- Chimera-users - RBVI Mailing Lists

Re: [Chimera-users] mac pro theoretical question
by Tom Goddard 29 Jan '08

29 Jan '08

Hi Jeff, I don't know much about any of the current Apple graphics cards (Radeon HD 2600 XT 256Mb, Nvidia GeForce 8800 GT 512Mb, Nvidia Quadro FX 5600 1.5Gb). My first concern would be whether bugs in the drivers for these cards are going to cause trouble in Chimera. We have a web page that lists problems with graphics drivers observed using Chimera: http://www.cgl.ucsf.edu/chimera/graphics/graphicsbugs.html#cards There was a recently reported problem (squashed atom spheres) with GeForce 8800 GTX (not GT) on Windows. The Apple driver may be much different. The Quadro FX 4500 which was Apple's high-end offering before the 5600 crashed with solid volume rendering. Don't know if the current 5600 would have the same problem -- but there is a fair chance it does. Chimera uses many different OpenGL calls and I'd guess 1 in 10 drivers exhibit bugs of varying severity when using Chimera. Some of Chimera performance is related to graphics card speed and memory, and some is related to CPU speed and memory. When the displayed models rotate slowly that is the graphics card except if you are using transparent volume surfaces or meshes which are sorted by depth on the CPU for every new view angle. Everything else is the CPU, main memory, memory cache performance, and hard-drive speed. Solid style volume rendering makes a lot of textures which usually will be stored in graphics card memory. More graphics card memory could make larger volume rotate faster because they fit on the card. 256^3 volumes (16 million voxels, 4 bytes per voxel) typically rotate at 10 frames per second with graphics memory sizes 128 Mb and up. Textures have to have power of 2 sizes so just knowing that can get you better performance if you trim your 550x550x100 region to 512x512x100 you'll get a factor of 4 speed-up. I don't know if a single application using one display can use more than one graphics card in Apple's multicard options. Chimera has no special code to do that. For large volume data, hard-drive speed, main memory size, and CPU speed are likely to be noticable factors in Chimera performance, each of about equal importance to graphics card speed. Tom > hi Tom, > > if you were considering a Mac Pro at some point in the future, what > graphics card would you pick when thinking of chimera (if you don't > mind peeking at some point at the list at > http://store.apple.com/1-800-MY-APPLE/WebObjects/AppleStore.woa/9354001/wo/… > > > are rendering bottlenecks in chimera largely a processor-limited thing > (and thus one of the 'lower end' cards listed would be fine), or > something that benefits dramatically from spending $2000 on a graphics > card (I am not in that kind of market, just wondering what kind of > decision you would make if evaluating that platform) > > best, > > -Jeff > > >

1 0

average vs ensemble
by Rachel Scherer 29 Jan '08

29 Jan '08

Given that chimera doesn't calculate average NMR structures, this query may be outside the groups interests, but can anyone suggest what the appropriate protocol is when making aa-substitutions into an NMR structure: should an "average" be calculated (by averaging the atom coordinates across all the models in a PDB file that predates inclusion of the average into the database entry) or, should one use a representative from a cluster generated with "ensemble cluster" be chosen? Or an average of the representatives of the clusters? Or, the cluster representative that best matches the calculated average? -- Rachel Scherer rachel(a)bio.umass.edu Biology Department (413)545-2825 vox UMASS Amherst (413)545-3243 fax

2 1

coordinates
by hsosa＠aecom.yu.edu 25 Jan '08

25 Jan '08

Hi, I am trying to rotate a molecule around a 3D density map. I had already saved the coordinates of the molecule relative to the map. However if I rotate the pdb file (creating a new pdb file with the rotated molecule) by rotating the coordinates around the origin (0,0,0) the new molecule is rotated around a a center that does not coincide with the center of the map. So I am confused about what are the relative positions of the origin for the map and pdb file. If I have an atom with pdb coordinates 0,0,0 should this atom appear at the center of the density map ?. Thanks Hernando -- ----------------------------------- Hernando Sosa Dept. of Physiology and Biophysics Albert Einstein College of Medicine 1300 Morris Park Av. Bronx NY 10461 phone (718) 430-3456 FAX (718) 430-8819 email hsosa(a)aecom.yu.edu -----------------------------------

