- ChimeraX-users - RBVI Mailing Lists

Re: [chimerax-users] Stereo mode error
by Tom Goddard 07 Jan '21

07 Jan '21

Hi J.C., Since Chimera works in stereo and ChimeraX does not I think the problem is almost surely that ChimeraX requires a "core" OpenGL context (ie using modern OpenGL) while Chimera requires a "legacy" OpenGL context. I assume ChimeraX is working for you not in stereo. So it can get a core context, but apparently not a core stereo context. It looks like the Nvidia driver is not providing a core OpenGL stereo context. But Nvidia has clearly stated that it does support this in their 3D Vision FAQ https://nvidia.custhelp.com/app/answers/detail/a_id/4845 "Without 3D Vision, will NVIDIA Quadro GPUs still support OpenGL Quad-buffered Stereo? Absolutely YES. OpenGL Quad-buffered Stereo is not affected by changes to the 3D Vision driver." ChimeraX is asking for quad-buffered stereo core context and "core" contexts are the standard of current OpenGL. So if we believe Nvidia's statement then their driver will give this kind of context. And that suggests that you have a driver configuration setting that prevents it. It is certainly also possible that the Nvidia driver you are using is simply broken in its ability to provide core stereo contexts. But I think it is more likely you have it configured to not give such a context, possibly your X server "Option "Stereo" "10"" setting is the cause. When the Nvidia driver provides a core OpenGL stereo context this does not mean your 3D Vision glasses will work. In fact the above 3D Vision FAQ very clearly says it will not work with a current Nvidia driver unless a separate 3D Vision USB driver is installed. "NVIDIA announced it is no longer supporting 3D Vision. What does this mean? Hardware: NVIDIA is no longer selling the 3D Vision emitters and glasses. Drivers: R418 graphics drivers will be the last driver to contain the 3D Vision driver package. The R418 branch will be supported until January 2020. In addition, NVIDIA will decouple the 3D Vision USB driver from the graphics driver. NVIDIA will provide the 3D Vision USB driver package as a separate installer on NVIDIA.com moving forward - see Quadro Advance Options <https://www.nvidia.com/object/quadro-advance-options.html>. This will allow users to upgrade the graphics driver while continuing to use the standalone 3D Vision USB driver." If this statement from Nvidia is true then the 3D Vision emitter on your computer can only work if you have installed a separate 3D Vision USB driver. If the emitter does not work the screen will show a double image and stereo software like ChimeraX, Chimera and Coot will not complain but you will not see it in stereo through the glasses because the glasses cannot synchronize to the interleaved in time left and right eye images without the infrared emitter. It is possible that that separate 3D Vision USB driver does not work with a core OpenGL context. Kenneth Satyshur has several legacy Nvidia stereo systems and may know how to make ChimeraX work in stereo with 3D Vision. Ultimately, you are trying to use no longer supported hardware. Honestly Nvidia never supported 3D Vision well, and Windows is much more likely to work than Linux. It probably can be made to work with ChimeraX if enough hours are sunk into it, and will certainly be easier if you use old graphics software such as Chimera or Coot which was written based on the legacy OpenGL that 3D Vision originally worked with. Tom > On Jan 7, 2021, at 5:53 AM, Jean-Christophe HAESSIG <haessigj(a)igbmc.fr> wrote: > > Le mardi 05 janvier 2021 à 10:31 -0800, Tom Goddard a écrit : >> Hi Jean-Christophe, > Hi, > > That was a *very* insightful answer, thank you vey much. But at the > same time it's a bit saddening because it sounds like there will be no > way to use ChimeraX fully on this setup. > > My tests indicate that for the time being, 3D Vision still works with > the latest Nvidia driver on Linux. Despite nvidias decision to retire > 3D Vision, it only means they will stop producing shutter glasses. The > following FAQ entry makes that a little clearer : > https://nvidia.custhelp.com/app/answers/detail/a_id/4845 > Still, that doesn't explain how standard OpenGL calls would not be > supported on a recent driver. Maybe using 'Option "Stereo" "10"' in X > selects an alternate code path where only legacy contexts are > available. On the other hand, it seems this is the only way to get the > glasses to work. In fact, I tried every single stereo mode of the > nvidia driver, and all of them except one gave the same results : the > programs using the legacy contexts worked and ChimeraX crashed. > I also tried Chimera and indeed, it works as you suggested. > > In mode 14 (dubbed "Generic Stereo for DP") however, I couldn't get any > of the programs to work. Stereo is reported as disabled in nvidia- > settings but that could simply be beacuse there is no DP screen > attached. Anyway, it sounds more like this mode could support one of > the more recent GL calls. > > I guess I will keep the older 3D vision sets to run coot and chimera, > and seek more recent hardware for chimeraX. > > J.C. Haessig -- systems administrator, IGBMC > >

