March 2007 - Chimera-users - RBVI Mailing Lists

sequence & structure in Chimera
by Elaine Meng 20 Mar '07

20 Mar '07

Hi Abhijit, If I understand correctly, there were two questions here: > One more question is it possible to take a protein sequence of a gene, > part of which is cloned, crystallized and 3d structure is available, > and superimposes the protein sequence with the reported crystal > structure? Or how can you model a protein sequence to generate a > putative structure, for the part that hasn’t been crystallized so far, In Chimera, (1) can you associate a structure with a sequence that is read in from a separate file? short answer: yes, if that sequence is in an alignment (2) can you do comparative modeling (create a 3D model of an unknown structure starting from a known structure that is thought to be similar)? short answer: no, except simplistic replacement of a few side chains with the command "swapaa" I'll start with #2: Chimera does not do comparative modeling or threading different sequences on a known structure. Besides the approaches mentioned by Eric (getting a model from MODBASE or using the MODELLER program to make the model yourself), listed below are several web servers that might be useful. I believe the first two replace side chains while the others perform more complete modeling (can handle insertions/deletions and adjust the backbone). It's been a while since I used any of these, however, so you should read any author-provided information. - SCWRL server http://www1.jcsg.org/scripts/prod/scwrl/serve.cgi - Maxsprout server http://www.ebi.ac.uk/maxsprout/ - ModWeb server (requires registration) http://alto.compbio.ucsf.edu/modweb-cgi/main.cgi - RAPPER server (registration encouraged) http://mordred.bioc.cam.ac.uk/~rapper/ - PSIPRED server http://bioinf.cs.ucl.ac.uk/psipred/ - Swiss-Model server http://swissmodel.expasy.org//SWISS-MODEL.html Now question #1. You already saw the Sequence tool, which shows the exact sequence of any peptide or nucleotide chains in structures open in Chimera. However, you might want to read in a sequence from a separate file and associate it with the structure to allow the same kinds of crosstalk (a selection on the sequence selects the structure and vice versa). You can do that, if you have an alignment of that sequence with one or more additional sequences. These other sequences don't need to have any structures, and your sequence of interest does not have to match the structure exactly to associate with it (often there are point mutations, or the sequence is for a close homolog of the structure, or the structure is missing a few residues here and there, or is only one domain from the whole sequence). Here's a list of the sequence alignment formats that can be read (you can just use File... Open in the Chimera menu): http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/filetypes.html#alignment and the "Sequence-Structure Association" section in this page may be useful: http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/ multalignviewer/framemav.html 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

1 0

Motion callback functions
by Tom Goddard 20 Mar '07

20 Mar '07

Hi Karin, I've attached some example Python code that monitors the motion of models in Chimera. It registers an OpenState trigger. For your case you might update the color on your plane when the plane moves but the map does not move. Let me know if more details would be helpful. Tom Karin Gross wrote: > Hi Tom, > > i have a question concerning rotation and translation of surface models in > chimera. > I want the user to be able to rotate/translate my plane via mouse, but run the > computation of coloring and do the coloring during the movement. > Is it possible to register a python callback for this, and get the > rotation/translation information? > > Thanks > -Karin > > # ------------------------------------------------------------------------------ # Example code that monitors the motion of models in Chimera. # # Each Chimera Model object has an openState attribute whose value is an # OpenState object. The OpenState object has attributes xform, active, # cofr (center of rotation), bbox (bounding box). These are defined in the # Chimera C++ header file OpenModels.h. (Chimera C++ headers can be downloaded # from http://www.cgl.ucsf.edu/chimera/sourcecode.html). # # This example code monitors changes in the OpenState objects for all models. # class Monitor_Motion: def __init__(self): from chimera import triggers h = triggers.addHandler('OpenState', self.xform_changed_cb, None) self.handler = h def xform_changed_cb(self, trigger_name, call_data, trigger_data): if 'transformation change' in trigger_data.reasons: # The transformation matrix of some model has changed. from chimera import openModels mlist = openModels.list() # All open models. for open_state in trigger_data.modified: models = filter(lambda m: m.openState == open_state, mlist) if models: names = ', '.join(map(lambda m: m.name, models)) print names, 'moved' print open_state.xform def stop_monitoring(self): from chimera import triggers triggers.deleteHandler('OpenState', self.handler) self.handler = None # ------------------------------------------------------------------------------ # mm = Monitor_Motion()

