project defaults ncs.params not read or overridden/overwritten by GUI?
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory. Here is what I did: 1. I set the NCS groups definition in the GUI to go from residues 35:445 , 2. I check that using the view pick . Everything looks fine 3. I click the Green update and exit button 4. I look at the ncs.params file in the project defaults directory and it indeed reflects the changes . 5. I start the refinement from the GUI and the ncs section seems to have been reset to 26:450 . I even tried adding the ncs.params file to the Input files section, but to no avail. The ncs seems to always get re-determined and I end up getting an excessive distance stoppage of refinement. I even made sure I dont have the Find NCS restraints automatically button checked. Any idea why I cannot get the gui run refinement to remember my ncs definitions. Thanks in advance Hari
On Mar 25, 2010, at 3:58 AM, hari jayaram wrote:
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory.
Here is what I did:
• I set the NCS groups definition in the GUI to go from residues 35:445 , • I check that using the view pick . Everything looks fine • I click the Green update and exit button • I look at the ncs.params file in the project defaults directory and it indeed reflects the changes . • I start the refinement from the GUI and the ncs section seems to have been reset to 26:450 . I even tried adding the ncs.params file to the Input files section, but to no avail. The ncs seems to always get re-determined and I end up getting an excessive distance stoppage of refinement. I even made sure I dont have the Find NCS restraints automatically button checked.
Not sure what's happening, but it looks like a bug - I'll try running something when I get into lab. Could you please send me (not the list) the .eff file written out by the GUI? (e.g. PROJECT/phenix/project_data/refine_1.eff, or something similar.) Also, a simple test: set up the refinement, then click the Save button, and send me the file that it writes. A possible fix: close the phenix.refine GUI, open a new window, add ncs.params to the input files (and make sure the find automatically option is unchecked). Does this keep the groups you've set? -------------------- Nathaniel Echols Lawrence Berkeley Lab 510-486-5136 [email protected]
Hi Hari, for whatever it's worth, if you run phenix.refine from the command line the parameter "ncs.find_automatically=False" will preserve your original NCS selections. Otherwise, "ncs.find_automatically=True", phenix.refine will use your original NCS selections as a starting point to determine the new ones. Pavel. On 3/25/10 3:58 AM, hari jayaram wrote:
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory.
Here is what I did:
1. I set the NCS groups definition in the GUI to go from residues 35:445 , 2. I check that using the view pick . Everything looks fine 3. I click the Green update and exit button 4. I look at the ncs.params file in the project defaults directory and it indeed reflects the changes . 5. I start the refinement from the GUI and the ncs section seems to have been reset to 26:450 . I even tried adding the ncs.params file to the Input files section, but to no avail. The ncs seems to always get re-determined and I end up getting an excessive distance stoppage of refinement. I even made sure I dont have the Find NCS restraints automatically button checked.
Any idea why I cannot get the gui run refinement to remember my ncs definitions.
Thanks in advance Hari
------------------------------------------------------------------------
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
I remember this. At least from running phenix.refine on command-line, there is a parameter called ncs.find_automatically. It is "True" by default. I remember spending several days trying to change the NCS selections, until I found about this parameter. So, unless you change this parameter to "False", your input is always written over. I am not sure where this might be in the GUI, but it could be the culprit. To see if this is the case, in the Refinement Settings tab, go to All Parameters -> View all settings. There, check under parameters -> refinement -> ncs. You can also find it in the refine_??.eff file, just search for this parameter. Best, Engin On 3/25/10 3:58 AM, hari jayaram wrote:
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory.
Here is what I did:
1. I set the NCS groups definition in the GUI to go from residues 35:445 , 2. I check that using the view pick . Everything looks fine 3. I click the Green update and exit button 4. I look at the ncs.params file in the project defaults directory and it indeed reflects the changes . 5. I start the refinement from the GUI and the ncs section seems to have been reset to 26:450 . I even tried adding the ncs.params file to the Input files section, but to no avail. The ncs seems to always get re-determined and I end up getting an excessive distance stoppage of refinement. I even made sure I dont have the Find NCS restraints automatically button checked.
Any idea why I cannot get the gui run refinement to remember my ncs definitions.
Thanks in advance Hari
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
-- Engin Özkan Post-doctoral Scholar Laboratory of K. Christopher Garcia Howard Hughes Medical Institute Dept of Molecular and Cellular Physiology 279 Campus Drive, Beckman Center B173 Stanford School of Medicine Stanford, CA 94305 ph: (650)-498-7111
Oh, I just saw, there is a Find NCS restraints automatically button in the Refinement Settings tab. Make sure that it is checked off. Engin On 3/25/10 9:19 AM, Engin Ozkan wrote:
I remember this. At least from running phenix.refine on command-line, there is a parameter called ncs.find_automatically. It is "True" by default. I remember spending several days trying to change the NCS selections, until I found about this parameter. So, unless you change this parameter to "False", your input is always written over. I am not sure where this might be in the GUI, but it could be the culprit. To see if this is the case, in the Refinement Settings tab, go to All Parameters -> View all settings. There, check under parameters -> refinement -> ncs. You can also find it in the refine_??.eff file, just search for this parameter.
