- ChimeraX-users - RBVI Mailing Lists

OpenGL error running ChimeraX under VNC on CentOS 7
by Tom Goddard 14 Mar '19

14 Mar '19

Hi Alexis, ChimeraX is crashing in the OpenGL graphics driver (libGL.so.1) when the Qt window toolkit is trying to use it. This is certainly an OpenGL graphics problem with your VNC setup. Sometimes VNC is run mirroring an actual display connected to the server computer you are connecting to, and sometimes it displays a virtual desktop, for example, when connecting to a server that has no physical display. ChimeraX will very likely work if VNC is mirroring an actual display on the server, but will almost surely fail if it is using a virtual desktop on the server. This is because OpenGL rendering only works to a physical display directly connected to the computer running the app (ChimeraX). I think there have been efforts to send the OpenGL commands to another computer and have that client computer's graphics card render, but that is a very obscure capability, and I have only heard of it failing, never saw it work. Tom > On Mar 14, 2019, at 10:40 AM, Alexis Rohou wrote: > > Hi Tom, > > I am attempting to get ChimeraX running on a CentOS7 workstation with a > nice discrete GPU card for running ISOLDE. Unfortunately, I am having no > luck so far. Tristan Croll, who is helping me get started with ISOLDE, > suggested I send you directly this report and ask for guidance. > > Some details: > - I installed the daily CentOS7 RPM from yesterday (13-March-2019) > - I am connecting to this workstation using a VNC server/client. I do > not have physical access to the workstation, at least not any time soon > - Starting the binary under gdb gives the error & stack trace copied below > - it could be that this is just related to OpenGL over VNC not working > properly. If you suspect this is true, could you suggest some test I > could run to confirm this is the case? > > Any other hints would be most welcome! > > Thanks, > Alexis > > >> WARNING: QOpenGLShader::compile(Vertex): failed >> WARNING: QOpenGLShader::compile(Fragment): failed >> WARNING: Could not link shader program: >> "" >> >> Program received signal SIGSEGV, Segmentation fault. >> 0x00007fffe6e6a013 in ?? () from /usr/lib64/nvidia/libGL.so.1 >> Missing separate debuginfos, use: debuginfo-install >> ucsf-chimerax-daily-2019.03.13-1.el7.x86_64 >> (gdb) bt >> #0 0x00007fffe6e6a013 in () at /usr/lib64/nvidia/libGL.so.1 >> #1 0x00007fffe54e268a in () at >> /usr/lib64/nvidia/libnvidia-glcore.so.340.107 >> #2 0x00007fffe6eac1a0 in () at /usr/lib64/nvidia/libGL.so.1 >> #3 0x00007fffe0e4fc6f in () at >> /opt/UCSF/ChimeraX-daily/lib/python3.7/site-packages/PyQt5/Qt/lib/libQt5Gui.so.5 >> #4 0x00007fffe0df1842 in () at >> /opt/UCSF/ChimeraX-daily/lib/python3.7/site-packages/PyQt5/Qt/lib/libQt5Gui.so.5 >> #5 0x00007fffe0df2143 in >> QPlatformBackingStore::composeAndFlush(QWindow*, QRegion const&, >> QPoint const&, QPlatformTextureList*, QOpenGLContext*, bool) () at >> /opt/UCSF/ChimeraX-daily/lib/python3.7/site-packages/PyQt5/Qt/lib/libQt5Gui.so.5 > > >

1 0

MOL2 import
by Marco J. Müller 14 Mar '19

14 Mar '19

Dear ChimeraX Team, when I try to load structure outputs of docking runs in .mol2 format into ChimeraX, it fails to load them, while the same can be imported by Chimera successfully. After a bit of playing around I found that the following section is relevant and fails in this format: @<TRIPOS>MOLECULE MYCOOLMOLECULE 57 62 1 SMALL USER_CHARGES The problem seems to be that ChimeraX requires explicit definitions of /#features/ and/#sets/ in the third line (apart from /#atoms #bonds #substructures/), while they are typically omitted in most programs. So the following lines could be imported: @<TRIPOS>MOLECULE MYCOOLMOLECULE 57 62 1 0 0 SMALL USER_CHARGES Can this case be addressed in a future version of ChimeraX? Thank you in advance, Marco -- Marco J. Müller PhD student Technische Universität Braunschweig Institute of Medicinal and Pharmaceutical Chemistry https://www.tu-bs.de/pharmchem Beethovenstr. 55 | 38106 Braunschweig | Germany