2 1

Stereo mode error
by Jean-Christophe HAESSIG 05 Jan '21

05 Jan '21

Hi, I'm in charge of software packaging for our users at IGBMC. I'm trying to make stereo mode work since we still have several 3D vision sets. Even if that hardware has been retired, it seems that there isn't a more practical solution to 3D visualization ATM. We are going to test how it goes with a VR headset, but it already seems that not being able to look at anything else while using it will be a hindrance. I found a message on that subject dating back to June 2017, but I get the same errors even with the 0.1alpha2 version which seems to match what would have been current at that time : An error occurred in drawing the scene. Redrawing graphics is now stopped to avoid a continuous stream of error messages. To restart graphics use the command "graphics restart" after changing the settings that caused the error. Traceback (most recent call last): File "/usr/lib/ucsf-chimerax/lib/python3.7/site- packages/chimerax/core/updateloop.py", line 73, in draw_new_frame view.draw(check_for_changes = False) File "/usr/lib/ucsf-chimerax/lib/python3.7/site- packages/chimerax/graphics/view.py", line 165, in draw self._draw_scene(camera, drawings) File "/usr/lib/ucsf-chimerax/lib/python3.7/site- packages/chimerax/graphics/view.py", line 205, in _draw_scene camera.set_render_target(vnum, r) File "/usr/lib/ucsf-chimerax/lib/python3.7/site- packages/chimerax/graphics/camera.py", line 442, in set_render_target render.set_stereo_buffer(view_num) File "/usr/lib/ucsf-chimerax/lib/python3.7/site- packages/chimerax/graphics/opengl.py", line 1206, in set_stereo_buffer GL.glDrawBuffer(b) File "src/errorchecker.pyx", line 58, in OpenGL_accelerate.errorchecker._ErrorChecker.glCheckError OpenGL.error.GLError: GLError( err = 1282, description = b'invalid operation', baseOperation = glDrawBuffer, cArguments = (GL_BACK_RIGHT,) ) The host computer runs ubuntu 20.04 with the latest nvidia drivers (455). ChimeraX is deployed through a singularity container based on ubuntu 18.04. I got glxgears and coot running on the same setup (in 0.8.2 since latest coot also has a known bug in hardware stereo rendering as of now) Can anyone make sense of the error above ? I have a programming background so I can have a look but I have no idea where to start. Thanks, J.C. Haessig -- systems administrator, IGBMC

