[Chimera-users] User-defined graphics in chimera: NMR tensor plotting

[Elaine Meng]

Hello Albert,
Is it possible to write the information as "volume data" (a 3D grid of values) such that isosurfaces would give the appearance that you want?  

Examples of this type of data include electron density maps from crystallography and EM, and electron localization function and spin density maps from QM calculations.  I often see published images of orbitals (drawn in Chimera) that look similar to the image you sent, except they are smooth isosurfaces.  I wanted to send a link but none of the examples I know are in open-access papers.

Maybe it would work with values on the grid -1, 0, 1 with negatives enclosed in one color of isosurface and positives enclosed in the other color.  However, I don't know how hard it would be to generate this from the information you have, or which format of volume data would be the simplest.

Chimera reads many different volume data formats:
<http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/filetypes.html#volume>

I hope this helps,
Elaine
-----
Elaine C. Meng, Ph.D.                       
UCSF Chimera(X) team
Department of Pharmaceutical Chemistry
University of California, San Francisco

> On Aug 28, 2020, at 1:54 AM, Albert Smith-Penzel <Albert.Smith-Penzel at medizin.uni-leipzig.de> wrote:
> 
> Hello,
> 
> I have a question about improving user defined graphics in Chimera.
> 
> For background (maybe not necessary for answering):
> I’m working with nuclear magnetic resonance and MD simulation, and so I want to plot dynamically averaged spin-spin (H–C) dipole tensors, obtained from MD, on top of the corresponding H–C bonds. I’ve been relatively successful, but while one has a very high-resolution picture of the molecule itself, the tensors themselves are visibly triangulated.
> 
> How I get this plot:
> Each tensor is expressed first in spherical coordinate (theta, phi, r). In python, I run Delauney from scipy.spatial on theta and phi to get a list of vertices that can be used for triangulation of the tensor. Then, for each tensor, I convert theta, phi, and r into cartesian coordinates, shift the center so that the tensor sits on top of the correct bond. Finally, I create a .bild file where I write out each triangle using the coordinates and vertices. Tensors have negative and positive values (red and blue in the plot), and so I only take triangles for which r for all vertices is positive, and then in a second step, take only triangles where all vertices are negative. This all gets written to the .bild file, and then opened in chimera on top of the molecule (see picture).
> 
> The question:
> The result is certainly acceptable, however, there is clearly a quality difference in the molecule rendering vs. the tensor rendering (obviously, because the tensors are just made of triangles that I’ve created in python). Is there a better way to bring the surface data describing the tensors into chimera that would then allow usage of some sort of smoothing within chimera for a nicer image? (And possibly with a relatively simple data format….)
> 
> Thanks for your help!
> 
> Cheers,
> Albert Smith
> 
> <example_triangulation.png>