Re: [Jmol-users] auto-orientation?

[Thomas S. <tho...@gm...>]

To follow up on my question from last week concerning "auto-orientation",
Rolf Huehne very graciously pointed me in the direction of a solution that
he had already crafted for JenaLib.  He uses the dimensions of the bounding
box, selected for the atoms of interest, and arrays the longest two vectors
in the plane perpendicular to the viewing axis.  This has the effect of
making the majority of the molecule visible to the viewer.

I think that this is a fast, elegant and effective solution would very much
like to see it included in Jmol as a selectable function -- perhaps even
added to the pop-up menu?  Rolf has indicated his support for this as well
(ie, he is happy to donate the code snippet).

As an aside, Jena3D is utilizing this for macromolecules & I have just
implemented it locally for small, organic molecules and both situations
appear to work quite acceptably well.  I have not tested it on molecules
such as intermetallics or NaCl, but don't expect any difference.

Cheers,
Tom



On Sat, Feb 13, 2010 at 3:04 AM, Rolf Huehne <rh...@fl...> wrote:

> Thomas Stout schrieb:
> > Hi All --
> >
> > Before I embark on re-inventing yet another wheel, I thought I would
> > ask: has anyone already created a script for choosing an optimal
> > orientation in which to display a given molecule?  I am thinking in
> > terms of minimizing the overlap of atoms along the vector of view (z
> > axis), i.e., the best fit of atom coordinates to the xy plane...
> >
> It is quite difficult to define an optimal orientation. Even if you do
> it manually you have to make compromises and it depends on what you want
> to show. Since many PDB structures are oriented initially rather bad for
> viewing, we tried to find a way of optimizing the view for our
> automatically generated images for all PDB structures in the JenaLib
> (http://www.fli-leibniz.de/IMAGE.html). I first tried to use distance
> measurements to determine a better orientation. But this was much too
> slow within Jmol. So finally we ended up using the boundbox plane sizes.
> The method is very fast and often results in a better or equally good
> orientation. I currently don't have time to explain it in more detail
> but you can take a look at it in our Jena3D viewer, e.g.:
> http://www.imb-jena.de/cgi-bin/3d_mapping.pl?CODE=1deh
> The structure is first rotated 180 degrees and then reoriented
> automatically. You can compare it to the initial orientation by entering
> "reset" into the command input field. You can take a look at the
> mechanism by looking at the scripts behind the standard view buttons.
>
> Regards,
> Rolf
>
> <https://lists.sourceforge.net/lists/listinfo/jmol-users>
>