Aha!! My colleague James Robertson and I found a solution to this problem.
If I insert a newline in the midst of the long sequence, then the segfault
doesn't occur. So once the "sequence" command gets longer than 200
characters or so, I can start a new line and the leaprc script will go
though OK.
It would seem that the buffer that reads the leap commands is not large
enough to accommodate really long commands in the leaprc file when they are
on a single line. For now, it's no problem for me to just add these
newlines.
Thanks for your help everyone!
-Lane
On Thu, Sep 22, 2016 at 11:51 AM, Daniel Roe <daniel.r.roe.gmail.com> wrote:
> Hi,
>
> On Thu, Sep 22, 2016 at 11:21 AM, Lane Votapka <lvotapka100.gmail.com>
> wrote:
> >
> > And as for using something else to build the input structure, I was
> > thinking the exact same thing. Do you know of another tool that can
> > generate a PDB of an unfolded structure? I was considering just whipping
>
> You could try using the 'makestructure' command in cpptraj to set
> backbone angles to PP2 (something like 'makestructure pp2:1-X', where
> X is last residue). The command doesn't do any checking for clashes
> but it could be a good start. Alternatively you could try the
> 'permutedihedrals' command with the 'random' option which does do some
> basic clash checking (see manual for more details).
>
> -Dan
>
> > one up myself (I already have a lot of the code to write a pdb file based
> > on a z-matrix). All I need is the z-matrices of the amino acids that
> tleap
> > uses to generate the unfolded sequence. Do you know where I could obtain
> > those z-matrices?
> >
> > -Lane
> >
> > On Thu, Sep 22, 2016 at 11:17 AM, Carlos Simmerling <
> > carlos.simmerling.gmail.com> wrote:
> >
> >> the structure might not be very relevant, but I've found that structures
> >> built in this way typically respond fine to minimization and MD. It
> might
> >> still be a good idea to use other methods to build input structures,
> since
> >> they would let you use a pdb input file and not crash the routine that
> >> builds structure from sequence.
> >> carlos
> >>
> >> On Thu, Sep 22, 2016 at 11:14 AM, Lane Votapka <lvotapka100.gmail.com>
> >> wrote:
> >>
> >> > Hi Bill,
> >> >
> >> > Thanks for your suggestion, but I'm afraid that running on OSX won't
> be
> >> > useful for my particular situation. I need to be able to run tleap on
> >> Linux
> >> > for the task I'm trying to accomplish.
> >> >
> >> > Yes, I realize that the structure generated by "sequence" will not be
> >> > physically realistic, but that's on purpose. I'm trying to generate a
> >> > completely unfolded structure and I run minimizations before starting
> any
> >> > sort of simulations. I guess it's understandable though that the bug
> has
> >> > gone unnoticed before now.
> >> >
> >> > Do you guys think that I can expect the bug to be addressed anytime
> soon,
> >> > or should I look for other solutions to this problem beyond tleap?
> >> >
> >> > -Lane
> >> >
> >> > On Wed, Sep 21, 2016 at 8:34 PM, Bill Ross <ross.cgl.ucsf.edu> wrote:
> >> >
> >> > > With the amber14 release of leap, it didn't coredump on OSX for me.
> >> > >
> >> > > You realize your structure is meaningless even if you can generate
> it,
> >> > > and will likely crash any program you feed it to, because 'sequence'
> >> > > does no steric checking or adjustment as it tacks residues together.
> >> > > That's why the bug you have likely found hasn't been seen before.
> >> > >
> >> > > Bill
> >> > >
> >> > >
> >> > > On 9/21/16 5:12 PM, Daniel Roe wrote:
> >> > > > Also, the command
> >> > > >
> >> > > >> sys = combine { mol_0 }
> >> > > > seems redundant if you're not actually combining it with anything.
> >> Try
> >> > > > just using mol_0; see if that helps.
> >> > > >
> >> > > > -Dan
> >> > > >
> >> > > > On Wed, Sep 21, 2016 at 8:11 PM, Daniel Roe <
> daniel.r.roe.gmail.com>
> >> > > wrote:
> >> > > >> Hi,
> >> > > >>
> >> > > >> Have you tried tleap from AmberTools 16?
