- Chimera-users - RBVI Mailing Lists

Re: [Chimera-users] chimera and hydrophobic potentials
by Thomas Goddard 09 Apr '04

09 Apr '04

Hi Miguel, I wrote a conversion program cns2delphi.py that takes the k2f output and produces a delphi file and tried it on the data you sent. The delphi file is read properly by Chimera DelphiViewer and can be used to color an MSMS surface. A problem with the data you sent is that the hydrophobicity map is zero over the entire MSMS surface. There are blobs of hydrophobicity a few angstroms away from the surface but they don't touch the surface. I observed this by showing the map with the Chimera volume viewer (just use File/Open...). You can change the probe radius for generating the MSMS surface by going to Favorites/Model Panel, selecting the MSMS surface, clicking the Attributes button, changing the probe radius setting, and pressing Enter. (You need to set the surface color to a solid color before changing the radius or an error message is generated.) Unfortunately that does not push the molecular surface out uniformly -- only the surface in the cavities is pushed further from the molecule. Apparently the probe radius is the radius of a ball rolled over the molecule with standard atom radii. The MSMS surface is the surface of the region that cannot be touched by the ball. So a bigger ball does not change the surface location at protrusions. Maybe a better approach is to specify larger atom radii. Below is the cns2delphi.py Python program. It requires the Numeric Python module which is not a standard part of Python. The Python that comes with Chimera includes Numeric. You can convert k2f output to delphi format as follows. % /usr/local/chimera/bin/python2.3 cns2delphi.py 1ltlA.k2f 1ltlA.phi This creates the 1ltlA.phi file. To use the Chimera DelphiViewer extension, start Chimera, open 1ltlA.pdb, make a molecular surface with Actions/Surface/Show, start Extensions/Delphi/DelphiViewer, enter Potential File 1ltlA.phi, click the Browse button next to MSMS Surface in the Molecular surface pane so the surface is listed, then press Apply in the Molecular surface pane. This will color the surface. You can also just use File/Open... on 1ltlA.phi to see an isosurface of the hydrophobicity map. You can drag the vertical bar on the histogram shown in the volume viewer dialog to change the isosurface threshold level. Tom ----- cns2delphi.py follows. Preserve indentation since it is important in Python. #!/usr/bin/env python # ---------------------------------------------------------------------------- # Convert CNS volume to Delphi format. # # ---------------------------------------------------------------------------- # Returns matrix, xyz_origin, and xyz_step # def read_cns_file(path): f = open(path, 'r') f.readline() # First line is blank ntitle_line = f.readline() # integer number of comment lines try: ntitle = int(ntitle_line.split()[0]) except ValueError: raise SyntaxError, ('Invalid CNS comment line count on line 2: %s' % ntitle_line[:80]) for t in range(ntitle): f.readline() extent_line = f.readline() extent = split_fields(extent_line, 8, 9) try: na, amin, amax, nb, bmin, bmax, nc, cmin, cmax = map(int, extent) except ValueError: raise SyntaxError, 'Invalid CNS grid size line: %s' % extent_line[:80] cell_line = f.readline() cell = split_fields(cell_line, 12, 6) try: cell_params = map(float, cell) except ValueError: raise SyntaxError, ('Invalid CNS cell parameters line: %s' % cell_line[:80]) cell_size = cell_params[0:3] cell_angles = cell_params[3:6] axis_order_line = f.readline() # Must be ZYX asize, bsize, csize = (amax - amin + 1, bmax - bmin + 1, cmax - cmin + 1) if asize <= 0 or bsize <= 0 or csize <=0: raise SyntaxError, ('Bad CNS grid size (%d,%d,%d)' % (asize, bsize, csize)) import Numeric data = Numeric.zeros((csize, bsize, asize), Numeric.Float32) for c in range(csize): read_floats(f, 1) # c section number values = read_floats(f, asize * bsize) array = Numeric.array(values, Numeric.Float32) plane = Numeric.reshape(array, (bsize, asize)) data[c,:,:] = plane f.readline() # footer - int value must be -9999 f.readline() # density average and standard dev f.close() xyz_step = (cell_size[0]/na, cell_size[1]/nb, cell_size[2]/nc) xyz_origin = map(lambda gmin, step: gmin * step, (amin, bmin, cmin), xyz_step) return data, xyz_origin, xyz_step # ----------------------------------------------------------------------------- # Read ascii float values on as many lines as needed to get count values. # def read_floats(f, count): values = [0.0] * count c = 0 while c < count: line = f.readline() if line == '': raise FileFormatError('File is shorter than expected') fields = split_fields(line, 12, 6) if c + len(fields) > count: fields = fields[:count-c] for v in map(float, fields): values[c] = v c = c + 1 return values # ----------------------------------------------------------------------------- # def split_fields(line, field_size, max_fields): fields = [] for k in range(0, len(line), field_size): f = line[k:k+field_size].strip() if f: fields.append(f) else: break return fields[:max_fields] # ---------------------------------------------------------------------------- # def write_delphi_file(path, matrix, xyz_center, scale): zsize, ysize, xsize = matrix.shape if xsize != ysize or xsize != zsize: raise ValueError, "Matrix size (%d,%d,%d) has to be the same in all 3 dimensions" % (zsize, ysize, xsize) size = xsize uplbl = delphi_record('Converted from ' + path) morelabels = delphi_record('%70s' % '') import Numeric data_flat = Numeric.ravel(matrix) data = delphi_record(floats_as_string(data_flat)) botlbl = delphi_record('') params = delphi_record(floats_as_string((scale,) + tuple(xyz_center)) + integer_as_string(size)) f = open(path, 'wb') f.write(uplbl + morelabels + data + botlbl + params) f.close() # ---------------------------------------------------------------------------- # def delphi_record(data): size_str = integer_as_string(len(data)) r = size_str + data + size_str return r # ---------------------------------------------------------------------------- # def integer_as_string(i): import Numeric return Numeric.array((i,), Numeric.Int32).tostring() # ---------------------------------------------------------------------------- # def floats_as_string(f): import Numeric return Numeric.array(f, Numeric.Float32).tostring() # ---------------------------------------------------------------------------- # def cns_to_delphi_file(cns_path, delphi_path): data, xyz_origin, xyz_step = read_cns_file(cns_path) if xyz_step[0] != xyz_step[1] or xyz_step[0] != xyz_step[2]: raise Value_Error, 'Step sizes (%.2g, %.2g, %.2g) are not the same along all axes' % xyz_step scale = 1.0 / xyz_step[0] size = list(data.shape) size.reverse() xyz_center = map(lambda o, st, sz: o + st * ((sz-1)/2), xyz_origin, xyz_step, size) write_delphi_file(delphi_path, data, xyz_center, scale) # ---------------------------------------------------------------------------- # if __name__ == '__main__': import sys if len(sys.argv) != 3: sys.stderr.write('Syntax: cns2delphi <cnsfile> <delphifile>\n') sys.exit(1) cns_path = sys.argv[1] delphi_path = sys.argv[2] cns_to_delphi_file(cns_path, delphi_path)

