- phenixbb - phenix-online.org

Re: [phenixbb] phenix.refine questions: weights, docs, etc.
by Ralf W. Grosse-Kunstleve 29 May '07

29 May '07

Hi Dale, > The sets of parameters that can be refined simultaneously is a result of the > choice of optimization method, not the choice of the function being optimized. > The crucial difference is how the second derivatives are handled. In the > methods used in Phenix and CNS the seconds derivatives of all parameters are > assumed to be equal and uncorrelated. This is true only for the first iteration of a minimization in a phenix macro cycle. After that the L-BFGS method builds up an approximation of the second derivative information. > To ensure that this assumption is true > only parameters of the same category can be varied in a single cycle. This > means that the coordinates can be varied, but the B factors and scale factors > have to be held fixed. When the B factors are varied the coordinates and > scale factors must be constant. When the scale factors are varied everything > else must be fixed. > > Other refinement programs use second derivatives in a more explicit way > than Phenix and do vary more types of parameters in a single cycle. Shelxd is > the most powerful and, by default, varies all parameters of all classes each > cycle. Refmac, to the best of my knowledge, refines both coordinates and ADPs > together, but does refine TLS parameters in a separate step. As far as I know, Shelx doesn't use second derivatives at all. Shelx is shaped completely by the least-squares approach, using the Jacobian. You couldn't plug another target function into Shelx as you can do in the common macromolecular programs. I.e. there is in fact a link between the minimization method and choice of the target function. REFMAC uses a block-diagonal approach that only takes the correlations of the parameters of one atom into account. phenix approximates these with the L-BFGS method. We have not actually tried to refine coordinates and B-factors simultaneously since we've been too busy with other priorities. In theory it should be possible, but I'm sure it will take a lot of fine-tuning. > Since all the parameters of our models are correlated with one another, it > is better to refine as many of them at once as possible. Implementing > the join refinement of all these kinds of parameters is difficult, so to save > programmers' time approximations are sometimes made. We'll spend the time when we find it. :-) But I don't expect spectacular gains, no matter how long we beat on the second derivatives. Ralf

2 1

phenix.refine questions: weights, docs, etc.
by Frank von Delft 29 May '07

29 May '07

Hi, four questions on phenix.refine: 1) What's the reason for refining bss, xyz and adp in separate cycles? CNS does that too while refmac (I believe) refines them simultaneously. I suspect I missed something, but I thought the power of ML refinement is that no refinable parameters are held constant if they're not supposed to be. (At least, wasn't that the big deal in Sharp when it came out?) 2) How does one figure the right restraint weights? (wxc and wxc). In refmac, I've always calibrated the weight with rmsd(bonds), tightening it until rmsd(bonds) is ~0.018 (or below for lower reso); invariably this is where Rfree will decrease furthest and smoothly as well (i.e. without that silly rise at the end.) With phenix.refine I'm finding that I only get Rfree to drop monotonously when I tighten the weights (both wxc and wxu) so much that rmsd(bonds) goes down to ~0.002. Would there be a different way to do this? 3) There isn't currently torsional refinement, is there? For high-temperature simulated annealing, I mean. I'm having the same weight-adjustment question for SA, except is harder to fiddle with because jobs take so long, but with torsional that becomes moot. Or should one always put very tight weights on for SA? 4) Is there good documentation for all the loads and loads of keywords? The names themselves are just descriptive enough to be interesting, but obviously not enough to be informative if I don't just want the black box. Nice program, though, got to say... phx.