2 3

How can I use command crosslinks with bonds?
by Jun Ha Song 04 Mar '19

04 Mar '19

Dear RBVI, I want to use few command lines that are exclusively available for pseudobonds. The problem is that my model is consisted of markers and links, and there seems like no possible way for me to simply transfer bonds into pseudobonds. What commands like *crosslinks histogram *provide seems way to good for it to be limited for pseudobonds only. Is there a technical reason why such command is only applicable with pseudobonds? Can commands in Chimera/ChimeraX that work with pseudobonds easily be adapted to work with bonds, or is that too much of technical difficulty? If it is not feasible, is there a way to turn bonds into pseudobonds? Best, Junha Song.

3 2

matchmaker RMSD in ChimeraX
by Tom Goddard 04 Mar '19

04 Mar '19

> From: Martin Moncrieffe <mcm35(a)cam.ac.uk> > Date: March 4, 2019 at 1:15:06 AM PST > To: Tom Goddard > Subject: matchmaker in ChimeraX > > Hi Tom, > > Is it possible to get the rmsd as a function of residue position when using mm in ChimeraX? I’ve turned on the verbose flag and can see the residues used to match but not their respective rmsd’s. >

3 3

Can we draw arrows in ChimeraX?
by Steve Chou 28 Feb '19

28 Feb '19

Dear ChimeraX team, I need to draw straight and curved arrows to show the straight and twisting forces on the molecules. See the attached images (cropped from ChimeraX interface) for details. Does ChimeraX have the functionality to do this? Steve -- Steve Chou

2 3

equivalent to "define"?
by Alexis Rohou 28 Feb '19

28 Feb '19

Hi there, Having looked through the docs a few times, I'm pretty sure I know the answer, but here goes... Is there a ChimeraX equivalent to define <https://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/define.html> in Chimera 1? Or any other facility to draw geometric shapes (I'm particularly interested in calculating and drawing planes)? Cheers, Alexis

3 4

Re: [chimerax-users] Color palettes for planes and solid data?
by Tom Goddard 28 Feb '19

28 Feb '19

Hi Oli, The single plane you are showing is the “solid” style, no different from if multiple planes are shown. You can use brightness value 1 for each color if you want, but it needs to be specified, it does not default to 1 (that usually is not what people want when using solid style rendering of more than 1 plane since it would be totally opaque). Also normally single planes are shown as partially transparent in both Chimera 1 and ChimeraX. Tom > On Feb 28, 2019, at 12:34 PM, Oliver Clarke <olibclarke(a)gmail.com> wrote: > > I’m also unclear why for an opaque plane it is necessary to set a brightness for each map value - surely a color palette with simple color,value pairs would be fine if there is no transparency issue as there is with a 3D solid representation? > > Oli > >> On Feb 28, 2019, at 3:06 PM, Tom Goddard <goddard(a)sonic.net> wrote: >> >> Hi Oli, >> >> To set the colors of different map values use the volume command >> >> volume #1 level 0,0 color white level 10,.9 color blue level 70,1 color green >> >> This says map value 0 has brightness 0 and color white, map value 10 has brightness 0.9 (on a 0 to 1 scale) and color blue, map value 70 has brightness 1 and color green. That is for using the “volume” or “plane” style as shown in the volume viewer panel. >> >> When you instead try to color a clipped map surface, your second screenshot has the surface threshold set to -10 while your map values range -.1 to .15 so you see just a cube surface — not sure what you are doing that for. In your third screen snapshot it looks better, but the colors are streaks on the clipped surface because the cap surface mesh subdivision is not fine. In Chimera you can set the cap subdivision in the Surface Capping dialog, but there currently is no control of the cap mesh subdivision in ChimeraX. Instead you could just crop the map either setting its bounds with the volume command “region” option or easier by using the crop mouse mode. Then the face of the cropped box will use a mesh that matches the map resolution and coloring should be ok. >> >> Tom >> >> >>> On Feb 28, 2019, at 11:35 AM, Oliver Clarke <olibclarke(a)gmail.com> wrote: >>> >>> Hi, >>> >>> Would it be possible to add support for more convenient coloring of solid representation data and planes of density maps? E.g. by specifying a color palette? Apologies if this already exists and I’m just not finding it! >>> >>> I would like to be able to make nice looking slices of a density map, without surface contours, colored by density value (e.g. attached). This can be a good way of showing disordered regions of a EM reconstruction. >>> >>> Currently, when a density map is loaded as a plane or solid represntation in Chimera, there is no way to automatically assign a predefined palette, and the defaults do not convey much information (at least for typical EM maps - example attached). >>> >>> Also, when I attempt another way to do this, coloring the surface by map value and then clipping the surface, the results are odd, unlike in Chimera (see attached at different contours). >>> >>> Cheers >>> Oli >>> >>> <Screen Shot 2019-02-28 at 2.34.47 PM.png><Screen Shot 2019-02-28 at 2.32.03 PM.png><Screen Shot 2019-02-28 at 2.25.00 PM.png><Screen Shot 2019-02-28 at 2.26.03 PM.png> >>> _______________________________________________ >>> ChimeraX-users mailing list >>> ChimeraX-users(a)cgl.ucsf.edu >>> Manage subscription: >>> http://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users >> > >