2 4

Re: [Chimera-users] Save a transformed density map
by Thomas Goddard 24 Jan '08

24 Jan '08

Hi Keren, Chimera has a resampling capability for extracting regions of a map that aren't aligned with the axes. It doesn't work well in the case you want. So I've made a Chimera command that will do what you want. If your the map you wish to save in the new orientation has id number 0 (listed next to the name of the map in volume dialog), and the map you fit into has id 1 you would use the command: vop #0 resample onGrid #1 The vop stands for "volume operation". (Use menu Favorites / Command Line to type commands.) This will create the new map which will appear in the volume dialog and can be saved with the volume dialog File / Save map as... (or using a command like "volume #2 save /tmp/newmap.mrc"). The vop command with the resample operation will be in tonight's builds of Chimera (assuming they succeed). http://www.cgl.ucsf.edu/chimera/alpha-downloads.html Tom Keren Lasker wrote: > hey Tom, > > I need to save the transformed map for other calculations. Is there a > resampling option in chimera? if not I guess that I can use the output > transformation matrix generated by fit map2map - right? > > thanks, > Keren. > On Jan 24, 2008, at 3:27 PM, Thomas Goddard wrote: > >> Hi Keren, >> >> You can save the Chimera session to preserve the orientation of one >> map fit into another. >> >> If you want to write a new map so that when it is loaded in Chimera >> and other software it will be at the correct place in the map it was >> fit into that is harder. The problem is that none of the map file >> formats provide for a positioning transformation in the header. So it >> would be necessary to resample the map that you fit on the grid of the >> map you fit into. Resampling interpolates data values and can degrade >> map quality. Do you want to do that? >> >> Tom >

2 1

volume visualization
by Enrico Malito 24 Jan '08

24 Jan '08

Hello, I am trying to visualize a map as Volume (ccp4 format map calculated with VOIDOO in NEWZD then converted with MAPMAN) and somehow the map in chimera results shifted with respect to the corresponding pdb model (in a completely different place). If I open same pdb and map file in other visualization programs they are actual in the right position (map into the cavity of the enzyme). I'm opening the map from the Volume Viewer-->File-->Open map Am I doing something wrong? Thanks, EM -- Enrico Malito Ben-May Department for Cancer Research Center for Integrative Science, W423J The University of Chicago, 929 E. 57th Street, Chicago, IL, 60637 Tel: 773-702-4608 Fax: 773-702-6260 http://home.uchicago.edu/~emalito

2 2

Save a transformed density map
by Keren Lasker 24 Jan '08

24 Jan '08

hello, How can I save a transformed map after using the fit map2map option? The "save map" option saves it in its original transformation. thank you, Keren.

2 1

Chimera batch surface exporter
by Trent Grover 24 Jan '08

24 Jan '08

Hi Everyone, I'm pretty new to Chimera, but I'm working on an animation project that requires quite a few proteins and molecules. To speed up the process, I'd like to create a batch converter/exporter script to do the following: Open a PDB file Generate a multiscale model at a specified resolution Export the model out as a VRML or OBJ file I think I can do this using a Chimera command file, but I was hoping someone could point me in the right direction for some of the relevant command line syntax. Is it possible to control the multiscale model tool via command line (I know I can launch the dialog, but can I change the parameters and execute)? Is there a reference somewhere for the command line parameters for the multiscale models tool? What is the command line export command? Where can I find an OBJ exporter (there are references to it in the mailing list archive)? Any help would be greatly appreciated. Thanks, Trent BTW - I have a suggestion for improvement of Chimera. Add a command history window (or modify the existing one) that echoes out the command line text that is equivalent to any command executed via the Chimera user interface. Maya has this functionality and it is wonderful for anyone who is trying to figure out how to script a process that they know how to execute via the various menus.

2 1

Re: [Chimera-users] Surface representation
by Elaine Meng 24 Jan '08

24 Jan '08

> Hi, > I would like to know if there is a way to build a surface > representation from a pdb file of a dummy atom saxs model in > chimera. The file consist only of CA atoms. I could simulate this > by increasing the vdw radius but the model becomes way to big. > Thanks > James Hi James, You can use Multiscale Models to make a contour surface that encloses the "atoms." The tricky part may be adjusting the contour surface parameters to get the desired smoothness and enclosed total volume. At least in my tests, Chimera did not like the columns in your file after the B-factor column, so I made a version with those removed and opened it with "File... Open" before proceeding. I will send you that file in a separate message. (1) for purposes of comparison to the surface, I displayed the atoms as spheres with the radius 3.75 (this value was given in comments in the file) using the following commands: setattr a radius 3.75 rep sphere You can show the command line with Favorites... Command Line. (2) started Multiscale Models (under Tools... Higher-Order Structure), clicked the "Make models" button near the bottom. This generates the contour surface but hides the spheres. To also show the spheres, Ctrl-click on the surface to select it, and then in the "Style" line of the dialog, choose "Show also.. Spheres". To see spheres inside the surface, make it transparent; set Transparency to some fraction, say 0.3. (3) Then you can fiddle with the contour surface parameters in the Multiscale Models dialog. Many are hidden but can be shown by checking the box next to "..." I got a fairly decent enclosure of the spheres with all default parameters except setting the "Threshold atom density" for "CA only" to 0.0006 instead of the default 0.002. You could also fiddle with the smoothing parameters and resolution if that is not good enough. If you have some idea of what total volume the surface should enclose, that can be evaluated with Measure Volume and Area (under Tools... Volume Data). I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. meng(a)cgl.ucsf.edu UCSF Computer Graphics Lab and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco http://www.cgl.ucsf.edu/home/meng/index.html