> >> > > >>
> >> > > >> -Dan
> >> > > >>
> >> > > >> On Wed, Sep 21, 2016 at 4:31 PM, Lane Votapka <
> >> lvotapka100.gmail.com>
> >> > > wrote:
> >> > > >>> Hi AMBER developers and users,
> >> > > >>>
> >> > > >>> When I run the following leaprc in the tleap that comes with
> >> Amber14,
> >> > > the
> >> > > >>> program fails in a segfault:
> >> > > >>>
> >> > > >>> set default PBradii mbondi3
> >> > > >>> source leaprc.ff14SBonlysc
> >> > > >>> source leaprc.gaff
> >> > > >>> mol_0 = sequence { NMET THR HIE GLN THR HIE ALA TYR HIE MET VAL
> ASN
> >> > > PRO SER
> >> > > >>> PRO TRP PRO LEU THR GLY ALA LEU SER ALA LEU LEU MET THR SER GLY
> LEU
> >> > > THR MET
> >> > > >>> TRP PHE HIE PHE ASN SER MET THR LEU LEU MET ILE GLY LEU THR THR
> ASN
> >> > > MET LEU
> >> > > >>> THR MET TYR GLN TRP TRP ARG ASP VAL ILE ARG GLU SER THR PHE GLN
> GLY
> >> > > HIE HIE
> >> > > >>> THR PRO ALA VAL GLN LYS GLY LEU ARG TYR GLY MET ILE LEU PHE ILE
> ILE
> >> > > SER GLU
> >> > > >>> VAL LEU PHE PHE THR GLY PHE PHE TRP ALA PHE TYR HIE SER SER LEU
> ALA
> >> > > PRO THR
> >> > > >>> PRO GLU LEU GLY GLY CYS TRP PRO PRO THR GLY ILE HIE PRO LEU ASN
> PRO
> >> > > LEU GLU
> >> > > >>> VAL PRO LEU LEU ASN THR SER VAL LEU LEU ALA SER GLY VAL SER ILE
> THR
> >> > > TRP ALA
> >> > > >>> HIE HIE SER LEU MET GLU GLY ASP ARG LYS HIE MET LEU GLN ALA LEU
> PHE
> >> > > ILE THR
> >> > > >>> ILE THR LEU GLY VAL TYR PHE THR LEU LEU GLN ALA SER GLU TYR TYR
> GLU
> >> > > ALA PRO
> >> > > >>> PHE THR ILE SER ASP GLY VAL TYR GLY SER THR PHE PHE VAL ALA THR
> GLY
> >> > > PHE HIE
> >> > > >>> GLY LEU HIE VAL ILE ILE GLY SER THR PHE LEU ILE VAL CYS PHE PHE
> ARG
> >> > > GLN LEU
> >> > > >>> LYS PHE HIE PHE THR SER ASN HIE HIE PHE GLY PHE GLU ALA GLY ALA
> TRP
> >> > > TYR TRP
> >> > > >>> HIE PHE VAL ASP VAL VAL TRP LEU PHE LEU TYR VAL SER ILE TYR TRP
> TRP
> >> > GLY
> >> > > >>> CSER }
> >> > > >>>
> >> > > >>> translate mol_0 { 0.0 0.0 0.0 }
> >> > > >>> sys = combine { mol_0 }
> >> > > >>> check sys
> >> > > >>> saveAmberParm sys system.top system.mdcrd
> >> > > >>> quit
> >> > > >>>
> >> > > >>> The output says:
> >> > > >>>
> >> > > >>> ERROR: syntax error
> >> > > >>> Segmentation fault (core dumped)
> >> > > >>>
> >> > > >>> If I enter the commands one by one, the segfault occurs on the
> >> > > "sequence"
> >> > > >>> command.
> >> > > >>>
> >> > > >>> If I change the sequence, but keep it the same length, the error
> >> > still
> >> > > >>> occurs. If I shorten the sequence, the segfault does not occur.
> I
> >> can
> >> > > even
> >> > > >>> use a long sequence of alanines and the segfault still happens.
> >> > > >>>
> >> > > >>> This makes me think that there is a memory overflow or something
> >> with
> >> > > >>> longer entries to the "sequence" command in the tleap code.
> >> > > >>>
> >> > > >>> Best,
> >> > > >>> -Lane
> >> > > >>>
> >> > > >>> --
> >> > > >>> Lane Votapka Ph.D.
> >> > > >>> Postdoctoral Associate, Dill Laboratory 115G
> >> > > >>> Laufer Center for Physical and Quantitative Biology
> >> > > >>> Stony Brook University
> >> > > >>> Stony Brook, NY 11794
> >> > > >>> _______________________________________________
> >> > > >>> AMBER mailing list
> >> > > >>> AMBER.ambermd.org
> >> > > >>> http://lists.ambermd.org/mailman/listinfo/amber
> >> > > >>
> >> > > >>
> >> > > >> --
> >> > > >> -------------------------
> >> > > >> Daniel R. Roe
> >> > > >> Laboratory of Computational Biology
> >> > > >> National Institutes of Health, NHLBI
> >> > > >> 5635 Fishers Ln, Rm T900
> >> > > >> Rockville MD, 20852
> >> > > >> https://www.lobos.nih.gov/lcb
> >> > > >
> >> > > >
> >> > >
> >> > >
> >> > > _______________________________________________
> >> > > AMBER mailing list
> >> > > AMBER.ambermd.org
> >> > > http://lists.ambermd.org/mailman/listinfo/amber
> >> > >
> >> >
> >> >
> >> >
> >> > --
> >> > Lane Votapka Ph.D.
> >> > Postdoctoral Associate, Dill Laboratory 115G
> >> > Laufer Center for Physical and Quantitative Biology
> >> > Stony Brook University
> >> > Stony Brook, NY 11794
> >> > _______________________________________________
> >> > 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
> >>
> >
> >
> >
> > --
> > Lane Votapka Ph.D.
> > Postdoctoral Associate, Dill Laboratory 115G
> > Laufer Center for Physical and Quantitative Biology
> > Stony Brook University
> > Stony Brook, NY 11794
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
> --
> -------------------------
> Daniel R. Roe
> Laboratory of Computational Biology
> National Institutes of Health, NHLBI
> 5635 Fishers Ln, Rm T900
> Rockville MD, 20852
> https://www.lobos.nih.gov/lcb
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Lane Votapka Ph.D.
Postdoctoral Associate, Dill Laboratory 115G
Laufer Center for Physical and Quantitative Biology
Stony Brook University
Stony Brook, NY 11794
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 22 2016 - 09:30:03 PDT