2 1

VR for DICOM images
by Tom Goddard 04 Jan '21

04 Jan '21

H Gabriel, I am glad you were able to get ChimeraX to work with the Oculus Rift to look at DICOM images. I think this could be a useful feature but it has a long ways to go. ChimeraX is primarily used for molecular visualization, electron microscopy, and 3d light microscopy. The DICOM medical imaging capabilities have had very little development. As you found out the "volume appearance" settings are almost all useless. That was a first try to borrow some image rendering settings from other software that is designed for medical imaging data, but as you saw it does not look good at all. The trouble is that the ChimeraX volumetric rendering is different than other medical imaging software and I have not had enough time to figure out the differences. I have not looked at Invesalius -- thanks for that suggestion. Although some of the volume appearance presets have the brightness curve as a horizontal line, they use different transparencies and colors along that horizontal line. Still I am not surprised that they don't work right. They probably should not have been added to ChimeraX because they are not working correctly. If you can figure out useful presets in Hounsfield units by manually adjusting the curve in ChimeraX Volume Viewer and setting colors for each node, I could add your specific presets to ChimeraX (and probably take out those old broken presets that were translated from other software). You can add vertices to the brightness curve on the histogram in Volume Viewer by right clicking and choosing menu entry "New Threshold". The color button above the histogram changes the color of the last vertex on the curve that you clicked on. This is easiest not using VR. If you get a useful setting you can use the command "volume settings #1" (if your DICOM model is #1) and it will report all the settings including "image levels" and "image colors" which describe the brightness curve. Those are the values I would need to add a volume appearance preset. Did you see this ChimeraX VR DICOM tutorial? Maybe it has some useful info for understanding the current limited capabilities. https://www.cgl.ucsf.edu/chimerax/docs/user/tutorials/dicom-vr.html <https://www.cgl.ucsf.edu/chimerax/docs/user/tutorials/dicom-vr.html> And this documentation page on ChimeraX DICOM page summarizes the features (not specific to VR). https://www.rbvi.ucsf.edu/chimerax/docs/user/dicom-quickref.html I agree with you that there is great potential for VR for medical imaging. It will take years of software improvement to make it work reasonably well, and since it is far from our primary audience for ChimeraX I cannot promise it will ever get those improvements. Tom > On Jan 4, 2021, at 12:31 PM, Biomodelos 3D <unidadbiomodelos3d(a)gmail.com> wrote: > > Tom, thanks for your answer, I finally was able to use ChimeraX with Oculus Rift and it's absolutely amazing!. (I may have issues with some USB contacts in my notebook, don't know really). > My first concern is about DICOM images, as we produce Biomodels (VR, AR and 3D printed). I've been playing with ChimeraX Volume Viewer and Oculus. I found a lot of advantages to accelerate and improve our workflow for VR Biomodels and online demonstrations, I'd like to share: > -We can enter a VR environment immediately, no need of complex segmentation and post processing. This is really fast! and allows us to face emergency cases (usually our job can't face emergencies). > -We can directly measure in 3D and see those measures in real time. > -Doctors can have a much more realistic sense of the problem, in a much faster way, and relate it to different tissue densities. > -Working in VR over original dicom images, many dimensional errors are avoided (any segmentation process implies dimensional errors). > -Also, interpretation errors (biomodelist's most important problems) are avoided, as doctors can see the study in VR by themselves, over an unaltered set of dicom images. > So I think ChimeraX VR capabilities are a complete game changer for doctors and surgeons. > > Beside, I have some questions: > I created a .txt to load a buttonpanel having Volume Appearances Presets (file attached). It works really fine, but I noticed some issues: > Black & White does not work > Glossy II and Glossy have all points over an horizontal line > High Contrast and Low Contrast have no difference (and have all points over an horizontal line) > Pencil does not work > Soft Skin II and Soft Skin III are the same > Soft on Blue and Soft on White have no difference (and all points over an horizontal line) > Appearances Presets look pretty different in ChimeraX compared to other Dicom Viewer software, can it be improved? > We use Invesalius, an open source brazilian software, as our mainstream workflow to produce volumetric renderings and .stl or .obj surfaces from dicom images. Unfortunately, it has no real time VR or AR tools. But is really good in Volumetric Rendering and segmentation of surfaces based in Hounsfield scale thresholds. So I wonder if some kind of knowledge-interchange may be useful for both developments? Please let me know if you think so. > > Thanks again! > > Gabriel Massano > Biomodelos 3D, Hospital de Ninos, Córdoba, Argentina > www.biomodelos3d.com.ar <http://www.biomodelos3d.com.ar/> > <ChimeraX Volume_Presets.txt>

1 0

Deleting lipid molecules
by Moradkhan, Tiglath A 04 Jan '21

04 Jan '21

Hi Elaine Thank you for the help. I figured out what I was doing wrong. I wasn't using the proper syntax to delete the lipid molecules. Thanks Tiglath

1 0

Bottleneck Large Surfaces
by Anderson, Jacob 04 Jan '21

04 Jan '21

With large colored surfaces of maps ~2.4GB, I experience a lag/delay of several seconds during rotation and translation. Is it obvious where the bottleneck might be that causes the lag? I am running ChimeraX on a 3090 gpu w/ 128gb ram, & 12 core CPU. Might there be ways to change other graphics settings to make the manipulations have a less observable delay? ~jacob

3 7

Sequence viewer Background color
by Y. Mutum 02 Jan '21

02 Jan '21

Hi This is a quick question, which I couldn't find online. I'm using ChimeraX-0.93 and was looking at the sequence of a protein chain. The background of the sequence viewer is black (picture attached) - is there any way to change it to some light color, as I can barely see the letters? I tried background color -but that changes for the molecule in display not the sequence. I also checked settings for the sequence viewer but there doesn't seem to be a setting specifically for the background sequence viewer either. Thanks Yaikhomba

