[Chimera-users] About electrostatic potentials

[Thomas Goddard]

Hi Xiaobin,

  In theory you could compute the electrostatic potential for the
entire virus capsid and then display it in Chimera.  But that is a big
calculation.  I tried calculating the electrostatic potential for just
the asymmetric unit of 1ny7 using a program called APBS (equivalent to
Delphi).  The 1.5 Gbytes of memory on my machine was not enough (grid
spacing ~0.7A).  For the full capsid with 60 copies the volume I think
it would be about 4 times larger in each dimension requiring 64
(4x4x4) times more memory.  So you would probably need a machine with
about 128 Gbytes of physical memory.  Maybe other software can break
down the calculation and use less memory.

  So the straight-forward approach seems quite difficult.  Here is a
simple approach you could use in Chimera on a normal computer.  You
can simply color the residues by charge (positive blue, negative red)
then map those colors onto the virus surface.  Of course that is not
as accurate as an electrostatics calculation.  Here is how you would
do this simple display in Chimera.  Open 1ny7.  The PDB version of
1ny7 only has the symmetry matrices to produce a pentamer.  So I
suggest using the version from the Virus Particle Explorer (VIPERdb)
web site:

	http://viperdb.scripps.edu/cgi-bin/viper_coord.cgi?VDB=1ny7

Note that the VIPERdb 1ny7 has 3 chains (A,B,C) while the PDB 1ny7 has
only 2 chains (1,2).  I think the VIPERdb version is more useful in
general.  The structure 1ny7 comes from Jack Johnson's lab which
created the VIPERdb web site.

  So open 1ny7.vdb.  With the ".vdb" suffix Chimera will recognize this
as a VIPERdb file and use the Chimera Multiscale tool to show the whole
virus capsid.  Now color the charged residues.  First show the atomic
model by pressing the select "with loaded atoms" button on the Multiscale
dialog, then choose the Style Show... / Wire option in the Multiscale dialog.
To do the coloring use Chimera menu entry

	Select / Residue / amino acid category / positive

then menu entry

	Actions / Color / blue

then

	Select / Residue / amino acid category / negative

and

	Actions / Color / red

  Now to use the atom coloring to color the multiscale surfaces you need
the Multiscale Coloring command tool that is not distributed with Chimera.
Get it from the Chimera experimental features page

	http://www.cgl.ucsf.edu/chimera/experimental/experimental.html

and install it according to the instructions given there.  You have to
restart Chimera after installing it so you should do this before the steps
I described above.

  Now to color the capsid surface using the atom colors use the Chimera
command

	msc #1 #0 5

which says to color Chimera model #1 (the multiscale surfaces) using atoms
from model #0 (1ny7.vdb) within a range of 5 angstroms.  To display the
Chimera command-line where you type this command use Chimera menu entry

	Favorites / Command line

  Now you probably want a higher resolution virus surface.  To get that
press the Select All button on the multiscale dialog.  Then change the
resolution value in the middle of that dialog from 8 angstroms to 3
and press the Enter key or the Resurface button.  You'll then have to
rerun the "msc" command to get the red/blue coloring by charge.

  Here's an image I made for 1ny7.vdb showing what this looks like:

	http://www.cgl.ucsf.edu/home/goddard/temp/1ny7-charges.png

I used some more tricks to make this image.  I decided the red/blue
coloring over the whole capsid was too confusing to look at all at
once, so I selected one copy of chains B and C with the mouse, then
pressed the Select Copies button in the multiscale dialog, then
changed their resolution to 8 angstroms which erased their red/blue
coloring for those chains.  Also I used resolution equal to 0 for
chain A.  That makes a solvent excluded molecular surface which is
higher quality and resolution than the multiscale surfaces made with
resolution > 0.  I did run into a problem that 1ny7 vdb failed to
create solvent excluded surfaces for chain C because two atoms (number
1481 and 1485 in the file) have identical coordinates.  I deleted atom
1481 from the file when I was experimenting with this model to
circumvent that error in the data.

  This is kind of a complex process.  Here's some more info on the
multiscale tool, the msc command, and electrostatic coloring.

	http://www.cgl.ucsf.edu/chimera/tutorials/virus-howto/showvirus.html

	http://www.cgl.ucsf.edu/chimera/experimental/multiscale_color/msc.html

	http://www.cgl.ucsf.edu/Outreach/Workshops/UCSF-Fall-2005/07-VolumeData/tutorial/chaperonin.html

	http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/multiscale/framemulti.html


   Tom