1 0

Electrostatic coloring on slice surface
by Tom Goddard 20 Mar '07

20 Mar '07

Hi Boaz, When you color a surface (e.g. GRASP surface) using electrostatic potential with the Surface Color tool it does show the correct colors on the slice if you are clipping the surface. This is described in the guide to volume data display http://www.cgl.ucsf.edu/chimera/tutorials/volumetour/volumetour.html#colorc… The colors are only assigned at the triangle vertices of the slice surface which can be seen by switching to "mesh" display in the Surface Capping dialog. To get finer resolution coloring increase the "Mesh subdivision factor" in the Surface Capping dialog from 1.0 to say 3.0 (giving ~9x more vertices). Press the Enter key after changing the number to see the effect. Tom Boaz Shaanan wrote: > Hi Tom, > > Thanks a lot. While we're at it I have another question related to > surface rendering. In the 'colour surface' window there is button that > says 'colour sliced ...'. Does that mean that if I covered a molecule > with surface and mapped a Delphi calculated potential map onto it, if > I then slice into the molecule, the surface covering this slice will > show the potential correctly ? That would be very nice but I'm just > wondering whether I can believe what I see. Grasp used to do this > nicely but I'd like to use chimera (and grasp is dead anyway as far as > I'm concerned - I don't have SGI's). I'd appreciate your input. > > Cheers, > > Boaz

1 0

Re: [Chimera-users] pdb-l: structure based sequence alignment
by Eric Pettersen 16 Mar '07

16 Mar '07

Hi Sona, Before getting to your question, just wanted to mention that the post you quoted isn't accurate anymore: Chimera can incorporate structureless sequences into alignments as of the last production release (October). Once you have a structure-based alignment (or any alignment really) showing in Multalign Viewer you can add sequences to it using MAV's Edit menu. I don't know of any non-madness-inducing way in Chimera of superimposing 50+ ligands without writing a script. There are two kinds of scripts you could write. One is a text file containing the exact same commands you would type at Chimera's command line. You would give the file a suffix of .com or .cmd and open it using File...Open or the "open" command. The other type is a Python script, also a text file, which requires a little bit of learning but offers the ability to use loops -- quite helpful in this scenario. A Python script file would have a .py suffix and can be opened the same way as a Chimera-command script. For purposes of the example scripts below, I'm going to assume the ligands have the same residue name (FMN in the example), that there is only one occurrence of the residue per file, and the ligand atoms have the same name. PDBs you want to use that don't match these criteria you could do by hand -- hopefully there would just be a few. On the assumption that you have a mix of NMR and crystallographic structures, the script will delete hydrogens, yielding a heavy-atom only RMSD (probably a good idea anyway if there are any methyl groups). Lastly, I'll assume you have a reference structure to match onto already open in model #0 before you run the script. The Chimera-command script would be something like this: open 1jhf; del H; match #1:FMN #0:FMN; close 1 open 1kef; del H; match #1:FMN #0:FMN; close 1 open 1rd9; del H; match #1:FMN #0:FMN; close 1 etc. For the Python version I'll make one further assumption: that you have the PDB IDs listed in a file named "pdbs". The Python version would be: from chimera import runCommand # use runCommand to execute Chimera commands f = open("pdbs", "r") # open file containing PDB IDs for line in f: # read through file pdbID = line.strip() # remove trailing newline from text runCommand("open " + pdbID + "; del H; match #1:FMN #0:FMN; close 1") f.close() # close file The indentation for the lines that follow the "for" statement indicate that they are to be executed as the loop. The lack of indentation for the f.close() indicates that it is not part of the loop -- it is executed once the loop finishes. --Eric On Mar 15, 2007, at 6:54 PM, Sona Vasudevan wrote: > Hello Eric, > > Have a few questions about Chimera. I would like to superimpose > just the ligands bound to complexes. In this case I would like to > basically get a rms deviation of 50+ ligand co-ordinates (same > ligand from different PDBs). I now Chimera's match could do it but > is there a way a easier way to superimpose ligands? > > Thanks! > > Sona > > > On 2/6/06, Eric Pettersen <pett(a)cgl.ucsf.edu> wrote: > Hi Sona, > UCSF Chimera (www.cgl.ucsf.edu/chimera) can certainly > construct a > structure-based sequence alignment, but can't incorporate > structureless > sequences into the alignment so maybe it's not what you're looking > for. > But just in case, here are the URLs for Chimera's Match->Align tool > (which creates the alignment) and MatchMaker tool (which superimposes > structures): > > http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/matchalign/ > matchalign.html > http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/matchmaker/ > matchmaker.html > > --Eric > > Eric Pettersen > UCSF Computer Graphics Lab >

