VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Thu Feb 16 2012 - 10:34:58 CST

Neither of those tools accounts for periodicity, they (currently) assume
that the part of the structure you care about does not cross the periodic
boundaries.

Cheers,
  John

On Fri, Feb 17, 2012 at 01:27:38AM +0900, Aric Newton wrote:
> Axel,
>
> I originally built the model in Materials Studio. I exported it as a
> *.car/ *.mdf file pair; made a LAMMPS data file with the msi2lmp tool;
> ran the simulation; and wrote an unscaled trajectory in the LAMMPS
> (*.atom) format and an unwrapped trajectory in the GROMACS (*.xtc)
> format.
>
> I generated a *.psf from the lammps datafile and loaded the trajectories
> (a different model for both trajectory formats). I can see the H-bonds
> via the Graphical Representation> Hbonds and generate the data files and
> a multiplot graph with the Analysis>Hydrogen Bonds extension. I do
> still receive the "Warning) multiplot: Data vector empty, ignoring
> plot!" message. I accept it as a warning not an error.
>
> I assume that the tool is functioning as it should (input parameters at
> the end). What I would like to ask is if either of the H-bonds methods
> is calculating H-bonds across the periodic boundaries? I ask because
> the anion starts in the center and during the trajectory migrates toward
> the cell boundary. The output indicates that as the trajectory
> progresses, the number of H-bonds decreases when the anion is near the
> periodic boundary. The graphs of the H-bonds v. frames for the wrapped
> and unwrapped trajectories are not equivalent, but the decrease in the
> number of bonds is present in both (i.e. unwrapped atoms can't H-bond,
> wrapped atoms can H-bond, but H-bonds can't cross the cell boundary).
> Visually, the trajectory appears to confirm this (no H-bond
> representations cross the boundary). Is this an artifact of the way that
> I have loaded the trajectory or something the tool doesn't do?
>
> Thank you,
>
> Aric
>
> Parameters used in the calculation of hydrogen bonds:
> - Atomselection 1: resid 407
> - Atomselection 2: not resid 407
> - Update selections every frame: yes
> - Initial frame: 0
> - Frame step: 1
> - Final frame: 5000
> - Donor-Acceptor distance: 3.0
> - Angle cutoff: 20
> - Type: none
> - Write a file with H bond/frame data: yes
> - Filename: ohz_wrap_hbonds.dat
>
> Printing frame/hbond data to file... Done.
> Warning) multiplot: Data vector empty, ignoring plot!
>
>
>
> On Wed, 2012-02-15 at 19:14 -0500, Axel Kohlmeyer wrote:
> > On Wed, Feb 15, 2012 at 7:11 PM, Aric Newton <agnewton_at_berkeley.edu> wrote:
> > > Dear John and Axel,
> > >
> > > Thank you for your replies and the more detailed info on how the Hbond
> > > algorithm does what it does. The naming conventions appear to be key.
> > > Your tips also suggest I should revisit topotools docs as the guessatoms
> > > options have been expanded since my last reading.
> > >
> > > I can render the Hbonds through the Graphical Representations dialog but
> > > haven't had luck (yet) with the Analysis> Hydrogen Bonds extension. 5
> > > minutes after I hit <send> I realized that I have the data file in other
> > > formats (*.car/ *.mdf) and can convert it to a *.pdb file. I'll use
> > > that as a starting point and keep at it.
> >
> > aric,
> >
> > it'll be better to generate a .psf file instead of .pdb file
> > since the .psf doesn't contain coordinates, but the .pdb does.
> > you usually don't want those "parasitic" coordinates in
> > your analysis, right?
> >
> > cheers,
> > axel.
> >
> >
> > >
> > > Thanks again to you both.
> > >
> > > Aric
> > >
> > >
> > > On Wed, 2012-02-15 at 10:13 -0600, John Stone wrote:
> > >> Hi,
> > >> I thought I'd chime in and amplify Axel's comment regarding atom
> > >> names/types. The hbond visualization and analysis functions of VMD
> > >> look for atom names/types that they recognize, if you're using
> > >> a naming scheme that VMD doesn't understand, then it won't be able to
> > >> do the visualization/analysis you want. The simplest way to fix this
> > >> is to rename the atoms, and then do a "mol reanalyze" so VMD re-tags
> > >> the atoms it is able to recognize.
> > >>
> > >> Cheers,
> > >> John Stone
> > >> vmd_at_ks.uiuc.edu
> > >>
> > >> On Tue, Feb 14, 2012 at 04:36:13PM -0500, Axel Kohlmeyer wrote:
> > >> > dear aric,
> > >> >
> > >> > On Tue, Feb 14, 2012 at 3:22 PM, Aric Newton <agnewton_at_berkeley.edu> wrote:
> > >> > > Dear VMD users,
> > >> > >
> > >> > > I am trying to use the Hydrogen Bonds extension to assess the hydrogen
> > >> > > bonding of a hydroxyl anion in SPC/E water. I have run the simulation
> > >> > > in LAMMPS and am loading it into VMD using the sequence 1) topotools
> > >> > > readlammpsdata <file>; 2) topo guessatom element mass; and 3) loading
> > >> > > the lammps trajectory (*.atom format) into the molecule.
> > >> >
> > >> > better yet. try:
> > >> >
> > >> > topo guessatom lammps data
> > >> >
> > >> > to import/reconstruct as much VMD information from the