3 3

Re: [Chimera-users] Chimera-users Digest, Vol 57, Issue 24
by Jonathan Hilmer 22 Jan '08

22 Jan '08

POV-Ray also works extremely well as a distributed job. This site has some basic discussion of some of the distributed POV options: http://columbiegg.com/httpov/?p=why Jonathan > Today's Topics: > > 1. (no subject) (jean-francois menetret) > 2. Re: Flip handedness of map (Tom Goddard) > 3. VRML alpha vs 1-alpha (Nils Becker) > 4. Re: VRML alpha vs 1-alpha (Greg Couch) > > > ------------------------------ > > Message: 4 > Date: Tue, 22 Jan 2008 10:37:54 -0800 (PST) > From: Greg Couch <gregc(a)cgl.ucsf.edu> > Subject: Re: [Chimera-users] VRML alpha vs 1-alpha > To: Nils Becker <nils.becker(a)ens-lyon.fr> > Cc: chimera-users(a)cgl.ucsf.edu > Message-ID: <Pine.OSF.4.63.0801220934340.1761700(a)guanine.cgl.ucsf.edu> > Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed > > On Tue, 22 Jan 2008, Nils Becker wrote: > > > Hi, > > > > I found information on the web site that VRML models use > > 1-alpha blending for transparency, and that that looks ugly. > > I agree. > > > > Now I have externally generated VRML files including transparency. > > > > Is there a way to get nice (normal) transparency rendering of these > > models in chimera? > > > > e.g. editing the vrml files, converting them to another format or some > > option inside chimera? > > > > Thanks for any help, cheers, Nils > > The VRML transparency is 1-alpha, because alpha is an opacity value. So > that is just an alternative way of saying that same thing, and is not the > reason it is worse. > > The reason VRML transparency is worse is that there is no sorting within > the VRML shapes, so the blending of transparent objects has to be > drawing-order independent, so VRML uses additive blending. > > The simpliest solution would be to live with chimera's current depiction > and use the Save Image dialog to generate a ray-traced version. The > ray-traced version will be much nicer (and take a lot longer to compute, > try using the POV-Ray 3.7 beta on a multi-core computer to speed things > up). > > The other solution, is to write some Python code and use the _surface > module's Surface_Model and Surface_Group classes. That is, convert each > VRML object into a triangulated surface, a Surface_Group, and put all of > the Surface_Groups into one Surface_Model. All surfaces within one > Surface_Model are sorted so that the transparency is correct for more than > one overlapping surface (unlike the transpareny in chimera between > models that only works for one level of transparency). You can see > examples that use the _surface module in the GraspSurface, IMOD, > Icosahedron, and Intersurf modules (and others) in the CHIMERA/share > directory of your installation. > > If you have any questions about the _surface module, please ask them on > the chimera developers mailing list, chimera-dev(a)cgl.ucsf.edu. > > Greg Couch > UCSF Computer Graphics Lab > > > ------------------------------ > > _______________________________________________ > Chimera-users mailing list > Chimera-users(a)cgl.ucsf.edu > http://www.cgl.ucsf.edu/mailman/listinfo/chimera-users > > > End of Chimera-users Digest, Vol 57, Issue 24 > ********************************************* >

1 0

VRML alpha vs 1-alpha
by Nils Becker 22 Jan '08

22 Jan '08

Hi, I found information on the web site that VRML models use 1-alpha blending for transparency, and that that looks ugly. I agree. Now I have externally generated VRML files including transparency. Is there a way to get nice (normal) transparency rendering of these models in chimera? e.g. editing the vrml files, converting them to another format or some option inside chimera? Thanks for any help, cheers, Nils

2 1