ESP grid is offset from original PDB coordinates??
I was playing with esp results yesterday from Delphi on several structures, and ran a quick calculation on 2ace. To my surprise, the results as visualized as a solid under volume viewer were very obviously offset from the structure itself. They stand alone to one side of the original. I've attached a small jpg of this. I'm using the latest Linux production version (beta version 1 build 2415) on an amd64 box (Debian Sid, kernel 2.6.20). I hope there is a simple explanation... ;-) I am no expert with Delphi, so something may be odd there. I used a couple of the siz/crg files, with the same results in each case, so I suspect something in the original pdb file is messing things up? Cheers! Kenward -- .'^~;,_ Dr. Kenward Vaughan `:,'~~~~~ Professor of Chemistry \;:/ Bakersfield College |,;| 1801 Panorama Drive / ', \ Bakersfield, CA 93305 / o O \ http://www2.bc.cc.ca.us/kvaughan (oOoOOoOo) ---========--- ???$$MM$$???
On May 29, 2007, at 8:52 AM, Kenward Vaughan wrote:
I was playing with esp results yesterday from Delphi on several structures, and ran a quick calculation on 2ace. To my surprise, the results as visualized as a solid under volume viewer were very obviously offset from the structure itself. They stand alone to one side of the original. I've attached a small jpg of this.
Hi Kenward, I tried calculating ESP for 2ace with DelPhiController in 1.2415 on a couple of platforms (albeit neither linux) but could not reproduce the problem. In this version of Chimera, Volume Viewer does not automatically appear when the ESP is calculated, but if I do open Volume Viewer and display surface or solid, the "blobs" are all on in the same place as the protein. I attached an example of the "surface" display in Volume Viewer at the default contour level (the "solid" display also overlaps the protein, but the spots are fainter at the default contour level). I just used the default radius and charge files and run parameters in delphi. Here are some things to check: - were the models separately transformed? Maybe if you use the "reset" command, they will come back into register. - is the grid box surrounding the protein or is it totally offset? Show a box outline in Volume Viewer (Features... Data display options, check the "Show outline box" option) to check this. - are you using DelPhiController? if so, the grid should automatically center on the molecule. If you are running delphi in standalone fashion, however, there are some parameters exposed that could give you an offset that is not centered on the protein - make sure you aren't specifying such offsets in the delphi input file. Other than that, I'm stumped... Elaine ----- Elaine C. Meng, Ph.D. meng@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 Tue, 2007-05-29 at 10:15 -0700, Elaine Meng wrote:
On May 29, 2007, at 8:52 AM, Kenward Vaughan wrote:
I was playing with esp results yesterday from Delphi on several structures, and ran a quick calculation on 2ace. To my surprise, the results as visualized as a solid under volume viewer were very obviously offset from the structure itself. They stand alone to one side of the original. I've attached a small jpg of this.
Hi Kenward, I tried calculating ESP for 2ace with DelPhiController in 1.2415 on a couple of platforms (albeit neither linux) but could not reproduce the problem. In this version of Chimera, Volume Viewer does not automatically appear when the ESP is calculated, but if I do open Volume Viewer and display surface or solid, the "blobs" are all on in the same place as the protein. I attached an example of the "surface" display in Volume Viewer at the default contour level (the "solid" display also overlaps the protein, but the spots are fainter at the default contour level). I just used the default radius and charge files and run parameters in delphi.
Here are some things to check: - were the models separately transformed? Maybe if you use the "reset" command, they will come back into register.
- is the grid box surrounding the protein or is it totally offset? Show a box outline in Volume Viewer (Features... Data display options, check the "Show outline box" option) to check this.
- are you using DelPhiController? if so, the grid should automatically center on the molecule. If you are running delphi in standalone fashion, however, there are some parameters exposed that could give you an offset that is not centered on the protein - make sure you aren't specifying such offsets in the delphi input file.
