- phenixbb - phenix-online.org

simulated annealing composite omit map or iterative build composite omit map at 2.7A resolution
by Xiao Lei 15 Jul '15

15 Jul '15

Dear Phenixbb members, I am thinking of generating a composite omit map for my whole structure model (the model is solved by phaser MR in CCP4 and refined by phenix.refine program, results of phenix.refine shows Rfree/Rwork is 27% / 20%. The resolution of data is 2.7A). From previous phenix posts and information from internet, I know there are four composite omit map types in Phenix: "simple"; "refine"; "simulated annealing" and "iterative build" composite omit map. Because I want to get rid of model bias as much as possible, I am thinking of choosing either "simulated annealing" or "iterative build" method, though I know it needs much longer time to run. I know "iterative build" composite map option is more powerful to get rid of model bias as I found in phenix manual that "A more powerful approach is to generate an "iterative-build" omit map (Terwilliger et al. 2008), which rebuilds the structure instead of annealing." But at the "autobuild-iterative guild omit map" manual it also said "The AutoBuild Wizard produces a refined model that can be nearly complete if the data are strong and the resolution is about 2.5 A or better. At lower resolutions (2.5 - 3 A) the model may be less complete and at resolutions > 3A the model may be quite incomplete and not well refined." My questions are: 1. As the resolution of my data is 2.7A, does this mean Simulated annealing composite map option is better for me in this case? 2. Inside phenix panel, there are two map tools for generating composite omit map, one is "Composite omit map" tool and the other is "Autobuild-iterative build omit map" tool. I tried the "composite omit map" tool with simulated annealing option with my refined model and original x-ray data, in the output file I only get 2mFo-DFc map (I know this is by default) but no pdb output, I am guessing this is because in this "composite omit map" tool my model is not modified (or refined) by the program, just the composite omit map is generated, am I right? If I am right, does this mean the following step is to open the output map in coot with my model, and do manual adjustment of the model in the regions which does not fit the map? 3. In the "autobuild-iterative build omit map" tool, there is also an option of "simulated annealing" composite omit map, is this option the same as the "simulated annealing" option in the "composite omit map" tool and will give the same output (just the 2mFo-DFc map by default, not modifying my model) ? 4. If I use iterative-build composite map with this 2.7A resolution, I'm guessing I will get a modified or refined pdb model output and a composite 2mFo-DFc map, but my model may not be complete because of the resolution, does this mean I have to manually add the missing part of model in the following step? 5. Is there an option to generate composite omit mFo-DFc map along with composite 2mFo-Fc map at the same time? Thanks ahead. Xiao

1 0

The difference of using data label "Imean and SigImean" or "F and SigF" in Phenix
by Xiao Lei 14 Jul '15

14 Jul '15

Dear All, I am using Phenix1.9-1962 version on Mac OS X 10.10.3. I notice in Phenix refinement or composite omit map GUI panel, there is an option for me to select data labels. I have two options of data labels: one is "Imean and SigImean"; the other is "F and SigF". My experimental data is from a single crystal diffraction set, after convert scale file (.sca) to mtz file using scalepack2mtz in ccp4, I have a mtz file output. I checked the mtz file using viewhkl program, it shows I have both Imean, SigImean, and F, SigF columns. Is there any difference between the data label choices in these phenix programs? Xiao

2 2

message after refinement
by Almudena Ponce Salvatierra 14 Jul '15

14 Jul '15

Dear all, after refinement of my nucleic acid structure with phenix.refine I get: "warning: rms deviation from ideal geometry (bonds and/or angles) is suspisciously high. You may need to optimize X-ray stereochemistry weighting" However when I move to the molprobity tab I see no outliers no anything... looking weird. I even more wonder where this comes from because the fact is that I had one refinement with no warnings or anything and today I repeated the same, exactly the same, except for the fact that I renumbered one of the chains. Now all out of a sudden (I know it doesn't sound scientific at all), I get this warning regarding the rmsd when I was actually considering that the refinement stage was over. Does anybody have a theory regarding where the problem might be here? Thanks a lot in advance. Best wishes, Almudena -- Almudena Ponce-Salvatierra Macromolecular crystallography and Nucleic acid chemistry Max Planck Institute for Biophysical Chemistry Am Fassberg 11 37077 Göttingen Germany

