|
Survey of the computer visualization environment at the IMV
Scientific Visualization of Virus Structures
Platforms used at the IMV
Programs used at the IMV
Other powerful packages
- Stanford based Molecular Applications Group MacImdad(R) program.
Info to obtain a demo by ftp is availabe here.
The data
Two types of data are visualized: x-ray crystallographic coordinates and computer-averaged
image reconstructions
of cryo-electron microscopy photographs. The former provides the
highest resolution, usually between 2 to 8 Angstroms, while cryoEM data is usually around
35 Angstroms.
X-ray Crystallographic Coordinates
Atomic coordinates of a virus are collected by
x-ray crystallography
and then visualized in various ways, such as molecular surfaces, ball-and-stick or
space-filling atomic models, and ribbon diagrams. For visualizing a virus, a molecular
surface representation is preferable, as it gives an overall view of the solvent-accessible
surfaces. Unfortunately, calculation of molecular surfaces is memory intensive and requires
sophisticated programs like Grasp for the SGI. Plotting of the atomic coordinates provides
details such as interactions between amino acid residues, but this detail becomes clutter
when visualizing an entire virus, which contains dozens of proteins. One can circumvent
this problem in either of two ways: plot a molecular surface (preferable) or simplify the
atomic plot to include only the alpha-carbon backbone of the molecule (i.e. c-alpha trace),
leaving out the amino acid residues (A c-alpha trace is better when computing power is
limited). Atoms are usually represented as
ball-and-stick
or space-filling
(Van der Waals surfaces). Midas, O or Sybyl on the SGI or MolView,
MacMolecule, or Kinemage on the Mac are excellent programs for this type of visualization.
Although not useful for visualizing the entire virus, ribbon diagrams display nicely a
protein's secondary structure (e.g. alpha-helices and beta-strands) and can be produced
by Ribbons on the SGI or
MolView for the Mac.
An example ribbon
diagram made with MolView shows the VP1 protein of rhinovirus 14 with the
WIN drug bound.
 Some examples can be
found on the Visualizations of Virus Crystal Structures page.
Cryo-electron microscopy and image reconstruction
The data is in the form of three dimensional scalar lattices, meaning that each
node (location in 3-D space) is associated with a scalar density value.
We visualize the datasets using
Explorer 2.2 on the SGI
(Nice description
of Explorer at NCSA Pathfinder Project). We may also start using
Data Explorer,
which would also be run on the SGI (I know of no analogous program for the Mac).
The structure of the virus can be presented as a
3-D isosurface
(a plot of all the nodes with a data variable equal to a chosen threshold value),
a collection of volume
elements or voxels (3-D volumetric rendering), or
2-D slices of a 3-D
scalar field. I'm in the process of implementing
volume rendering
with Explorer of the datasets, as it will allow visualization of gradations
in density in the virus particle.
Some examples
can be found on the 3D CryoEM Reconstructions page.
(Other examples
of 3D Image Reconstructions)
Be sure to check out the Scientific Visualization and Graphics page at the University of Minnesota.
 Go to the
Web Server for Virology home page
Content: ©1994-1997 Stephan Spencer &
Jean-Yves Sgro
Design: ©1997
Internet Concepts, LLC
Web Design by
 |
