can phenix.refine create a diff of the current input file with the pdb from the last run before refining?
Hi all, Just a feature request that phenix.refine make a diff comparing the current input file with the pdb from the last run and then doing a refinement on the input file and then allow a phenix.refine --revert to allow the user to backtrack incase the refinement isn't suitable. Right now I edit the pdb, save it as xxxx_001-input.pdb and specify xxxx_001-input.pdb as the input file in xxxx_002.def. (Yes I'm using the command line, does the GUI allow this? (i.e. right click on a refinement run then click revert?) ) F --------------------------------------------- Francis Reyes M.Sc. 215 UCB University of Colorado at Boulder gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D 8AE2 F2F4 90F7 9640 28BC 686F 78FD 6669 67BA 8D5D
Hi Francois, you have your two PDB files: before and after phenix.refine run. So in case you did not like the refinement outcome you always have your input PDB file, the one you fed into phenix.refine. I'm not sure what exactly you mean by "--revert"... Could you please explain in more details? Thanks! Pavel. On 4/21/10 6:02 AM, Francis E Reyes wrote:
Hi all,
Just a feature request that phenix.refine make a diff comparing the current input file with the pdb from the last run and then doing a refinement on the input file and then allow a phenix.refine --revert to allow the user to backtrack incase the refinement isn't suitable.
Right now I edit the pdb, save it as xxxx_001-input.pdb and specify xxxx_001-input.pdb as the input file in xxxx_002.def.
(Yes I'm using the command line, does the GUI allow this? (i.e. right click on a refinement run then click revert?) )
F
--------------------------------------------- Francis Reyes M.Sc. 215 UCB University of Colorado at Boulder
gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D
8AE2 F2F4 90F7 9640 28BC 686F 78FD 6669 67BA 8D5D
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
Pavel Well here's a workflow. phenix.refine model.pdb creates model_001.pdb . I edit model_001.pdb ... the question is what do I save it as? If I were to follow the convention of the model_002.def file, the suggestion is to save it as model_001.pdb (so I don't have to edit the input on model_002.def). I run model_002.def and realize the edits that I made in model_001.pdb made my statistics worse. How do I get back model_001.pdb ? do I rerun phenix.refine model.pdb to get model_001.pdb? Seems like a lot of work / time . My current workflow for being able to revert is to save model_001.pdb as model_001-input.pdb and then change model_002.def to compensate. Lots of vim keystrokes, and the possibility of forgetting to edit model_XXX.def. If I had a diff file that allowed me to revert back to obtain the original model_001.pdb, this would be ideal Does this make sense? F On Apr 21, 2010, at 7:10 AM, Pavel Afonine wrote:
Hi Francois,
you have your two PDB files: before and after phenix.refine run. So in case you did not like the refinement outcome you always have your input PDB file, the one you fed into phenix.refine. I'm not sure what exactly you mean by "--revert"... Could you please explain in more details?
Thanks! Pavel.
On 4/21/10 6:02 AM, Francis E Reyes wrote:
Hi all,
Just a feature request that phenix.refine make a diff comparing the current input file with the pdb from the last run and then doing a refinement on the input file and then allow a phenix.refine --revert to allow the user to backtrack incase the refinement isn't suitable. Right now I edit the pdb, save it as xxxx_001-input.pdb and specify xxxx_001-input.pdb as the input file in xxxx_002.def. (Yes I'm using the command line, does the GUI allow this? (i.e. right click on a refinement run then click revert?) )
F
--------------------------------------------- Francis Reyes M.Sc. 215 UCB University of Colorado at Boulder
gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D
8AE2 F2F4 90F7 9640 28BC 686F 78FD 6669 67BA 8D5D
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
--------------------------------------------- Francis Reyes M.Sc. 215 UCB University of Colorado at Boulder gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D 8AE2 F2F4 90F7 9640 28BC 686F 78FD 6669 67BA 8D5D
Hi Francois, It is a bit more work we have time for right now. But if I understand you correctly this would almost be the same: cp model_001.pdb model_001.pdb_as_refined edit model_001.pdb phenix.refine model_002.def # ... doesn't go well cp model_001.pdb_as_refined model_001.pdb I.e. just make a backup of model_001.pdb before you modify it. Ralf
Ralf Thanks for the advice. I admit the request should go to the bottom of any priority list. F On Apr 21, 2010, at 11:08 AM, Ralf W. Grosse-Kunstleve wrote:
Hi Francois,
It is a bit more work we have time for right now. But if I understand you correctly this would almost be the same:
cp model_001.pdb model_001.pdb_as_refined edit model_001.pdb phenix.refine model_002.def # ... doesn't go well cp model_001.pdb_as_refined model_001.pdb
I.e. just make a backup of model_001.pdb before you modify it.
Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
--------------------------------------------- Francis Reyes M.Sc. 215 UCB University of Colorado at Boulder gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D 8AE2 F2F4 90F7 9640 28BC 686F 78FD 6669 67BA 8D5D
I see your point. I have a similar conundrum. It is a very minor point, and I'm sure there are many other pressing issues for the developers. phenix.refine model_refine_001.def creates model_refine_001.pdb I edit in Coot and save as model_001-coot-0.pdb model_refine_002.def is expecting to see: file_name = "model_refine_001.pdb" Therefore I have to edit the def file to: file_name = "model_refine_001-coot-0.pdb" Perhaps there is some way to tell phenix.refine the preferred filename pattern. Like adding an addition output value (i.e. suffix). ~~~~~~~~~ snippet from theoretical def file ~~~~~~~~~~~~~~~ output { prefix = "model_refine" pdb_input_suffix = "editted" serial = 1 serial_format = "%03d" would produce: file_name = "model_refine_001_editted.pdb" in the next def file Scott On Apr 21, 2010, at 6:18 AM, Francis E Reyes wrote:
Pavel
Well here's a workflow.
phenix.refine model.pdb creates model_001.pdb .
I edit model_001.pdb ... the question is what do I save it as? If I were to follow the convention of the model_002.def file, the suggestion is to save it as model_001.pdb (so I don't have to edit the input on model_002.def). I run model_002.def and realize the edits that I made in model_001.pdb made my statistics worse. How do I get back model_001.pdb ? do I rerun phenix.refine model.pdb to get model_001.pdb? Seems like a lot of work / time .
My current workflow for being able to revert is to save model_001.pdb as model_001-input.pdb and then change model_002.def to compensate. Lots of vim keystrokes, and the possibility of forgetting to edit model_XXX.def.
If I had a diff file that allowed me to revert back to obtain the original model_001.pdb, this would be ideal
Does this make sense?
F
On Apr 21, 2010, at 7:10 AM, Pavel Afonine wrote:
Hi Francois,
you have your two PDB files: before and after phenix.refine run. So in case you did not like the refinement outcome you always have your input PDB file, the one you fed into phenix.refine. I'm not sure what exactly you mean by "--revert"... Could you please explain in more details?
Thanks! Pavel.
On 4/21/10 6:02 AM, Francis E Reyes wrote:
Hi all,
Just a feature request that phenix.refine make a diff comparing the current input file with the pdb from the last run and then doing a refinement on the input file and then allow a phenix.refine --revert to allow the user to backtrack incase the refinement isn't suitable. Right now I edit the pdb, save it as xxxx_001-input.pdb and specify xxxx_001-input.pdb as the input file in xxxx_002.def. (Yes I'm using the command line, does the GUI allow this? (i.e. right click on a refinement run then click revert?) )
F
-----Original Message----- From: [email protected] [mailto:phenixbb- [email protected]] On Behalf Of Scott Classen Sent: Thursday, April 22, 2010 1:55 PM To: PHENIX user mailing list Subject: Re: [phenixbb] can phenix.refine create a diff of the current inputfile with the pdb from the last run before refining?