2 1

Refinement with new mtz.file
by Aleksandar Bijelic 14 Jul '15

14 Jul '15

Dear CCP4 user and experts, I refined (with PHENIX) a 3.0 A dataset obtaining Rfree of about 0.24 (with good geometry according to Ramachandran, Beta Outliers, etc.) .... Everything seems to be ok (especially in relation to the resolution) .... because the mtz.file I used was quite old and I cannot find my xscale.hkl file, I processed the data set again (this time with optimizng and polishing) and received a "better" file according to almost everything (resolution limit, I/sigma, CC(1/2), Rmeas), thus I used this new mtz.file and put it in my last refinement step (the refinement which led to the above mentioned Rfree = 0.24). Suprisingly, the refinement starts at Rwork = 0.18 and Rfree = 0.19 but ending up with 0.20 and 0.22, respectively. So I wanted to know if this is usual? I was expecting my data to become slightlly better but what is irritating me is the starting R-values of the refinement and that it get worse during refinement. Maybe I did something wrong? Is it reasonable to replace the mtz.file with a new one in the last refinement step or should I start the refinement from the scratch? In the sake of completeness, I deleted the header of the used pdb.file because of the R-flag error which occurs, since Phenix reconizes that the pdb.file was already used with other Rflags. Sorry, but I am still a beginner in this field, so I would be very grateful if somebody could explain me this situation and my mistake and if I need to start refinement from the beginning. Thank you in advance! Best Regards. Aleksandar -- ------------------------------------------- Aleksandar Bijelic, MSc. Institut für Biophysikalische Chemie Universität Wien Althanstrasse 14 A-1090 Wien Tel: +43 1 4277 52536 e-Mail: aleksandar.bijelic(a)univie.ac.at --------------------------------------------

7 10

Set up Simulated Annealing starting tempreture
by Xiao Lei 13 Jul '15

13 Jul '15

Dear Phenixbb members, I'm learning phenix this days. I use phenix refinement and found there is a simulated annealing option (SA Cartesian or SA torsion option). However, I found the SA cartesian starts at 5000K and SA torsion starts at 2500K by default, and I'd like to change the starting temperature by myself. I search phenix manual and found: To change starting SA temperature: % phenix.refine data.hkl model.pdb simulated_annealing=true \ strategy=individual_sites simulated_annealing.start_temperature=10000 I do not understand this, does this mean I have to type the above command in somewhere in phenix? is there a place I can simply put SA start temperature numbers in? Xiao

2 2

NCS and R-Values Hung up at 0.42/0.46 for 2.1Å resolution
by John Bruning 11 Jul '15

11 Jul '15

I collected two datasets of the same ligand bound structure at 2.4 and 2.1 Å resolution. The apo structure has already been solved to 2.2Å resolution in P31 (and P3121 for SeMet). After processing each dataset as both P31 and P3121 in MOSFLM, I proceeded with MR with Phaser using sequential partial search models to place either 3 copies in the ASU for P3121, or 6 copies for P31. (The previously solved apo structure had 6 copies of the same protein in the ASU). *** in summary two datasets, processed as both P31 or P3121 After running xtriage for either .mtz, I get this warning about NCS: Distance to origin : 56.728 Height relative to origin : 26.388 % Translational pseudo-symmetry is very likely present in these data. Be aware that this will change the intensity statistics and may impact subsequent analyses, and in practice may lead to higher R-factors in refinement. Regardless, I can proceed with the MR using sequential partial search models to place either 3 copies (P3121) or 6 copies (P31) with decent solutions (i.e. packing looks good upon inspection, scoring looks appropriate -> SOLU SET RFZ=2.8 TFZ=8.3 PAK=13*LLG*=-37 *LLG*=2199 as an example). However, I did get the following somewhat expected warnings: Number of known components (1) not divisible by 2: Translational *NCS* correction not applied Large non-origin Patterson peak indicates that translational *NCS* is present Upon any attempt at initial refinement (with a number of different strategies), the R-factors are hung up > 0.42 (for example in the P3121, 2.1Å case): Start R-work = 0.4966, R-free = 0.5133; Final R-work = 0.4231, R-free = 0.4477 My concern is that there is some pathology that is not being recognized (merohedral twinning) or that the NCS/pseudosymmetry is not being addressed appropriately. -- Cheers, John

1 0

Re: [phenixbb] Fwd: [ccp4bb] Phenix; converting a map to structure factors
by Pavel Afonine 11 Jul '15

