VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Mon Apr 12 2010 - 08:58:28 CDT

On Mon, 2010-04-12 at 15:45 +0200, Ajasja Ljubetič wrote:
> By IUPAC angles I meant the amino acid side-chain (chi_n) angles as
> defined by the IUPAC body by the rules defined here (this is quite an
> old document, but I believe the rules are still valid). The problem is
> choosing the "principal torsion angle" as defined here.
> Basically, heavier atoms have priority, so the principal torsion angle
> around one bond is the one formed by the heaviest atoms.
>
>
> I think I will just have to sit down and define the principal torsion
> angle of each amino acid by hand. Then I can define a TCL function
> that returns a list quadruples of atoms names comprising all the
> side-chain principal torsion angles and use that in a selection macro.

that sounds awfully complicated. since you have the mass information,
why not take the list of all dihedrals, sort them by the third entry
(lsort should be able to do that) and then weed out all the dihedrals
about the same bond that have lower mass. for that you just need to
generate a "mass map".

> Or even more likely, I will just export the 40000 frames I currently
> have to pdb files and analyse them using bioshell. This will be less
> efficient in term of processor time, but probably it will be quicker
> than developing a new tcl script in terms of my time.

only, if you don't have to do this or something similar again.
investing some effort into well written VMD/tcl code will always
pay off in the long run. even if only providing you with a library
of "solutions" for common problems that you don't have to figure
out the next time. many people are always looking for "the one and
only" solution for the problem at hand and are not thinking in
writing reusable components.

> Since atom names differ among various forcefields I think this would
> be difficult to implement in a consistent manner using topotools.

yes. name based processing would be bad. but masses are often available
and can be guessed based on some heuristics, if needed (topotools has
some features for that, too).

cheers,
   axel.

>
> thank you for your help,
> Ajasja
>
> On Mon, Apr 12, 2010 at 15:08, Axel Kohlmeyer <akohlmey_at_gmail.com>
> wrote:
> On Mon, 2010-04-12 at 13:38 +0200, Ajasja Ljubetič wrote:
> > Yes, I have tried that, actually. But I get too many
> dihedrals (not
> > just the IUPAC Chi1, Chi2 .. Chi4 but all possible dihedral
> angle
>
>
> you get all the dihedrals that you feed VMD. like all _good_
> programs, VMD does not make many assumptions about your data.
> it just processes what you give it to work with and then it
> is up to you to tell it how to trim it down. can you please
> explain what you understand under IUPACK dihedrals?
>
> if it is easy enough to implement, i could consider adding
> a special case version for that. or even add some commands
> to the high-level tools that allow to do the whole shebang.
>
> > combinations). I have not yet figured out how to filter
> them.
> > What is the first value in the list of dihedral angles? The
> 'topo
> > getdihedrallist -sel "protein and resid 18"' returns a list
> of lists
> > {unknown 215 217 223 224} {unknown 215 217 219 220} ....
> Does
> > "unknown" refer to the dihedral name? Or is the type meant
> to be
> > improper or dihedral angle?
>
>
> "unknown" refers to the force field type. in charmm class
> force field
> this type is not stored in the topology file but inferred from
> the
> atom type and the corresponding parameters are read from the
> parameter
> file that you have to feed the simulation software. with 'topo
> retypedihedrals' you can generate canonical dihedral types
> from
> the atom types (note that those are canonicalized, i.e. not
> simple
> the corresponding atom types concatenated. in other MD
> programs
> dihedrals are identified with distinct type names or numbers
> and
> thus VMD allows to store those.
>
> hope that helps,
> axel.
>
>
> > I'm using the 1.1 version of topotools. And I have loaded
> the molecule
> > from a psf/pdb combination.
> >
> >
> > Best regards,
> > Ajasja Ljubetič
> >
> > On Mon, Apr 12, 2010 at 12:43, Axel Kohlmeyer
> <akohlmey_at_gmail.com>
> > wrote:
> > On Mon, Apr 12, 2010 at 4:43 AM, Ajasja Ljubetič
> > <ajasja.ljubetic_at_gmail.com> wrote:
> > > Dear all,
> > > I hope this does not count as a cross post (as I
> have asked
> > a similar
> > > question on the NAMD mailing list as well).
> > > For each frame of my trajectory I would like to
> calculate
> > all the
> > > side-chain dihedral angles. Similar to the output
> of
> > > this webservice. Currently I'm planing to export
> each frame
> > of
> > > the trajectory to PDB and using the
> > excellent Bioshell to calculate dihedral
> > > angles, but probably it would be faster to do this
> without
> > all the temporary
> > > PDBs (I will have to analyse a lot of short 10 ns
> > trajectories).
> > > A possible way would be to add labels for all the
> > side-chains using "label
> > > add Dihedrals". Could anyone suggest a simpler
> way?
> >
> >
> > not sure about simpler, but definitely more
> efficient would be
> > to use tcl scripting in VMD.
> >
> > you can get the list of dihedrals from the topotools
> plugin
> > (or even use the generate those, in case you don't
> have
> > a suitable topology file with existing definitions.
> topotools
> > commands accept a -sel option that you can give a
> selection
> > string or a selection command as argument, which
> would
> > restrict the list to dihedrals fully contained in
> the
> > sidechain.
> >
> > then you can loop over those and either use the
> measure
> > or label command to compute the dihedral angle
> values
> > for your trajectory. probably label would be more
> efficient.
> >
> > cheers,
> > axel.
> >
> >
> > > Thank you,
> > > Ajasja Ljubetič
> > > Young researcher at Jozef Stefan Instiute,
> > > Slovenia, Ljubljana
> >
> >
> >
> >
> > --
> > Dr. Axel Kohlmeyer akohlmey_at_gmail.com
> > http://sites.google.com/site/akohlmey/
> >
> > Institute for Computational Molecular Science
> > Temple University, Philadelphia PA, USA.
> >
> >
>
>
> --
>
> Dr. Axel Kohlmeyer akohlmey_at_gmail.com
> http://sites.google.com/site/akohlmey/
>
> Institute for Computational Molecular Science
> Temple University, Philadelphia PA, USA.
>
>
>
> 

-- 
Dr. Axel Kohlmeyer  akohlmey_at_gmail.com
http://sites.google.com/site/akohlmey/
Institute for Computational Molecular Science
Temple University, Philadelphia PA, USA.