1 0

Re: [Chimera-users] chimera and hydrophobic potentials
by Thomas Goddard 08 Apr '04

08 Apr '04

Hi Miguel, Send me some example k2a output and I'll send you a little Python program that will convert it into a Delphi file readable by Chimera DelphiViewer. Tom > Date: Thu, 8 Apr 2004 19:33:31 +0100 > From: Miguel Ortiz =?iso-8859-15?Q?Lombard=EDa?= <mol(a)ysbl.york.ac.uk> > Subject: Re: [Chimera-users] chimera and hydrophobic potentials > > Hi Tom, > > Thank you very much for your comments. > Part of the problem is that I don't trust the "delphi" maps that > GRID produces. Actually, I've been able to open some of them with > chimera. Other programs ablo to open such maps (like vmd) can't > open any of them. > > Yes, I've used k2a, and it works: it produces an ascii map that > looks sensible. My problem here is that my programming skills are > very reduced! I could work it out on a "model" program, but not > from scratch. Have you something like that? > > Thank you again, your answer has been very useful. > > Best regards, > > Miguel > > -- > Miguel Ortiz Lombardía > email: mol(a)ysbl.york.ac.uk > http://www.ysbl.york.ac.uk/~mol > ----------------------------------------------------------------------- > Our business in life is not to succeed, but > to continue to fail in good spirits. R.L. Stevenson >