11 Jul '15

Hi Mark, Tom copied me your email. > As I'm sure you are aware there is tremendous interest in building > into and refining against low resolution (by X-ray standards) EM maps. > Could you comment to the phenix BB about whether its appropriate to > convert maps to reciprocal space and use ML refinement? While it is a possible way of doing this (and this is how it's been done in the past) I don't think it is a good route to follow. In case of cryo-EM your data is map, so why not just refine against the map? Why take a detour through reciprocal-space by converting map into structure factors (=manipulating your original data) and doing reciprocal space using refinement programs that were designed for crystallography? With this in mind we have developed a new refinement tool in Phenix (it is still in active development!) called phenix.real_space_refine that is designed to refine atomic models directly into maps (any map: X-ray, neutron, cryo-EM, low or high resolution), for quick summary see http://phenix-online.org/presentations/real_space_refine.pdf as well as documentation page (at the bottom): http://phenix-online.org/documentation/ and newsletter article (page 43): http://phenix-online.org/newsletter/CCN_2013_07.pdf Let me know if you have any questions or comments! Pavel

2 4

atom vdW dictionary for Python
by David Shin 11 Jul '15

11 Jul '15

Hi - I'm writing a Python script to measure the volume of pockets within protein structures - plus extra bells and whistles... thus a custom job. I was wondering just for speed, if someone could point me to values for van der Waals radii that are somewhat agreed upon for protein atoms, or if there is some dictionary already made that someone could point to that resides in phenix or some other structural biology package that would be ok for me to copy/use. Thanks, Dave

4 3

Re: [phenixbb] [ccp4bb] Phenix; converting a map to structure factors
by Terwilliger, Thomas Charles 10 Jul '15

10 Jul '15

Hi Reza, Just use: phenix.map_to_structure_factors map.ccp4 d_min=2.5 or something like that, All the best, Tom T On Jul 10, 2015, at 11:27 AM, Reza Khayat wrote: Hi, Is it possible to convert a map to structure factors using Phenix? Thanks. Best wishes, Reza Reza Khayat, PhD Assistant Professor City College of New York Department of Chemistry New York, NY 10031

3 4

NCS and R-Values Hung up at ~0.42/0.46
by John Bruning 10 Jul '15

10 Jul '15

Hi Everyone, I collected two datasets of the same ligand bound structure at 2.4 and 2.1 Å resolution. The apo structure has already been solved to 2.2Å resolution in P31 (and P3121 for SeMet). After processing each dataset as both P31 and P3121 in MOSFLM, I proceeded with MR with Phaser using sequential partial search models to place either 3 copies in the ASU for P3121, or 6 copies for P31. (The previously solved apo structure had 6 copies of the same protein in the ASU). *** in summary two datasets, processed as both P31 or P3121 After running xtraige for either .mtz, I get this warning about NCS: Largest Patterson peak with length larger than 15 Angstrom: Frac. coord. : 0.333 -0.333 -0.153 Distance to origin : 56.728 Height relative to origin : 26.388 % p_value(height) : 2.659e-03symmetry. Translational pseudo-symmetry is very likely present in these data. Be aware that this will change the intensity statistics and may impact subsequent analyses, and in practice may lead to higher R-factors in refinement. Regardless, I can proceed with the MR using sequential partial search models to place either 3 copies (P3121) or 6 copies (P31) with decent solutions (i.e. packing looks good upon inspection, scoring looks appropriate -> SOLU SET RFZ=2.8 TFZ=8.3 PAK=13*LLG*=-37 *LLG*=2199 as an example). However, I did get the following somewhat expected warnings: Number of known components (1) not divisible by 2: Translational *NCS* correction not applied Large non-origin Patterson peak indicates that translational *NCS* is present Upon any attempt at initial refinement (with a number of different strategies), the R-factors are hung up > 0.42 (for example in the P3121, 2.1Å case): Start R-work = 0.4966, R-free = 0.5133 Final R-work = 0.4231, R-free = 0.4477 *** best R-factors from ~ 15 different strategies My concern is that there is some pathology that is not being recognized (merohedral twinning) or that the NCS/pseudosymmetry is not being addressed appropriately. Any suggestions in troubleshooting this would be much appreciated, as I have pretty much exhausted everything I could think of. -- Cheers, John

2 1