3 2

Death by Big-ass structure
by Frank von Delft 25 May '07

25 May '07

Hi, we get the memory error below for a structure of 30603 atoms 273217 reflections 2.1A H3 193.868 193.868 352.223 90.00 90.00 120.00 10x NCS Anything I said? It's not THAT large a problem, I'd have thought, or should we go to the rhombohedral setting instead? (Hmmmm, this rings a bell, wasn't there a posting about this 3 days ago? Nothing in the archives, was it CCP4bb?) Cheers phx Traceback (most recent call last): File "/usr/local/phenix/phenix-1.26b/phenix/phenix/command_line/refine.py", line 5, in ? command_line.run(command_name="phenix.refine", args=sys.argv[1:]) File "/usr/local/phenix/phenix-1.26b/phenix/phenix/refinement/command_line.py", line 75, in run refine_object.run( File "/usr/local/phenix/phenix-1.26b/phenix/phenix/refinement/driver.py", line 1236, in run log = log) File "/usr/local/phenix/phenix-1.26b/phenix/phenix/refinement/strategies.py", line 273, in refinement_machine log = log) File "/usr/local/phenix/phenix-1.26b/phenix/phenix/refinement/weight_xray_chem.py ", line 111, in __init__ self.fmodel_copy = self.fmodel.deep_copy() File "/usr/local/phenix/phenix-1.26b/mmtbx/mmtbx/twinning/twin_f_model.py", line 1003, in deep_copy map_types = self.map_types File "/usr/local/phenix/phenix-1.26b/mmtbx/mmtbx/twinning/twin_f_model.py", line 888, in __init__ self.update_f_mask() File "/usr/local/phenix/phenix-1.26b/mmtbx/mmtbx/twinning/twin_f_model.py", line 1370, in update_f_mask self.f_mask_array = mask.structure_factors( self.miller_set ) File "/usr/local/phenix/phenix-1.26b/mmtbx/mmtbx/masks.py", line 103, in structure_factors flex.grid(fft_manager.m_real()).set_focus(fft_manager.n_real())) MemoryError

4 6

Re: [phenixbb] Question about restraints in individualB-factor re finement
by Schubert, Carsten [PRDUS] 25 May '07

25 May '07

Thanks Pavel,Nigel: so it looks as if neither approach works and I would need to split the ligands into seperate PDB files for a truly independent B-factor refinement. Thanks for the hint about the non-bonded interactions. That was useful. Cheers Carsten -----Original Message----- From: phenixbb-bounces(a)phenix-online.org [mailto:[email protected]]On Behalf Of Pavel Afonine Sent: Friday, May 25, 2007 12:51 PM To: PHENIX user mailing list Subject: Re: [phenixbb] Question about restraints in individualB-factor refinement Hi Carsten, for the refinement of b-factors there is no difference between two options. We use the sphere b-factors restraints which are calculated based on interatomic distances and the information about alternative conformations is not used at that point. However if you refine the coordinates, the difference is significant: if you refine it as two separate residues then they will "see" each other for non-bonded interactions term calculation and hence will be pushed apart. Please let me know if you still have questions! Pavel. Schubert, Carsten [PRDUS] wrote: Hi, I have a question about the handling of restraints in the individual B-factor refinement routine. What I'd like to do is to refine a ligand, which can be present in 2 or more different conformations/orientations in its binding site. I'd like to use B-factor refinement on the various instances of the ligand, which one is the most relevant one, assuming that the most relevant conformation/orientation is associated with the lowest B-factor (Validity of that assumption set aside ...) My question is now is there any difference in the restraints applied to the b-factors in the scenarios where A) the ligand is modeled as alternate conformations i.e. ATOM 2724 C01AINH I 1 27.808 26.376 23.301 0.50 27.77 I C ATOM 2733 C01BINH I 1 30.898 22.496 17.340 0.50 22.15 I C ... vs. scenario B) where the same ligand is modeled as 2 different residues ATOM 2724 C01 INH I 1 27.808 26.376 23.301 0.50 27.77 I C ATOM 2733 C01 INH I 2 30.898 22.496 17.340 0.50 22.15 I C .... Basically, what I am trying to achieve is to uncouple the B-factor refinement of each individual instance of the ligand from its other instance(s). Are there any hidden pitfalls between these 2 scenarios I should be aware of? Many thanks for any input. Cheers Carsten _______________________________________________ phenixbb mailing list phenixbb(a)phenix-online.org http://www.phenix-online.org/mailman/listinfo/phenixbb

1 0

Question about restraints in individual B-factor refinement
by Schubert, Carsten [PRDUS] 25 May '07

25 May '07

Hi, I have a question about the handling of restraints in the individual B-factor refinement routine. What I'd like to do is to refine a ligand, which can be present in 2 or more different conformations/orientations in its binding site. I'd like to use B-factor refinement on the various instances of the ligand, which one is the most relevant one, assuming that the most relevant conformation/orientation is associated with the lowest B-factor (Validity of that assumption set aside ...) My question is now is there any difference in the restraints applied to the b-factors in the scenarios where A) the ligand is modeled as alternate conformations i.e. ATOM 2724 C01AINH I 1 27.808 26.376 23.301 0.50 27.77 I C ATOM 2733 C01BINH I 1 30.898 22.496 17.340 0.50 22.15 I C ... vs. scenario B) where the same ligand is modeled as 2 different residues ATOM 2724 C01 INH I 1 27.808 26.376 23.301 0.50 27.77 I C ATOM 2733 C01 INH I 2 30.898 22.496 17.340 0.50 22.15 I C .... Basically, what I am trying to achieve is to uncouple the B-factor refinement of each individual instance of the ligand from its other instance(s). Are there any hidden pitfalls between these 2 scenarios I should be aware of? Many thanks for any input. Cheers Carsten

