[Chimera-users] Calculating distances in and between sheets

[Elaine Meng]

Hi Einav,
It all depends on what atoms are used to define the planes, of course.  Those two numbers really aren’t that far apart, unless you and the collaborator used exactly the same atoms to define the planes.  I guess the Chimera measurement is also the mininum distance, not the average distance, between the plane objects.

I opened 5i55 and expanded to unit cells 1 4 1.  Then for the plane definitions, I tried 4 variations, Chimera commands:

define plane #0,2,4,6:2-20 at n,ca,c,o
define plane #1,3,5,7:2-20 at n,ca,c,o

define plane #0,2,4,6:2-20 at ca
define plane #1,3,5,7:2-20 at ca

define plane #0,2,4,6:3-19 at n,ca,c,o
define plane #1,3,5,7:3-19 at n,ca,c,o

define plane #0,2,4,6:3-19 at ca
define plane #1,3,5,7:3-19 at ca

Visually, the planes for each helix “sheet" are rather close together and appear to well represent the locations of the helices.  I measured the minimum distance for each pair, giving (in the same order as the above):

minimum distance from #0:? to #1:? = 13.192
minimum distance from #0:? to #1:? = 13.221
minimum distance from #0:? to #1:? = 12.632
minimum distance from #0:? to #1:? = 12.624

Personally I would be comfortable saying that the layers are spaced ~13 angstroms apart.  I changed to orthographic view to avoid distortion from perspective, and saved a session.  Image and session attached.  I happened to be using the daily build, but the session restores OK in Chimera 1.12 when in response to the warning I say to open it anyway.  However, the named selections aren’t saved in the session file (I guess because they aren’t made of atoms), so you have to again select each plane (easy way is to show Axes/Planes/Centroids GUI and click it in the list of objects) and and use “namesel” to give it a name that can be used in “measure distance”.

After using namesel to name the first two planes p1 and p2 (coincidentally the same as their IDs in the Axes/Planes/Centroids GUI), the blue line in the image shows the result of command:

measure distance p1 p2 show true

You could do it similarly for the other 3 pairs of planes.  (The lines showing the distance measurements are listed separately in the Model Panel and could be hidden or closed using that dialog.)  The visual feedback may help you decide how much you trust the number.
I hope this helps,
Elaine
-----
Elaine C. Meng, Ph.D.
UCSF Chimera(X) team
Department of Pharmaceutical Chemistry
University of California, San Francisco

-------------- next part --------------
A non-text attachment was scrubbed...
Name: planes.png
Type: image/png
Size: 139645 bytes
Desc: not available
URL: <http://plato.cgl.ucsf.edu/pipermail/chimera-users/attachments/20180205/49a4b207/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: planes.py
Type: text/x-python-script
Size: 550246 bytes
Desc: not available
URL: <http://plato.cgl.ucsf.edu/pipermail/chimera-users/attachments/20180205/49a4b207/attachment.bin>
-------------- next part --------------

> On Feb 5, 2018, at 12:36 PM, Einav Tayeb-Fligelman <einavt at gmail.com> wrote:
> 
> Hi Elaine,
> 
> Thank you very much for your elaborated answer!
> The thing is that I tried, as a reference, your method on a structure that a collaborator of us already calculated everything using a few PyMol scripts, and I get a totally different number (12.6 vs. 11.3 A). I am just not sure what it more accurate.
> If you don't mind to try your method on my structure maybe you could offer me some specific tips for these kind of calculations (since I have to do it on many similar structures now).
> The PDBID is 5i55. If you go to the "unit cell" window and set the copy number on the Y axis to an higher value (I used at least 4) then you should get two "sheets" of alpha-helices. I am trying to calculate the distance between those sheets and the problem is that the planes are really not parallel.
> I used for the calculation only the backbone atoms and omitted the flanking residues, but I am not sure that this is the best way to go.
> 
> I would really appreciate your help on that.
> 
> Thank you,
> Einav  
> 
> On Mon, Feb 5, 2018 at 9:40 PM, Elaine Meng <meng at cgl.ucsf.edu> wrote:
> A minor limitation of my “measure distance” method of getting the closest distance between two plane objects is that it only uses the disc surface vertices.  (Each plane display is actually a triangulated surface.)
> 
> It is very minor in my opinion, really only affecting planes that actually intersect or are very close together.  For example, the closest distance between two intersecting planes might be reported as a small nonzero number because the surface vertices don’t exactly coincide.  I couldn’t figure out a way to show the discs as mesh to reveal the locations of the vertices, but my test of intersecting discs gave just 0.22 angstroms.
> 
> Elaine
> 
> > On Jan 27, 2018, at 9:27 AM, Elaine Meng <meng at cgl.ucsf.edu> wrote:
> >
> > Dear Einav,
> > There may not be a way to do exactly what you want, since for distance measurements the planes are treated as infinite.  However, axes are treated as finite line segments for distance measurements (although infinite for angle measurements), so I would have guessed it would work, but maybe not with the specific geometry of your system.
> >
> > One idea is that you can use command “measure distance” to report the minimum distance between one set of atoms and another, and in that case I'd use only CA atoms or N,CA,C,O to avoid the variability from sidechains.  I would do it several times (say for different regions of the sheets) to see the spread of the resulting values.
> > <http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/measure.html#distance>
> >
> > Oooh, I just thought of a trick: use this command to measure the minimum distance between the discs of defined planes (but treating them as finite) and to my surprise, it actually worked!!
> >
> > (1) define two planes
> > (2) Ctrl-click one to select it, command: namesel sel1
> > (3) Ctrl-click the other plane to select it, command: namesel sel2
> > (4) command: measure distance sel1 sel2
> >
> > (I thought it wouldn’t work because these plane surfaces are hidden from the model panel, but as far as I can tell, the distance reported in the status line and Reply Log is correct for my two planes)