Re: "atoms moving too fast" error in bilayer simulation

From: Himanshu Khandelia (hkhandel_at_memphys.sdu.dk)
Date: Thu Feb 08 2007 - 15:04:12 CST

Hi Peter,

The bilayer got equilibrated finally. I had increased the pair-list-dist
as per your recommendation, and minimized a lot more (~ 100,000 steps).

Thank you,

-Himanshu

On Sun, 4 Feb 2007, Peter Freddolino wrote:

> Ok, sorry for my misunderstanding. Given that I'd also suggest two
> additional things:
> -set your pairlistdist to be at least 2-3A larger than your cutoff (to
> avoid atoms drifting into nonbond range)
> -Expand your pme grid to be at least 1 point per angstrom
>
> You'll then just need to look at the trajectories, preferably with the
> proper wrapping on. I'll be interested to hear what's going on, because
> I've never seen a problem like this with the pope bilayer patch.
>
> Peter
>
> Himanshu Khandelia wrote:
> > I am using the the vmd POPE bilayer patches. Will try to look at the dcd
> > with a write frequeny of 1 and update,
> >
> > Thank you,
> >
> > -himanshu
> >
> >
> >
> > On Sat, 3 Feb 2007, Peter Freddolino wrote:
> >
> >
> >> I should clarify a bit more -- you're not using one of the POPC or POPE
> >> bilayers that are built in to VMD right? I remember you were recently
> >> adding in a different lipid with the membrane plugin. Are you currently
> >> using your new bilayer or one of the ones that comes with VMD? In any
> >> case, you should rerun the dynamics portion with a dcd frequency of 1,
> >> and have a look at the atoms that namd reports as moving too fast to see
> >> what the problem is (remember that NAMD atoms are 1-indexed, whereas VMD
> >> atoms are 0-indexed).
> >> Peter
> >>
> >> Himanshu Khandelia wrote:
> >>
> >>> The membrane was made from the vmd membrane plugin. When I said bad
> >>> contacts, I meant that the initial energy of the system was diverging
> >>> (99999999.9999)
> >>>
> >>> Towards the end of minimization, the gradient tolerance is 0.015, but the
> >>> VDW energy term is still infinite (99999999.9999). I guess I need more
> >>> minimization ?
> >>>
> >>> -Himanshu
> >>>
> >>>
> >>> On Fri, 2 Feb 2007, Peter Freddolino wrote:
> >>>
> >>>
> >>>
> >>>> Hi Himanshu,
> >>>> is this using a home-built lipid bilayer? The bilayers that come with
> >>>> VMD's membrane plugin should not have any bad contacts...
> >>>>
> >>>> What is the last gradient tolerance reported by NAMD during the
> >>>> minimization process? If it's much larger than, say, 1, you may just not
> >>>> be minimizing the system for long enough to get rid of all the clashes,
> >>>> so the system explodes as soon as you allow it to move freely. Have you
> >>>> looked to see if the bad contacts appear to be eliminated at the end of
> >>>> minimization?
> >>>>
> >>>> Peter
> >>>>
> >>>> Himanshu Khandelia wrote:
> >>>>
> >>>>
> >>>>> Hi,
> >>>>>
> >>>>> I am using VMD's membrane plugin to construct a membrane, solvating it,
> >>>>> and then trying to equilibrate it. The membrane generated by VMD does have
> >>>>> a lot of bad contacts, but I thought NAMD would handle them.
> >>>>>
> >>>>> I keep getting the "atoms moving too fast" error as soon as I start
> >>>>> heating the bilayer after minimzation. I have tried the following, but
> >>>>> nothing works so far:
> >>>>>
> >>>>> - heating in very small increments (~ 3 degrees every 3000 steps)
> >>>>> - larger cutoffs (16 A)
> >>>>> - with/without constraints on head groups
> >>>>> - larger box sizes (than suggested by the minmax of water)
> >>>>>
> >>>>> Any suggestions into what might be going wrong ? The config file is
> >>>>> attached with this email,
> >>>>>
> >>>>> Thank you !
> >>>>>
> >>>>> -Himanshu
> >>>>> ------------------------------------------------------------------------
> >>>>>
> >>>>> set MOL pope-water
> >>>>>
> >>>>> set temp 500
> >>>>>
> >>>>> set xbox 70
> >>>>> set ybox 70
> >>>>> set zbox 94
> >>>>>
> >>>>> structure ${MOL}.psf
> >>>>> coordinates ${MOL}.pdb
> >>>>> temperature 0
> >>>>> parameters par_all27_prot_lipid.prm
> >>>>> paraTypeCharmm on
> >>>>>
> >>>>> outputEnergies 100
> >>>>> outputTiming 100
> >>>>> xstFreq 100
> >>>>> dcdFreq 500
> >>>>> wrapAll on
> >>>>>
> >>>>> timestep 1
> >>>>> nonBondedFreq 2
> >>>>> fullElectFrequency 4
> >>>>> stepsPerCycle 20
> >>>>>
> >>>>> exclude scaled1-4
> >>>>> 1-4scaling 1.0
> >>>>> switching on
> >>>>> switchDist 14
> >>>>> cutoff 16
> >>>>> pairlistdist 16
> >>>>>
> >>>>> cellBasisVector1 $xbox 00.00 00.00
> >>>>> cellBasisVector2 00.00 $ybox 00.00
> >>>>> cellBasisVector3 00.00 00.00 $zbox
> >>>>> cellOrigin -0.58 -0.78 0.056
> >>>>>
> >>>>> Pme on
> >>>>> PmeGridsizeX 60
> >>>>> PmeGridsizeY 60
> >>>>> PmeGridsizeZ 96
> >>>>>
> >>>>> #########################################
> >>>>> # CONSTRAINTS
> >>>>> #########################################
> >>>>> constraints on
> >>>>> consRef ${MOL}-cons.pdb
> >>>>> consKFile ${MOL}-cons.pdb
> >>>>> consKCol B
> >>>>>
> >>>>> #########################################
> >>>>> # PRESSURE AND TEMPERATURE CONTRON
> >>>>> #########################################
> >>>>>
> >>>>> langevin on
> >>>>> langevinDamping 5
> >>>>> langevinTemp $temp
> >>>>> langevinHydrogen no
> >>>>>
> >>>>> useflexiblecell yes
> >>>>> langevinPiston on
> >>>>> langevinPistonTarget 1.01325
> >>>>> langevinPistonPeriod 100
> >>>>> langevinPistonDecay 50
> >>>>> langevinPistonTemp $temp
> >>>>>
> >>>>> binaryoutput off
> >>>>> outputname ${MOL}-equil
> >>>>>
> >>>>> # GET INIT. ENERGY
> >>>>> minimize 0
> >>>>>
> >>>>> ###########
> >>>>> # MINIMIZE
> >>>>> ###########
> >>>>>
> >>>>> minimize 10000
> >>>>> output ${mol}-cons
> >>>>>
> >>>>> #######
> >>>>> # HEAT
> >>>>> #######
> >>>>>
> >>>>> set tem 1
> >>>>> while { $tem <= 500 } {
> >>>>> langevinTemp $tem
> >>>>> run 3000
> >>>>> set tem [expr $tem + 3]
> >>>>> }
> >>>>>
> >>>>>
> >>>>>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:20:02 CST