2 1

probable RTFM question
by William Scott 17 May '07

17 May '07

Howdie: I just used phenix.geometry_minimization to try to see the effect of deleting some disulphide bonds in a protein structure. Is there a way to do the equivalent of CNS's model_anneal (i.e., simulated annealing with just a model)? It wasn't apparent from reading the phenix.refine docs. Thanks. Bill William G. Scott Contact info: http://chemistry.ucsc.edu/~wgscott/

3 2

Re: [phenixbb] statistics output by phenix.refine
by Ralf W. Grosse-Kunstleve 17 May '07

17 May '07

> Are these maximum likelihood targets the normal maximum likelihood > function or the minus one? We are using the maximum likelihood targets described here: Lunin, V.Y., Afonine, P.V. & Urzhumtsev, A. (2002). Acta Cryst. A58, 270-282. Lunin, V.Y. & Skovoroda, T.P. (1995). Acta Cryst. A51, 880-887. > Should I expect an increase or decrease of the target? Decrease. But see below... > If the part of the restraints is not combined with > the x-ray part, the target is showing is not the actual target the > program is trying to maximize or minimize. Yes, that's right. > The values don't have a trend, at least I can't see a trend in my > log file. The x-ray target may go up since we refine x-ray target + geometry target. The total will always go down (that's how the minimizer works), but an individual contribution may go up a bit. Ralf

1 0

statistics output by phenix.refine
by Jianghai Zhu 16 May '07

16 May '07

Hi, At the end of the log file output by phenix.refine, only R and Rfree are output. Is it possible to generate a statistical table which includes R, Rfree, FOM, LLG (total), LLG (free), rmsd (bond), rmsd (angle), ... for every refinement cycle? I want to monitor the progress of my refinement. Thanks. Jianghai +++++++++++++++++++++++++++++++ Jianghai Zhu, Ph.D CBR Institute for Biomedical Research Department of Pathology Harvard Medical School 200 Longwood Ave., Boston, MA 02115 Ph: 617-278-3211 Fx: 618-278-3232 +++++++++++++++++++++++++++++++

2 3

Using Phenix to calculate Atom Maps
by James Fraser 14 May '07

14 May '07

Hello, Is there a way to use Phenix to calculate maps that are just representative of the model? This would be similar to the ATMMAP mode of SFALL or using model_fcalc.inp/model_phase.inp in CNS. Thanks, James

3 3

Reflection output
by Rife, Christopher L 09 May '07

09 May '07

Hi, In trying to get a reflection file written out after refinement with phenix.refine (version 2007_03_13_2155), we found a couple of problems. 1. The documentation gives the example of including "export_final_f_model=mtz" in the command line input, however, phenix complains that it does not know this command. It is possible to work around this by changing "write_refined_fmodel_to_file" in the param file to true. 2. When the model is refined with twinning, trying to write out the reflection file causes it to crash with the following error: > Writing refined fmodel to plain text file: > /data5/Completed/PH10062D-47717/Phenix/model_001_fmodel.hkl > Traceback (most recent call last): > File > "/home/sdcsoftware/linux/cci_apps_sources/phenix/phenix/comman > d_line/refine.py", line 5, in <module> > command_line.run(command_name="phenix.refine", args=sys.argv[1:]) > File > "/home/sdcsoftware/linux/cci_apps_sources/phenix/phenix/refine > ment/command_line.py", line 86, in run > if(output.write_refined_fmodel_to_file): > refine_object.write_refined_fmodel(f_hkl) > File > "/home/sdcsoftware/linux/cci_apps_sources/phenix/phenix/refine > ment/driver.py", line 927, in write_refined_fmodel > self.fmodel.dump(out = f) > AttributeError: 'twin_model_manager' object has no attribute 'dump' We are trying to prepare files for deposition, so any assistance would be welcome. Thanks, Chris ----------------------------------------- Structure Determination Core Joint Center for Structural Genomics Stanford Synchrotron Radiation Laboratory

2 1