2 2

Re: [chimerax-users] Color palettes for planes and solid data?
by Tom Goddard 28 Feb '19

28 Feb '19

Hi Oli, There is no capability to automatically use a palette and min/max range of map values that sets the levels and colors. That might be useful although when I have done this by hand I think equally spaced colors rarely produces what I want. The cryosparc image you attached is a good example of the color spacing not producing a very informative view. At any rate, if you want this feature in ChimeraX use the Help / Report a Bug… menu entry which is also for requesting features. I just tried setting the map threshold below the lowest value in the map and clipping in Chimera 1 and it only clips a cube for me. So I’m not sure what you mean when you say that works in Chimera 1. The streaks when you color the clipped and capped surface are because the cap is composed of triangles and only the vertices of the triangles are colored and the triangles are very long and skinny. Read about the Surface Capping mesh subdivision option in Chimera 1 if you want more explanation of it. ChimeraX does not yet support the cap subdivision, but if you follow my previous suggestion it is an alternative method that can work. Tom > On Feb 28, 2019, at 12:31 PM, Oliver Clarke <olibclarke(a)gmail.com> wrote: > > Hi Tom, > > Yes I know it is possible to do manually, thank you, but it requires quite a lot of futzing about with brightness/color pairs to get nice looking results. I was wondering if it would be possible to add support for palettes, with color thresholds automatically distributed between a minimum and a maximum, so that one can for example color a map slice with a rainbow palette between a minimum and a maximum density value (e.g. as per the density slices from cryosparc that I attached to the previous email). > > Re the cube, I set the threshold to -10 so that when I clip the map, it exposes a face that I can color by density value. I do this in Chimera and it works as expected without altering anything, here I just see streaks. I’m still unclear why I just see streaks on the clipped face no matter the threshold - this doesn’t seem to be the intended behavior? > > Oli > >> On Feb 28, 2019, at 3:06 PM, Tom Goddard <goddard(a)sonic.net> wrote: >> >> Hi Oli, >> >> To set the colors of different map values use the volume command >> >> volume #1 level 0,0 color white level 10,.9 color blue level 70,1 color green >> >> This says map value 0 has brightness 0 and color white, map value 10 has brightness 0.9 (on a 0 to 1 scale) and color blue, map value 70 has brightness 1 and color green. That is for using the “volume” or “plane” style as shown in the volume viewer panel. >> >> When you instead try to color a clipped map surface, your second screenshot has the surface threshold set to -10 while your map values range -.1 to .15 so you see just a cube surface — not sure what you are doing that for. In your third screen snapshot it looks better, but the colors are streaks on the clipped surface because the cap surface mesh subdivision is not fine. In Chimera you can set the cap subdivision in the Surface Capping dialog, but there currently is no control of the cap mesh subdivision in ChimeraX. Instead you could just crop the map either setting its bounds with the volume command “region” option or easier by using the crop mouse mode. Then the face of the cropped box will use a mesh that matches the map resolution and coloring should be ok. >> >> Tom >> >> >>> On Feb 28, 2019, at 11:35 AM, Oliver Clarke <olibclarke(a)gmail.com> wrote: >>> >>> Hi, >>> >>> Would it be possible to add support for more convenient coloring of solid representation data and planes of density maps? E.g. by specifying a color palette? Apologies if this already exists and I’m just not finding it! >>> >>> I would like to be able to make nice looking slices of a density map, without surface contours, colored by density value (e.g. attached). This can be a good way of showing disordered regions of a EM reconstruction. >>> >>> Currently, when a density map is loaded as a plane or solid represntation in Chimera, there is no way to automatically assign a predefined palette, and the defaults do not convey much information (at least for typical EM maps - example attached). >>> >>> Also, when I attempt another way to do this, coloring the surface by map value and then clipping the surface, the results are odd, unlike in Chimera (see attached at different contours). >>> >>> Cheers >>> Oli >>> >>> <Screen Shot 2019-02-28 at 2.34.47 PM.png><Screen Shot 2019-02-28 at 2.32.03 PM.png><Screen Shot 2019-02-28 at 2.25.00 PM.png><Screen Shot 2019-02-28 at 2.26.03 PM.png> >>> _______________________________________________ >>> ChimeraX-users mailing list >>> ChimeraX-users(a)cgl.ucsf.edu >>> Manage subscription: >>> http://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users >> > >