Other than that, I'm stumped... Elaine
Hmm. Did I mention IANADPE (I Am Not A Delphi Expert.)? I turned around and tried to duplicate it, with no success. Since I know I didn't hallucinate the event, I backtracked a bit. Changing the settings in DelphiController didn't seem to help. I then went further back into my memory of that particular run, and realized where the issue lay. I had done the same on another structure, then closed the session, then opened 2ace.pdb. Repeating this sequence recreated the problem. I have not exhaustively tried various examples, but for at least a subset of files, the following replicates the problem: 1) open a pdb and run Delphi, followed by Volume Viewer to see the results, 2) close the session (but not Chimera), 3) repeat #1 on another pdb file. I don't know if it has to do with not clearing out the old default.phi file or perhaps some other internal variable, but at least I'm at ease now... ;-) BTW, I've found that selecting a chain the first time after opening a pdb file causes the backgound to turn green (with envy?). This clears after a second chain selection... HTH, Kenward -- .'^~;,_ Dr. Kenward Vaughan `:,'~~~~~ Professor of Chemistry \;:/ Bakersfield College |,;| 1801 Panorama Drive / ', \ Bakersfield, CA 93305 / o O \ http://www2.bc.cc.ca.us/kvaughan (oOoOOoOo) ---========--- ???$$MM$$???
Hi Kenward, I suspect that your DelPhi run on the first PDB worked, but the second one with 2ace failed. But Chimera mistakenly thinks the second run worked and opens the DelPhi map from the first run. I checked the Chimera DelPhi Controller code and unfortunately it does just check if the map file exists rather than checking the exit status of DelPhi. I'll file this as a Chimera bug. Tom
Hi Kenward, Aha! It turns out that when you use "Close session," the Input "Molecule" in DelPhiController is not automatically cleared. When you open subsequent structures, you have to go to the Input section of DelPhiController and click "Browse" again to use a structure that is actually open. Otherwise, it will keep calculating the ESP for your original model even though it was closed already (I guess the information has been stored somewhere). If you didn't reset the Input, you are probably seeing the ESP for the first structure on which you used DelPhiController since Chimera was started, even though it's a new map with the new name indicated in the Output section. I would classify this as a Chimera bug, or at least bad behavior, rather than user error! Thanks for reporting this, Elaine ----- Elaine C. Meng, Ph.D. meng@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 May 30, 2007, at 12:03 AM, Kenward Vaughan wrote:
Hmm. Did I mention IANADPE (I Am Not A Delphi Expert.)?
I turned around and tried to duplicate it, with no success. Since I know I didn't hallucinate the event, I backtracked a bit. Changing the settings in DelphiController didn't seem to help.
I then went further back into my memory of that particular run, and realized where the issue lay. I had done the same on another structure, then closed the session, then opened 2ace.pdb. Repeating this sequence recreated the problem.
I have not exhaustively tried various examples, but for at least a subset of files, the following replicates the problem:
1) open a pdb and run Delphi, followed by Volume Viewer to see the results,
2) close the session (but not Chimera),
3) repeat #1 on another pdb file.
I don't know if it has to do with not clearing out the old default.phi file or perhaps some other internal variable, but at least I'm at ease now... ;-)
BTW, I've found that selecting a chain the first time after opening a pdb file causes the backgound to turn green (with envy?). This clears after a second chain selection...
HTH,
Kenward
On May 30, 2007, at 12:03 AM, Kenward Vaughan wrote:
BTW, I've found that selecting a chain the first time after opening a pdb file causes the backgound to turn green (with envy?). This clears after a second chain selection...
We haven't seen this behavior on any machine we have. It almost certainly is a graphics driver problem. You should check to see if you have the latest driver for your card. If you have a laptop, you should check your machine's manufacturer's web site -- otherwise check the graphic card's manufacturer's web site. If it turns out that you do have the latest driver, it may be possible to improve things by changing settings on your card. I can provide some guidance if it turns out you need to do this. --Eric Eric Pettersen UCSF Computer Graphics Lab pett@cgl.ucsf.edu http://www.cgl.ucsf.edu
participants (4)
-
Elaine Meng -
Eric Pettersen -
Kenward Vaughan -
Tom Goddard