2 2

Computing protein cavities
by Y. Mutum 29 Dec '20

29 Dec '20

Hi This question is on calculating empty cavities of large protein complexes (~2 MDa) For example, the pdb-id: 6J5K. I have tried a few webservers and they are ignoring to calculate such large protein-pdb files as a whole. Does ChimeraX have tools that can be installed from the 'More-Tools' or an inbuilt one to calculate protein cavities in ChimeraX? If you could suggest external programs for computing cavities, that are reliable for such large protein-inputs, it would be also of a great help. The cavities need not be sealed by proteins all around - they can still be open to the bulk solvent. Thanks Yaikhomba

2 2

ChimeraX "color by conservation" tutorial
by Elaine Meng 29 Dec '20

29 Dec '20

In case anyone is interested, there's a new tutorial on coloring by sequence conservation using ChimeraX (requires version 1.1 or newer). It has click-to-execute links if you view it in the ChimeraX built-in browser. <https://www.rbvi.ucsf.edu/chimerax/data/conservation-coloring/conservation-…> Happy holidays! Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco

3 3

setting up mousemode - translate z - set pivot
by Helge Paternoga 22 Dec '20

22 Dec '20

Hi Elaine and Tom, Thanks for the explanations, that was very helpful. The build-in shift functionality indeed explains the translate/rotate behaviour! For me the shift button + "reversed" translate/rotate functions work perfectly, so I guess I will keep it that way. Is there anything else that speaks against it? For fitting, I am dealing with ribosomes and I have come to align the map I am fitting into in a way that makes the features clearly distinguishable (i. e. the straight solvent side of the large subunit viewed from the side), then I just move in my new map rotate to align the features and dump it using the z translate until it is in the other map, followed by fit in map. I was contemplating on how much I should stick to my old habits vs. trying new behaviours that might be more effcient in the long run. It is clear that ChimeraX offers many new opportunities for improving the workflow, but for now I like to keep my old logic (which cleary comes from using Chimera for a long time). I like to have three separate keys for rotating - xy translating - z translate that can be modified for the selected model (or dragged, will look into that). Therefore, I came up with the following setup: mousemode shift leftMode 'translate selected models' mousemode shift middleMode 'rotate selected models' mousemode shift wheel contour mousemode setting contour speed 0.1 mousemode control wheel 'translate selected models' mousemode setting zoom speed -1 fun F1 cofr sel camera ortho set bgColor grey lighting soft This actually leaves the right mouse button completely free and I was already wondering what could be a good function to map here! An unrelated observation that I had was that with grey background the selection rectangle while holding control is not visible on the background anymore (It works with white and black). Helge

2 1

setting up mousemode - translate z - set pivot
by Helge Paternoga 21 Dec '20

21 Dec '20

Dear ChimeraX users, I am currently using the homeoffice time to finally get into ChimeraX, coming from Chimera. I just discovered "mousemode" and am trying to set up the controls to my liking. My goal is to make the fitting process of models and volumes into cryo-EM maps as smooth as possible. Currently, I use the following commands which are entered in the "execute these commands at startup" option in the settings => Is this the best way to permanently change the key-mapping? I could not find an .ini file, yet. mousemode shift leftMode 'translate selected models' mousemode shift middleMode 'rotate selected models' mousemode rightMode zoom mousemode shift wheel contour mousemode setting contour speed 0.1 My idea is to have the following set up: left mouse - turn all models (default) middle mouse - translate xy (default) right mouse - zoom shift left mouse - turn selected model shift middle mouse - translate xy selected model shift right mouse - translate z selected model scroll wheel - contour volume control right mouse - set pivot on selected model Turning and translating the models works well, also while holding shift for the selected model. As a side note, 'translate selected models' rotates the selected models in place whereas 'rotate selected models' translates in xy, is this intentional? => I would like to move the selected model in the z direction with the right mouse button while holding shift, is there an option for "translate z" ? => I am often re-setting the pivot point, therefore I would like to bind this function ("Set Pivot") to the mouse as well, does this option exist? I tried "pivot" but it did not work for cryo-EM volumes. Many thanks for reading, Helge

3 2