1 0

Color palettes for planes and solid data?
by Oliver Clarke 28 Feb '19

28 Feb '19

Hi, Would it be possible to add support for more convenient coloring of solid representation data and planes of density maps? E.g. by specifying a color palette? Apologies if this already exists and I’m just not finding it! I would like to be able to make nice looking slices of a density map, without surface contours, colored by density value (e.g. attached). This can be a good way of showing disordered regions of a EM reconstruction. Currently, when a density map is loaded as a plane or solid represntation in Chimera, there is no way to automatically assign a predefined palette, and the defaults do not convey much information (at least for typical EM maps - example attached). Also, when I attempt another way to do this, coloring the surface by map value and then clipping the surface, the results are odd, unlike in Chimera (see attached at different contours). Cheers Oli

1 0

Missing features in ChimeraX
by Tom Goddard 26 Feb '19

26 Feb '19

Hi Junha, Glad to hear you are trying VR with ChimeraX. Per-model clipping in ChimeraX is done with the clip command, or with the clipping mouse modes, atom radii are set with the size command, visibility is set with the show/hide commands, color is set with the color command, there are volume tracer mouse modes, and saving marker files should work, although you hit some bug and should press the Report a Bug… button when you get that traceback. ChimeraX does not have the selection inspector window yet. ChimeraX uses a different window toolkit Qt while Chimera used Tk so it requires all new code to move any graphical user interface from Chimera to ChimeraX. ChimeraX is a completely new program, modernizing every component of the 20-year-old design of Chimera so even moving commands from Chimera to ChimeraX requires substantial work. If you don’t find ChimeraX ready for your use, of course you should continue to use Chimera. We expect ChimeraX to replace Chimera in a few years. You might want to spend an hour skimming Elaine’s thorough ChimeraX documentation to understand what is in ChimeraX, including many new features like virtual reality that are not in Chimera. Tom > On Feb 26, 2019, at 2:03 AM, Jun Ha Song wrote: > > Tom, > > I recently got HTC Vive, and therefore started using ChimeraX. I have to say; the VR functionality is amazing. The ability to VR alone is sufficient reason for me to move over to ChimeraX. > > However, I noticed that key, crucial, irreplaceable features are (as far as I am aware) non-existent on ChimeraX. Below are function/tools that exist in Chimera that I use (literally) everyday, and I am wondering if these function/tools can also be implemented on ChimeraX or not. > > - per-model clipping (mclip) > - inspect selection (ability to change atom/bond/surface radius, visibility, color) > - volume tracer (ability to save markerset or atom/bond as .cmm) > > If the circumstance is so that implementing these onto ChimeraX is not feasible, then I must remain with Chimera until similar tools become available on ChimeraX. > > Also, is there any particular reason why all the tools of Chimera are not copied onto ChimeraX? Is ChimeraX trying to be another program that is different from Chimera? > > > > As for "ability to save .cmm", I tried saving atom/link using command > > save E:\LBNL\Stereocilia\ChimeraX_Markerset.cmm format markers models #1 > > but I got error saying that > > NotImplementedError: Slices are not implemented. > File "C:\Program Files\ChimeraX\bin\lib\site-packages\chimerax\markers\cmmfiles.py", line 39, in markerset_as_xml > rgb = tuple(r/255 for r in m.color[:3]) > See log for complete Python traceback. > > is this a simple bug during development, or have I typed wrong command? > > Best, > Junha Song.

1 0