I see your point. I have a similar conundrum. It is a very minor
and I'm sure there are many other pressing issues for the developers.
phenix.refine model_refine_001.def creates model_refine_001.pdb
I edit in Coot and save as model_001-coot-0.pdb
model_refine_002.def is expecting to see:
file_name = "model_refine_001.pdb"
Therefore I have to edit the def file to:
file_name = "model_refine_001-coot-0.pdb"
Perhaps there is some way to tell phenix.refine the preferred filename pattern. Like adding an addition output value (i.e. suffix).
~~~~~~~~~ snippet from theoretical def file ~~~~~~~~~~~~~~~ output { prefix = "model_refine" pdb_input_suffix = "editted" serial = 1 serial_format = "%03d"
would produce:
file_name = "model_refine_001_editted.pdb" in the next def file
Scott
On Apr 21, 2010, at 6:18 AM, Francis E Reyes wrote:
Pavel
Well here's a workflow.
phenix.refine model.pdb creates model_001.pdb .
I edit model_001.pdb ... the question is what do I save it as? If I were to follow the convention of the model_002.def file, the suggestion is to save it as model_001.pdb (so I don't have to edit the input on model_002.def). I run model_002.def and realize the edits that I made in model_001.pdb made my statistics worse. How do I get back model_001.pdb ? do I rerun phenix.refine model.pdb to get model_001.pdb? Seems like a lot of work / time .
My current workflow for being able to revert is to save model_001.pdb as model_001-input.pdb and then change model_002.def to compensate. Lots of vim keystrokes, and the possibility of forgetting to edit model_XXX.def.
If I had a diff file that allowed me to revert back to obtain the original model_001.pdb, this would be ideal
Does this make sense?
F
On Apr 21, 2010, at 7:10 AM, Pavel Afonine wrote:
Hi Francois,
you have your two PDB files: before and after phenix.refine run. So in case you did not like the refinement outcome you always have your input PDB file, the one you fed into phenix.refine. I'm not sure what exactly you mean by "--revert"... Could you please explain in more details?
Thanks! Pavel.
On 4/21/10 6:02 AM, Francis E Reyes wrote:
Hi all,
Just a feature request that phenix.refine make a diff comparing
I think the underlying problem is that Phenix ADDS on PDB files defined at the commandline to whatever is defined in the .def file. I mentioned this a couple of year back to Ralf as an improvement request, but did not get very far. If one could force Phenix only to only read the pdb file defined on the command line and ignore the one in the *002.def file these kind of minor headaches would go away. point, the
current input file with the pdb from the last run and then doing a refinement on the input file and then allow a phenix.refine --revert to allow the user to backtrack incase the refinement isn't suitable.
Right now I edit the pdb, save it as xxxx_001-input.pdb and specify xxxx_001-input.pdb as the input file in xxxx_002.def. (Yes I'm using the command line, does the GUI allow this? (i.e. right click on a refinement run then click revert?) )
F
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb
I think the underlying problem is that Phenix ADDS on PDB files defined at the commandline to whatever is defined in the .def file. I mentioned this a couple of year back to Ralf as an improvement request, but did not get very far. If one could force Phenix only to only read the pdb file defined on the command line and ignore the one in the *002.def file these kind of minor headaches would go away.
It is tricky to support both input of multiple pdb files and easily overriding the pdb file name in the .def file. Maybe I've made the wrong choice when deciding which of the two features is more important? How often do people want to give multiple pdb files to phenix.refine? If that's less common than the wish to override the name in the .def file we could change phenix.refine to only allow one file name. It would only use the last one given (like any other non-multiple parameter). Multiple pdb files would then need to be concatenated externally. It would be an easy change. Ralf
participants (5)
-
Francis E Reyes -
Pavel Afonine -
Ralf W. Grosse-Kunstleve -
Schubert, Carsten [PRDUS] -
Scott Classen