1 0

chimera and hydrophobic potentials
by Miguel Ortiz Lombardía 08 Apr '04

08 Apr '04

Hi, I would like to display the hydrophobic potential of a protein onto its MSMS surface. One program that calculates such potentials is GRID (from Molecular Discovery: http://www.moldiscovery.com/ ) GRID can produce the map in different formats. One of them is like delphi (grasp mode) map. This has the drawback that the map has to be 65x65x65 grid points and the grid points separation has to be an integer starting with 1 Ansgtroem. Hence, it's not appropriate for medium-big proteins. The GRID format hasn't those limitations. I can __sometimes__ read in chimera (throught the DelphiViewer) the delphi maps produced by GRID. Sometimes they are unreadable (and this seems to me more a problem with GRID than with chimera: something wrong in the header) But even when they are readable I'm not sure they are covering the region they are supposed to be covering. Does anyone know if chimera can read any other of the GRID map formats: GRID (this would be the optimal) Chem-X Frodo Insight Molcad Quanta Sybyl Otherwise: does anybody how to the kind of representation I'm looking for in any other way? Thank you for any advice! Miguel -- Miguel Ortiz Lombardía email: mol(a)ysbl.york.ac.uk http://www.ysbl.york.ac.uk/~mol ----------------------------------------------------------------------- Our business in life is not to succeed, but to continue to fail in good spirits. R.L. Stevenson

2 1

compiling minrms on mac os10.2.8
by Chris Adams 02 Apr '04

02 Apr '04

Hi: I am also trying to get minrms on my OS10.2.8 machine. I followed the commands you recommended to Howard to get the "make" command to work. However, my ls output was not at all like what you said would happen in the message. I get this: ls -1 /usr/bin/*make /usr/bin/automake and am unable to compile the minrms program. I am not very unix savvy but am learning, perhaps you can help? thank you, Chris Adams UW-Biochemistry Madison, WI 53704

2 1

correction re sessions
by Elaine Meng 31 Mar '04

31 Mar '04

Hello, in my previous message about saving parts of PDB files, I included some text from older replies to users, and this part about sessions is no longer true: >The command "save" (or File... Save Session As... ) doesn't save coordinates, >only how they are positioned/colored/displayed, etc. In fact, restarting >a saved session will use the original input file(s) for coordinates. Actually in the last few versions, the session files do include coordinates and do not look for the original input files. Sorry about that, Elaine

1 0

writing out part of a PDB
by Elaine Meng 31 Mar '04

31 Mar '04

Hi Shuanghong, Question (1), surface problem: I don't think the problem where nothing happens when you type the command "surface" has been reported. Sometimes there is a delay while the surface is being calculated, but it shouldn't be terribly long. Problems have been reported in some cases, but they generally cause some sort of error message. To register a problem more formally, you could send mail to chimera-bugs(a)cgl.ucsf.edu and be sure to include what version of Chimera and what kind of computer you are using, and what structure(s) cause the problem. Question (2), writing out part of a PDB file: If you are on a **non-Windows** system, you can use the command "pdbrun" to save just the atoms that are displayed. You could display only the atoms you want by first selecting what you want to measure the distance to. Then, use Select...Zone... to select all the atoms in the region you want to write out. Then the command "show sel" will undisplay the rest of the stuff. (Or, instead of selecting you could do it all in one command, for example show :212.a za<4.5 for all atoms within 4.5 A of any atom in residue 212 of chain A. Specifying zones in the command line is described in the manual page on atom specification, which is listed in the Basic Functions index.) Next, use pdbrun (see man page using the command "help pdbrun"), for example: pdbrun nouser cat > myzone.pdb Pdbrun literally sends the information for only the displayed atoms in their displayed positions to another command, in this case "cat" to send the information to the file myzone.pdb. If you want the coordinates to be the same as in the original file, first use the command "reset." The command "save" (or File... Save Session As... ) doesn't save coordinates, only how they are positioned/colored/displayed, etc. In fact, restarting a saved session will use the original input file(s) for coordinates. The command "write" (same as the function "write PDB" in the Model Panel) saves coordinates into a file. However, these will save the entire coordinates whether displayed or not. If you are on **Windows** the only way I can think of is to actually delete the atoms/bonds you don't want to write out ("delete" command or Actions... Atoms/Bonds... delete). For example, when only what you want to write out is displayed, delete @/!display actually deletes the other stuff. Subsequently, a file created with "write" would only contain the atoms/bonds that had not been deleted. I hope this helps, Elaine =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Elaine C. Meng, Ph.D. Computer Graphics Lab and Babbitt Lab meng(a)cgl.ucsf.edu http://www.cgl.ucsf.edu/home/meng/index.html =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

