[Chimera-users] Edge visualisation

[Tom Goddard]

Hi Petr,

  Chimera doesn’t have much to help you measuring nano-crystals, since it is designed for molecular visualization.

  First I think the chance that Laplacian filtering will help you find edges of the crystals is small, because even a small amount of noise will be amplified greatly by the second derivative calculation.  Perhaps by doing Gaussian smoothing first, followed by Laplacian filtering you can get something useful, but it depends on your data.  The Chimera “measure and color blobs” tool only finds a bounding box aligned with the data set axes (density map x, y, and z axes).  It has no capability to rotate to find the tightest box.  There is a “measure intertia” command that computes the principle moment axes of atoms or a surface, which might find the nanocrystal axes if you could produce a surface of just one nano-crystal.  But then there is no tool to use those axes and find the tightest box.  Chimera does not have a specific capability to find a surface of a density map located where second derivative is near zero and first derivative is large.  You could attempt to code something like this as a Python script for Chimera but I think it would be unlikely to work unless you have perfect data.  I think you will get holes in the surface because of places where the second derivative becomes too large or first derivative too small.   Your idea is good in theory, but I just don’t think it is likely to work with real noisy data.

  A few more comments about looking at a surface where the second derivative is near zero.  If you want to show a range of second derivative values near zero, then that defines a volume, maybe a thin shell, not a surface.  You can show that as bounding isosurfaces at the minimum and maximum allowed second derivative values.  The second derivative is a 3x3 matrix (Hessian) so I think you probably are actually interested in the second derivative along the gradient direction being zero.  Chimera doesn’t have any tool that analyzes the second derivative along the gradient direction.

  Your requests are all specialized and outside the realm of what Chimera is designed for.  So Chimera doesn’t do these things, and quite likely no other software does either.  To try your ideas will likely require that you do some programming.

	Tom



On Aug 25, 2014, at 5:33 AM, Petr.Brazda at seznam.cz wrote:

> Hello Tom,
> 
> thank you for the answer. I have now a more complex question. I would like to extract dimensions and volumes of the nanocrystals embedded in a matrix. The problem is that the edges of the crystals are blurred and the density of the whole particle (crystals+matrix) increases slightly towards the centre of the particle. I started to use inflection point as a point where the crystal ends/begins. I found out that there is a Laplacian volume filter in Chimera, which would help me with the analysis. However, there are some problems.
> 
> 1) Theoretically, the second derivative of the inflection point (where the particle ends/begins) is zero, but unfortunately it is also (close to) zero for the matrix and crystals in their "cores". Therefore, is it possible to visualize points, which have nearly zero (including the adjustment of the density (value) limits how far I want to go from zero) second derivative while having the first derivative "far" from zero (with possible adjustment of the density (value) "far")? I think that this filtering would show the real edges of the crystals (maybe there is a better option, as I said I am a complete newbie).
> 
> 2) I would like to use "measure and color blops tool" for measuring the dimensions of the crystals. The crystals have developed crystal faces. Crystal structure is orthorhombic, so using three orthogonal axes for the measurement is optimal for me. However, the setting of the box axes is usually useless for me, because the axes of the box are not parallel with the crystallographic ones. Is it possible to rotate/adjust the box so it would measure the distances I want? Or is there an other option how to measure dimensions of a crystal in three orthogonal directions?
> 
> 3) As I said in the beginning I also need the volume of the particles. Therefore, it would be optimal if Chimera would connect voxels with 2nd derivative about zero and 1st derivative far from zero and define them as the "surface" of the particle. Is such tool available in Chimera? There is a problem when the maximum in the 1st derivative is not so sharp and "noisy" so it is possible that there will be some range where the 2nd derivative will be close to zero/oscillate about zero so some "real minimum on the 2nd derivative" will be maybe necessary.
> 
> Petr
> 
> 
> ---------- Původní zpráva ----------
> Od: Tom Goddard <goddard at sonic.net>
> Komu: Petr.Brazda at seznam.cz
> Datum: 22. 8. 2014 17:58:21
> Předmět: Re: [Chimera-users] Intensity range shift
> 
> 
> Yes, you can shift the density values with volume dialog menu Tools / Volume Filter, filter type Scale. Or you can do it with the volume operation command, e.g. "vop scale #0 shift 100” described in the manual
> 
> http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/vop.html#scale
> 
> Tom
> 
> On Aug 22, 2014, at 5:22 AM, <Petr.Brazda at seznam.cz> <Petr.Brazda at seznam.cz> wrote:
> 
> > Hello,
> > 
> > I would like to ask whether there is a possibility in Chimera to shift the intensity range. I have tomography data, which range from -32000 to 10000 in intensity. A would like to gradually erase parts of the tomogram with volume eraser but it is not possible because the intesity range does not start at zero. It is not possible to do the job with subregion selection. Is there a way to shift the intensity range?
> > 
> > Petr
> > _______________________________________________
> > Chimera-users mailing list
> > Chimera-users at cgl.ucsf.edu
> > http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://plato.cgl.ucsf.edu/pipermail/chimera-users/attachments/20140826/f4302b4f/attachment.html>