VMD-L Mailing List

From: Leonardo Trabuco (ltrabuco_at_ks.uiuc.edu)
Date: Tue Jun 15 2010 - 02:54:08 CDT

Dear Esmael,

In the future, please send MDFF questions to the vmd-l mailing list instead.

On Tue, Jun 15, 2010 at 1:50 AM, esmael <haddadian_at_uchicago.edu> wrote:
> Hi Leonardo,
>
> Is there a way to change the grid spacing when converting the electron density map to an MDFF potential?

Currently, there is no direct option to do this. It's now in my TODO
list. But here are a couple of workarounds that would accomplish this:

1) If you only want to downsample (divide total number of voxels by 8)
or supersample (multiply total number of voxels by 8), you can use the
volutil command.

2) For an arbitrary grid spacing, you can create a dummy map with the
desired grid spacing and filled with 1s, and then use volutil to
multiply this dummy map together with your original EM map. It should
work at least for orthogonal cells aligned to x,y,z. Here's an
example:

# Create a simulated map from a pdb structure with grid spacing of 4A
mol new 1ubq waitfor all
set sel [atomselect top {protein}]
package require mdff
mdff sim $sel -spacing 4 -o foo.dx

# Create a dummy map filled with 1s with grid spacing of 1A
# Note that the volmap command uses the term "resolution" for what we
call "grid spacing" in MDFF
# You can optionally specify the target dimensions of the final map by
using the -minmax option of volmap
volmap mask $sel -res 1.0 -o bar.dx
package require volutil
volutil -range 1:1 bar.dx -o bar1.dx
volutil -mult bar1.dx foo.dx -o foo1.dx

# The file foo1.dx should have the simulated density map with grid
spacing of 1A.

Please let me know if any of this is not clear.

Best,
Leo

> Thanks a lot,
>
> Esmael Haddadian
> Postdoctoral Scholar
> Department of Biochemistry & Molecular Biology
> University of chicago 

-- 
Leonardo Trabuco, Ph.D.
CellNetworks, BioQuant
University of Heidelberg, Germany