1 0

Problems using Collaboratory
by slaton 14 Mar '07

14 Mar '07

Hi, Having some problems using Collaboratory mode. Some scenarios: 1. When attempting to host a collaboratory session between linux hosts on the same subnet, with another linux box running the hub, no one can join the session. I get the error connecting that suggests checking your firewall settings. None of the machines have iptables running, and since they are on the same subnet that should rule out border router/firewall issues. 2. If using a Mac OS X machine (PPC) to run the hub -- on a different subnet than the linux boxes, but within the same building -- the first linux box can join the hub. A second linux box on the same subnet fails with the same error as in 1) above. This can be replicated every time. Neither linux boxes are running iptables or any other firewall, they have nearly identical hardware and are running the same distro/version/patchlevel. 3. If the MOSX box (where the hub is running) now joins the session with the 1st linux box, they can chat with each other and see either others' pointers. However if the connected user on the linux box opens a model (for my tests, i used a SPIDER volume), the connected user on the MOSX box cannot see it. An error pops up on the linux box reading: "Error processing trigger: "Camera": AttributeError: '_chimera.Camera' object has no attribute 'viewDistance' Rotating/moving the model on the linux box triggers this error on the MOSX box: "Error processing trigger: "collaboratory": ValueError: unopen model Additionally, while sending chat messages from the MOSX box works, it also throws errors like this: "Error processing trigger; "Chat command": TclError: invalid command name ".38462968.47381600.47395144.47395584....[etc] 4. If i reverse scenarios here -- hosting on Linux and joining from MOSX (and the host machine) -- the results are nearly the same. Each user can chat and see the other's cursor, but as soon as one user loads a model it starts throwing errors like the ones mentioned above. After which, chat mode no longer works. So.....any known problems with collaboratory in this release, or any tips on troubleshooting these problems? All tests were performed using Chimera snapshot release 1.2352, i386 version (linux) and X11 version (MOSX). thanks slaton Slaton Lipscomb Nogales Lab, Howard Hughes Medical Institute http://cryoem.berkeley.edu

1 0

Re: [Chimera-users] Chimera - Question
by Elaine Meng 13 Mar '07

13 Mar '07

Hi Peter, No problem - the features are documented, but there is no tutorial that includes adding a sequence. I tested the following with the most recent snapshot (1.2352, March 2007). (1) Open your sequence alignment in Chimera. I just found the following problem when testing the procedure: *if you created the sequence alignment with Match->Align, it is necessary to save it to a file (File... Save As from the Multalign Viewer menu, pick any format), quit from the alignment, and reopen it in Chimera (File... Open in the main menu) before you can add a sequence to it* ...this will be fixed in the next release, however. If the sequence alignment was already read from a file, no extra step is needed. (2) Choose "Edit... Add Sequence" from the Multalign Viewer menu and paste the new sequence in the dialog. Enter a name for the sequence, adjust the alignment parameters as desired, and click OK. Look at the alignment and decide whether it seems reasonable. If not, you can use "Edit... Delete Sequences/Gaps" to remove the sequence and try realigning it with "Edit... Add Sequence" and different alignment parameters. For example, if the sequence is fairly distantly related to the others, you may have better luck with blosum-30 than blosum-62. Editing is described here http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/ multalignviewer/multalignviewer.html#editing (3) If they are not already open, open one or more structures that associate with the sequence alignment. (4) Once your new sequence is aligned with the others to your satisfaction, you can draw out boxes on the alignment using the mouse. If the boxes enclose parts of the sequence(s) that are associated with the structure(s), the corresponding parts of the structures will be highlighted (selected). Or, you can go the other direction by selecting parts of the structures and seeing where they are in the alignment. These forms of crosstalk are described here http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/ multalignviewer/multalignviewer.html#crosstalk and in the "Sequences and Structures" tutorial. You can also show the secondary structure assignments on the alignment with "Structure... Secondary Structure... show actual" (for those sequences associated with structures) or "show predicted" (for sequences without structures) Alternatively, if you already have a sequence alignment that includes your sequence of interest plus others that will associate with structures, you can just open them all in Chimera and skip straight to #4. Sometimes I get alignments from databases such as Pfam. 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 On Mar 13, 2007, at 2:27 PM, Peter Murphy wrote: > Thanks for the quick response Dr. Meng, > > Thanks to your helpful tutorial, I now know how to align two > structures and get an alignment. But I am interested in the ability > to add a sequence (that > doesn't have a structure) to the alignment and see if/where parts > of it map onto the structures. Is there such a tutorial on your > website, or might you be able to give me a quick rundown :P > > Thanks again > > -- > Peter Murphy, B.Sc., M.Sc Candidate > Atlantic Research Center > Dalhousie University > Halifax, Nova Scotia, > Canada

