[Chimera-users] matrices

[Tom Goddard]

Hi Matt,

   My mistake, matrixset with a file containing an inversion

Model 0.0
       -1  0  0  0
        0 -1  0  0
        0  0 -1  0

does not give an error, but also changes the matrix so it is the proper 
rotation

       -1  0  0  0
        0 -1  0  0
        0  0   1  0

   I tried a few other tests with Python code and all gave an error when 
I tried to even create an inversion Xform.  Also I am quite sure an 
interpolation with "reset" would not work going from an inversion matrix 
to an identity matrix because I know that code gets the rotation axis 
and angle relating the initial and final orientation and simply takes 
small rotation steps.  That clearly won't work if the two states are not 
related by a rotation.

   If you still are interested in doing this in Chimera it could be done 
with some Python code by transforming the atom coordinates.  Instead of 
changing the model Xform you change the Atom (x,y,z).  You could make a 
trajectory that gradually flips the z-coordinate over 100 frames.  Could 
be done in Chimera Python or an external program could generate 100 PDB 
files with the z-flip and those could be loaded as a trajectory using 
the MD Movie dialog (Tools / MD).

     Tom



Dougherty, Matthew T wrote:
> Hi Tom,
>
> can you send me the session?  Having difficulty understanding what you did.  I thought I did the same.
> When I create a inverted matrix -1[000] and read it in using a matrixset command, it immediately transforms it to a 2[001] matrix, that is it it creates a positive one for z.
>
> What I would expect for a m[001] matrix interpolating from an identity matrix using a reset command of 100 frames is the object would flatten out or scale from 1 to 0 on the z axis from frames 1-50, and invert 50-100, returning to full scale, but inverted at frame 100.
>
> I should be able to do this in Blender, but would rather do it in Chimera.
>
> thanks,
>
> Matthew Dougherty
> National Center for Macromolecular Imaging
> Baylor College of Medicine
> ________________________________________
> From: Tom Goddard
> Sent: Monday, September 26, 2011 12:36 PM
> To: Dougherty, Matthew T
> Cc: chimera-users at cgl.ucsf.edu
> Subject: Re: [Chimera-users] matrices
>
> Hi Matt,
>
>     You can use an inversion or inversion combined with a rotation in
> Chimera and at least some things work correctly.  I tried the matrixset
> command to apply an inversion and the inverted molecule displayed and
> rotated correctly, and then I created a surface for the inverted
> molecule which worked.  When I then tried to align the inverted molecule
> to a copy of the original uninverted molecule, it inverted it back to
> give perfect alignment (didn't just do the best pure rotation).
>
>     As far as I know, no thought has gone into making sure Chimera works
> with inverted spatial transformations, so some operations with inverted
> models may misbehave and you'll have to test and see.
>
>       Tom
>
>
> Dougherty, Matthew T wrote:
>> I was doing something with crystallographic point group matrices.
>>
>> Does Chimera allow for improper rotations and inversions?
>> Or does it rectify them as proper rotations?
>>
>>
>> I would like to go from
>>
>> Model 0.0
>>        1 0 0 0
>>        0 1 0 0
>>        0 0 1 0
>>
>> TO
>>
>> Model 0.0
>>        -1  0  0  0
>>         0 -1  0  0
>>         0  0 -1  0
>>
>>
>>
>> Matthew Dougherty
>> National Center for Macromolecular Imaging
>> Baylor College of Medicine
>>
>
> _______________________________________________
> Chimera-users mailing list
> Chimera-users at cgl.ucsf.edu
> http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users
>