Best, Engin
On 3/25/10 3:58 AM, hari jayaram wrote:
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory.
Here is what I did:
1. I set the NCS groups definition in the GUI to go from residues 35:445 , 2. I check that using the view pick . Everything looks fine 3. I click the Green update and exit button 4. I look at the ncs.params file in the project defaults directory and it indeed reflects the changes . 5. I start the refinement from the GUI and the ncs section seems to have been reset to 26:450 . I even tried adding the ncs.params file to the Input files section, but to no avail. The ncs seems to always get re-determined and I end up getting an excessive distance stoppage of refinement. I even made sure I dont have the Find NCS restraints automatically button checked.
Any idea why I cannot get the gui run refinement to remember my ncs definitions.
Thanks in advance Hari
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
-- Engin Özkan Post-doctoral Scholar Laboratory of K. Christopher Garcia Howard Hughes Medical Institute Dept of Molecular and Cellular Physiology 279 Campus Drive, Beckman Center B173 Stanford School of Medicine Stanford, CA 94305 ph: (650)-498-7111
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
-- Engin Özkan Post-doctoral Scholar Laboratory of K. Christopher Garcia Howard Hughes Medical Institute Dept of Molecular and Cellular Physiology 279 Campus Drive, Beckman Center B173 Stanford School of Medicine Stanford, CA 94305 ph: (650)-498-7111
On Mar 25, 2010, at 9:19 AM, Engin Ozkan wrote:
I remember this. At least from running phenix.refine on command-line, there is a parameter called ncs.find_automatically. It is "True" by default. I remember spending several days trying to change the NCS selections, until I found about this parameter. So, unless you change this parameter to "False", your input is always written over. I am not sure where this might be in the GUI, but it could be the culprit. To see if this is the case, in the Refinement Settings tab, go to All Parameters -> View all settings. There, check under parameters -> refinement -> ncs. You can also find it in the refine_??.eff file, just search for this parameter.
For the record, a related point: if you're used to the command-line version and/or want to confirm that the final parameters are exactly what you intended, you can choose to edit the final .eff file generated by the GUI *before* it actually runs phenix.refine. Go to Preferences->Refinement, and check the box "Edit parameters before running." It will open a text editor with the .eff file, and you should be able to change anything in it. (One word of caution: if you have checked "Add hydrogens automatically", this will be done before you can edit the file, and the PDB file(s) in refinement.input.pdb.file_name will be different than the file(s) shown in the GUI. This is intentional, but potentially confusing. It also means that if you restore the job later, the PDB file that will be listed is now the one with hydrogens, not what was originally entered. Also confusing, but it shouldn't break anything - hopefully the handling of hydrogens will be cleaned up in the next month or two.) -Nat PS. I will see if I can make it a little more obvious what to do about the "Find automatically" option. ------------------- Nathaniel Echols Lawrence Berkeley Lab 510-486-5136 [email protected]
On Mar 25, 2010, at 3:58 AM, hari jayaram wrote:
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory.
Any idea why I cannot get the gui run refinement to remember my ncs definitions.
Aha - the problem is that the parameters I'm getting from the NCS group editor are incomplete. The file it writes out will look like this: restraint_group { reference = "chain A and resseq 1:100" selection = "chain B and resseq 1:100" coordinate_sigma = 0.05 b_factor_weight = 10 } It needs the fully qualified "path" of the restraint_group section, however. Edit ncs.params to look like this: refinement.ncs { restraint_group { [snip] } } That should fix the problem (and I'll see if we can get a patched installer built ASAP). Sorry, I think I broke this last month attempting to repair something else. -Nat ------------------- Nathaniel Echols Lawrence Berkeley Lab 510-486-5136 [email protected]
Thanks a lot Nat , I will manually edit the ncs.params file to add the
correct syntax for the ncs group
Also thanks for the tips on checking and editing the *.eff file before
running.
Once you have a patched nightly , I will switch to that build..
Thanks again
Hari
On Thu, Mar 25, 2010 at 1:01 PM, Nathaniel Echols
On Mar 25, 2010, at 3:58 AM, hari jayaram wrote:
Hi I am quite sure I am doing something wrong , but the refinement seems to always re-determine the NCS definitions despite it being set in the ncs.params file in the project defaults directory.
Any idea why I cannot get the gui run refinement to remember my ncs definitions.
Aha - the problem is that the parameters I'm getting from the NCS group editor are incomplete. The file it writes out will look like this:
restraint_group { reference = "chain A and resseq 1:100" selection = "chain B and resseq 1:100" coordinate_sigma = 0.05 b_factor_weight = 10 }
It needs the fully qualified "path" of the restraint_group section, however. Edit ncs.params to look like this:
refinement.ncs { restraint_group { [snip] } }
That should fix the problem (and I'll see if we can get a patched installer built ASAP). Sorry, I think I broke this last month attempting to repair something else.
-Nat
------------------- Nathaniel Echols Lawrence Berkeley Lab 510-486-5136 [email protected]
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
participants (4)
-
Engin Ozkan -
hari jayaram -
Nathaniel Echols -
Pavel Afonine