'Tasks' tab missing from 1.3b GUI and HL coefficients from a model
Hi all, I recently (well, yesterday) installed the latest beta of phenix and everything is working great from the command line (which is how I use phenix 99.9% of the time). However, I fired up the GUI and found that the 'Tasks' tab appears to be missing from the 'Strategy Interface'. Is this a bug, a manifestation of an installation problem, or a feature? And... I was trying to use the GUI for prime-and-switch phasing to improve the maps for a partly-refined MR solution. What is the recommended path for creating HL coefficients from a model (from simulated annealing in phenix.refine) so that I can use the GUI's resolve 'prime-and-switch' strategy? Or is there an even better way of removing model bias and density-improving a big, low-res, MR-solved structure with 4-fold NCS in phenix? Thanks, Stephen -- Dr Stephen Graham Nuffield Medical Fellow Division of Structural Biology Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Phone: +44 1865 287 549
I think that the missing 'task' tab is a feature. Look under 'projects' -> 'preferences' to have the task window displayed. One way of improving maps is by using the autobuild wizard. If you run the gui, you might (or not) want to specify that you want to use 'rebuild in place' (builds 'around' your current chains) and that you would like to use the 'full omit' procedure to reduce model bias. This last option might take some time though! The wizard might be smart enough to pick up on the 4 fold NCS by the presence of the model and use that during density modification, although I am not sure about that. I think it will use the NCS information during refinement. HTH Peter Stephen Graham wrote:
Hi all,
I recently (well, yesterday) installed the latest beta of phenix and everything is working great from the command line (which is how I use phenix 99.9% of the time).
However, I fired up the GUI and found that the 'Tasks' tab appears to be missing from the 'Strategy Interface'. Is this a bug, a manifestation of an installation problem, or a feature?
And... I was trying to use the GUI for prime-and-switch phasing to improve the maps for a partly-refined MR solution. What is the recommended path for creating HL coefficients from a model (from simulated annealing in phenix.refine) so that I can use the GUI's resolve 'prime-and-switch' strategy? Or is there an even better way of removing model bias and density-improving a big, low-res, MR-solved structure with 4-fold NCS in phenix?
Thanks,
Stephen
Hi Peter,
I think that the missing 'task' tab is a feature. Look under 'projects' -> 'preferences' to have the task window displayed.
Thanks for that. Yes, now I can see my 'Tasks' tab (perhaps the tutorial at http://www.phenix-online.org/download/documentation/phenix/prerelease/introd... could be updated to reflect the fact that 'Tasks' are hidden by default?). Unfortunately, however, I was still unable to work out how I can use phenix to generate phases and FOMs from a PDB model (from simulated annealing in phenix.refine) for use in the prime-and-switch strategy. The map_coeffs.mtz from phenix.refine does not seem to have all of the requisite columns... Is there a 'phenix' way to do this, or should I just use SFALL? Also, while we are talking about map_coeffs.mtz, is there a command-line flag to turn off the automatic map generation by phenix.refine? I use coot to generate my maps from map_coeffs.mtz on the fly, so I would prefer not to have phenix.refine make maps for me. And one last question, just for fun. Can phenix.refine handle 'hard' (constrained, strict) NCS? I am working at low res with 4 molecules per ASU and would prefer to just have one chain modelled plus an NCS matrix to generate the other chains. However, all the documentation I have read leads me to believe that only 'soft' (restrained) NCS is possible at the moment. Is this the case? Thanks again, Stephen -- Dr Stephen Graham Nuffield Medical Fellow Division of Structural Biology Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Phone: +44 1865 287 549
Stephen Graham wrote:
Also, while we are talking about map_coeffs.mtz, is there a command-line flag to turn off the automatic map generation by phenix.refine? I use coot to generate my maps from map_coeffs.mtz on the fly, so I would prefer not to have phenix.refine make maps for me.
Set write_maps=false to suppress maps output in phenix.refine. Look at output {...} scope of parameters for more options for output suppressing.
And one last question, just for fun. Can phenix.refine handle 'hard' (constrained, strict) NCS?
In phenix.refine we have NCS restraints only. To emulate NCS constraints you can play with weights for NCS term. Pavel.
In phenix.refine we have NCS restraints only. To emulate NCS constraints you can play with weights for NCS term.
There is a limit to what you can do with the weights. It is clearly worth trying different values, but if they are on a vastly different scale compared to the rest of the restraints the structure may get very distorted (because the minimizer only pays attention to the overwhelming restraints). We (will have to) add strict NCS in the future. Ralf
I think that the missing 'task' tab is a feature. Look under 'projects' -> 'preferences' to have the task window displayed.
Yes, a feature...
One way of improving maps is by using the autobuild wizard. If you run the gui, you might (or not) want to specify that you want to use 'rebuild in place' (builds 'around' your current chains) and that you would like to use the 'full omit' procedure to reduce model bias. This last option might take some time though!
If the plain run with 4-fold ncs does not work, you might try the full-omit procedure. I would expect the ncs to help a lot here.
The wizard might be smart enough to pick up on the 4 fold NCS by the presence of the model and use that during density modification, although I am not sure about that. I think it will use the NCS information during refinement.
The wizards should find the NCS from the model and apply it in both density modification and refinement, yes... -Tom T
HTH
Peter
Stephen Graham wrote:
Hi all,
I recently (well, yesterday) installed the latest beta of phenix and everything is working great from the command line (which is how I use phenix 99.9% of the time).
However, I fired up the GUI and found that the 'Tasks' tab appears to be missing from the 'Strategy Interface'. Is this a bug, a manifestation of an installation problem, or a feature?
And... I was trying to use the GUI for prime-and-switch phasing to improve the maps for a partly-refined MR solution. What is the recommended path for creating HL coefficients from a model (from simulated annealing in phenix.refine) so that I can use the GUI's resolve 'prime-and-switch' strategy? Or is there an even better way of removing model bias and density-improving a big, low-res, MR-solved structure with 4-fold NCS in phenix?
Thanks,
Stephen
_______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
participants (5)
-
Pavel Afonine -
Peter Zwart -
Ralf W. Grosse-Kunstleve -
Stephen Graham -
Thomas C. Terwilliger