> > >> > LAMMPS data file as possible. that should set atom names
> > >> > as well.
> > >> >
> > >> >
> > >> > > The Hydrogen Bonds extension executes with what appears to be valid
> > >> > > input data, but returns a "Warning) multiplot: Data vector empty,
> > >> > > ignoring plot!" Obviously, the output is not reliable. However, there
> > >> > > is no overlap in atom selections. Each water molecule has a unique
> > >> > > resid but they all have the same molecule id: 0 and chain: X. I have
> > >> > > provided the Hydrogen Bonding scripts parameters from the Terminal
> > >> > > window below and the output from a TkConsole command line attempt as
> > >> > > well. I've looked through the archive to try to see if there is
> > >> > > something obvious I am missing, but I have yet to locate it.
> > >> > >
> > >> > > I have used the brH.pdb file in the /protein directory distributed with
> > >> > > VMD to test if the problem is user-error or something else. The
> > >> > > Graphical Representations> Drawing Method> Hbonds technique depicts
> > >> > > H-bonds in the brH.pdb model. Although,the TkConsole indicates that
> > >> > > there are 441 hbonds found and provides a list when I use the Hydrogen
> > >> > > Bonds Extension, the "Warning) muliplot: Data vector empty, ignoring
> > >> > > plot!" message persists and no graph is generated.
> > >> > >
> > >> > > I've tried this with both v.1.9 and the new v.1.9.1 of VMD on Ubuntu
> > >> > > 10.04 LTS and v.1.9 on Mac OS X 10.7. Are hydrogen bonds absent because
> > >> > > there are no "explicit" hydrogen atoms in the structure? The "topo
> > >> > > guessatom element mass" correctly assigns the elements and I can color
> > >> > > the model (correctly) by element. But there is no indication what the
> > >> > > element is when I pick the atoms in the Display (names=types). Where is
> > >> > > this info stored and how can I get the Hbonds tools to see it?
> > >> >
> > >> > each atom has various properties (name, type, element, mass, charge,
> > >> > x, y, z, resid, residue, resname, fragment, index, serial, ...).
> > >> > those are stored with the molecule topology information.
> > >> >
> > >> > this can be accessed from script level with atom selections.
> > >> > e.g. to give all atoms of type 1 the name "OW" you would do
> > >> >
> > >> > set sel [atomselect top {type 1}]
> > >> > $sel set name OW
> > >> > $sel delete
> > >> >
> > >> > and so on. the hbond plugin assumes a combination of
> > >> > PDB/NAMD/CHARMM/Amber naming conventions to
> > >> > identify hydrogens, and hbond donors and acceptors.
> > >> > this may not be 100% consistent with the Hbond representation.
> > >> >
> > >> > so if, using the more extended: topo guessatoms lammps data
> > >> > doesn't help to get you up to speed, you can just write a little
> > >> > script that renames the atoms by lammps type number to something
> > >> > that will be properly recognized by VMD and then write it
> > >> > out as a .psf file for later re-use.
> > >> >
> > >> > due to the way, how lammps strips of all information about
> > >> > conventional naming from its input, you have to resolve to
> > >> > these tricks. to do analysis.
> > >> >
> > >> > life is tough. :-(
> > >> >
> > >> > cheers,
> > >> > axel.
> > >> >
> > >> >
> > >> > >
> > >> > >
> > >> > >
> > >> > > Sorry if I have overloaded you with diagnostic info but I am stumped.
> > >> > > Thank you for reading.
> > >> > >
> > >> > > Sincerely,
> > >> > >
> > >> > > Aric
> > >> > >
> > >> > >
> > >> > >
> > >> > > Parameters used in the calculation of hydrogen bonds:
> > >> > > - Atomselection 1: type 3 4
> > >> > > - Atomselection 2: type 1 2
> > >> > > - Update selections every frame: yes
> > >> > > - Initial frame: 0
> > >> > > - Frame step: 1
> > >> > > - Final frame: 5001
> > >> > > - Donor-Acceptor distance: 3.0
> > >> > > - Angle cutoff: 20
> > >> > > - Type: none
> > >> > > - Write a file with H bond/frame data: no
> > >> > >
> > >> > >
> > >> > > Found 0 hbonds.
> > >> > > donor acceptor occupancy
> > >> > > Warning) multiplot: Data vector empty, ignoring plot!
> > >> > >
> > >> > > Alternate TkConsole output type 1 and type 2 are O(water) and H(water),
> > >> > > respectively.
> > >> > > Main console display active (Tcl8.5.6 / Tk8.5.6)
> > >> > > (NPT_iso500ps) 1 % set sel1 [atomselect top "type 1"]
> > >> > > atomselect20
> > >> > >>Main< (NPT_iso500ps) 2 % set sel2 [atomselect top "type 2"]
> > >> > > atomselect21
> > >> > >>Main< (NPT_iso500ps) 3 % measure hbonds 3.5 20 $sel2 $sel1
> > >> > > {} {} {}
> > >> > >>Main< (NPT_iso500ps) 4 %
> > >> > >
> > >> > >
> > >> > >
> > >> > >
> > >> >
> > >> >
> > >> >
> > >> > --
> > >> > Dr. Axel Kohlmeyer
> > >> > akohlmey_at_gmail.com http://goo.gl/1wk0
> > >> >
> > >> > College of Science and Technology
> > >> > Temple University, Philadelphia PA, USA.
> > >>
> > >
> > >
> >
> >
> >
> 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
http://www.ks.uiuc.edu/~johns/           Phone: 217-244-3349
http://www.ks.uiuc.edu/Research/vmd/       Fax: 217-244-6078