1 0

Chimera - Question
by Peter Murphy 13 Mar '07

13 Mar '07

Hi, I have a quick question about your program. Normally (on a day to day basis) I use fairly basic features of chimera (coloring, etc) to produce pictures of crystal structure images (for presentations, papers, etc). Rarely do I have to use the more advanced features of chimera (Solvent accessible surface area, distance calculations, etc) so I don`t have a good idea of the limits of your program. Thus, I figured I would check to see if what I wanted to do is out of the scope of your program. I have a few crystal structures which I want to align (I now understand how to do this based on your online tutorial (thanks)), but what I really want to do is feed a third structure into the consensus structure (or if there is no consensus structure, just feed a known sequence into a known crystal structure to see if the sequence could form that structure. Is there anyway to do this in your program.. Thanks for any help you may be able to provide, -- Peter Murphy, B.Sc., M.Sc Candidate Atlantic Research Center Dalhousie University Halifax, Nova Scotia, Canada

2 1

Rotated density maps
by Thomas Goddard 13 Mar '07

13 Mar '07

I added two Chimera commands "newmap" and "addmaps" that allow extracting a rectangular piece of a density map with axes that are rotated relative to the original map. http://www.cgl.ucsf.edu/chimera/experimental/resample_map/resamplemap.html This is useful for extracting substructures from electron microscope tomograms. It is not part of current Chimera distributions but can be obtained as a separate plug-in from the experimental features web page: http://www.cgl.ucsf.edu/chimera/experimental/experimental.html It is called "Resample Map Commands" on that web page because data values from the original map are interpolated (i.e. resampled) on the new rotated grid. Tom

1 0

~longbond
by Boaz Shaanan 12 Mar '07

12 Mar '07

Hi, I'm trying to use ~longbond to hide a long bond in the Calpha trace of a model (270-280 in ijsc_a) but the bond refuses to disappear. The bond is not there when atoms are shown in full and only shows up in the Calpha trace (atoms->show chain trace). I type ~longbond in the command line after the trace is on and nothing happens. What am I missing ? I'd appreciate your advice. Thanks, Boaz Boaz Shaanan, Ph.D. Associate Professor Dept. of Life Sciences Ben-Gurion University of the Negev Beer-Sheva 84105 Israel Phone: 972-8-647-2220 ; Fax: 646-1710 Skype: boaz.shaanan‎

2 1

Re: [Chimera-users] independent movement
by Elaine Meng 12 Mar '07

12 Mar '07

Hi Michel, namesel is a newer command added after version 1.2304 - get the recent snapshot to use this command: http://www.cgl.ucsf.edu/chimera/download.html#snapshots The release notes for this snapshot list what has changed since 1.2304: http://www.cgl.ucsf.edu/chimera/1.2349/docs/relnotes/snapshot.html The links I included in previous messages were to "development" documentation, which will always be newer than what you have (although mostly the same if you have a recent version). The documentation included with a download is more synchronized - you can access and search your local copy of the documentation using the Chimera Help menu. ----- 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 On Mar 11, 2007, at 3:32 PM, michel dedeo wrote: > Elaine and Tom > > Thanks for your help! Savepos and reset will let me do exactly what I > want. It is good to know there are two selection commands as well. > > I have one more question. I can name a selection using the menu but > when I try to use the 'namesel' command I get an error: Unrecognized > command: "namesel". Does this command still exist? > > I am using production version 1 build 2304 2006/10/23 on Windows. > > Thanks again, > > Michel Dedeo > > > On 3/9/07, Tom Goddard <goddard(a)cgl.ucsf.edu> wrote: >> Hi Michel, >> >> I think Elaine's suggestion about using the savepos and reset >> commands >> is the best approach. >> >> http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/savepos.html >> http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/reset.html >> >> >> The "movemodel" command envisioned in a previous mailing list entry >> >> http://www.cgl.ucsf.edu/pipermail/chimera-users/2006-June/000843.html >> >> does not exist. I think I would make two simpler commands if some new >> ones were needed that look like: >> >> spin #1 0.3 0.4 0.5 90 20 >> >> (spin model number 1 about axis (.3,.4,.5) by 90 degrees over 20 >> frames) >> >> slide #1 5.2 8.2 3.1 20 >> >> (translate model number 1 by (5.2,8.2,3.1) over 20 frames). These >> would >> work in screen coordinates by default. But using local coordinates of >> the model or local coordinates of a second model would also be handy: >> >> spin #1 0.3 0.4 0.5 90 20 local #2 >> >> (spin about axis specified in local coordinate system of model #2) >> >> slide #1 5.2 8.2 3.1 20 local >> >> (translate with vector specified in local coordinates of the model). >> >> These local coordinate versions allow the motion to stay the same >> if you >> rotate the whole scene. Neither the spin nor slide command currently >> exists. There are move, turn, and roll commands in Chimera that have >> more limited capabilities, only allowing rotation axes and shifts >> to be >> x, y, or z screen axes. >> >> Tom >> >> >> >> >> >> >>

1 0

Chimera-users March 2007