1 0

crashing when calculating larger surfaces
by Elaine Meng 31 Mar '04

31 Mar '04

>From eperozo(a)virginia.edu Wed Mar 31 15:15:53 2004 >Sorry to bother you, but I'm having terrible problems with the generation of >surfaces. The program crashes as soon as I request a surface to be built (a >tetramer with ~250 residues/monomer). Interestingly, I don't seem to have >problems with a smaller protein (120 residues). When it crashes it gives me >exit code - 1073741819. Tried two versions of chimera in two different >computers. Same result. >Is this a known bug? >Thanks in advance. >Eduardo Hi Eduardo, It is a known problem that the surface routine is a memory hog. If you are on a unix-type machine it might help to type "unlimit" in the shell before starting Chimera. Otherwise, it might just be trying to use more memory than your machines have. We have been using code developed elsewhere, MSMS, for these surfaces. Due to memory/robustness problems, we are generating some code in-house which will eventually replace MSMS. However, this is still in progress, and has not been incorporated into any Chimera release. You could try deleting extra monomers before using surface or instead show the low-resolution surface generated by the Multiscale Models tool (although it builds multimers with matrix info, even if there is no matrix info, it will still show the low-res surface for the structure you have open). Tools...Multiscale... Multiscale Models, click the "Make models" button. Then if you select All chains (button near the top) you can change the Resolution value and then click Resurface to see different levels of detail. There is a man page for this tool in the Tools section of the User's Guide if you want more info. If you want to register the problem more formally, please send a message to chimera-bugs(a)cgl.ucsf.edu including version, platform, and an example or few of which structures (e.g. their PDB IDs) cause the problem. Best, Elaine

1 0

Re: [Chimera-users] Rotating molecules
by Elaine Meng 30 Mar '04

30 Mar '04

Hi Eduardo, yes, the command "move" does translations by a specified amount and the command "turn" rotates by a specified amount. Type help move in the command line to see the man page for move, and help turn for the man page for turn. You might also want to take a look at the commands "rock" and "roll" Hope this helps, Elaine =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Elaine C. Meng, Ph.D. Computer Graphics Lab and Babbitt Lab meng(a)cgl.ucsf.edu http://www.cgl.ucsf.edu/home/meng/index.html =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

1 0

Rotating molecules
by Eduardo Perozo 30 Mar '04

30 Mar '04

Hello, Is there a way to rotate/translate a molecule by a specific amount? Clearly I can do "freehand" rotations with the mouse but I will like to define rotations and translation numerically. Thanks! Eduardo

1 0

Re: [Chimera-users] export pdb file
by Elaine Meng 29 Mar '04

29 Mar '04

Hi Shuanghong, there are two ways to export a PDB file. One is by choosing the "write PDB" function from the Model Panel. The other is by using the "write" command. Each way has an option to save the transformed coordinates or the coordinates relative to the untransformed coordinates of a particular model (take a look at the User's Guide for detailed descriptions). Hope this helps! Elaine =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Elaine C. Meng, Ph.D. Computer Graphics Lab and Babbitt Lab meng(a)cgl.ucsf.edu http://www.cgl.ucsf.edu/home/meng/index.html =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

2 1