step
> > minimisation to try and avoid the minimiser stopping:
> >
> > ####################
> > step1 minimisation - hold all atoms with weak force
> > &cntrl
> > imin=1, ntmin=1, maxcyc=350, ncyc=200,
> > ntb=1, ntpr=5, ntwr=100000, drms=0.05,
> > cut=12.0, ntxo=2, ioutfm=1, NTR=1,
> > ntc=1, ntf=1,
> > ntb=1, ntp=0,
> > &end
> > Group input for restrained atoms
> > 0.5
> > RES 1 8866
> > END
> > END
> > ###################
>
> What's the point of this step? These structures come from an MD
> simulation, right? So the structure shouldn't be strained, which makes
> me think this step isn't doing anything.
>
> >
> > Then:
> >
> > ####################
> > step2 minimisation - unrestrained min
> > &cntrl
> > imin=1, ntmin=1, maxcyc=350, ncyc=200,
> > ntb=1, ntpr=5, ntwr=100000, drms=0.05,
> > cut=12.0, ntxo=2, ioutfm=1,
> > ntc=1, ntf=1,
> > ntb=1, ntp=0,
> > &end
> > &wt
> > type='END',
> > &end
> > DISANG=restraint_list_${frame}.dist
> > ###################
> >
> >
> > I'm hoping that you might have some suggestions regarding a better
> > minimsation protocol, or alternatively a better way to create the inpcrds
> > for umbrella sampling?
>
> I would suggest leaving the box information alone (unless I'm really
> missing what you're trying to do).
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
>
>
> _______________________________________________
> 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 Aug 27 2014 - 13:30:02 PDT
