- phenixbb - phenix-online.org

R factor
by zhangh1＠umbc.edu 19 Sep '10

19 Sep '10

Hi, I want to calculate Fc based on the PDB file and reproduce the R factor in it. I first tried phenix.model_vs_data. It optimized ksol bsol and b_cart, and the output R factor (0.2143) is close enough to that in PDB (0.215). However, it didn't generate mtz files containing the calculated Fc. Then, I take the optimized ksol bsol and b_cart as the input and tried phenix.fmodel. I wrote my own code to calcuate the R factor. However, I got a lower R factor (0.2078). I think the reason might be my R calculation formula (sigma(fo-fc)/sigma(fo)) is wrong; or model_vs_data did more than just did flat solvent correction and anisotropic scaling. So, I am wondering whether phenix can: (1) Do the similar thing as model_vs_data (hope can be somehow faster:), but generate Fc-containing mtz file; (2) Do a simple R factor calculate to see whether my code is wrong (maybe answered already:) Thanks! Best Regards, Hailiang

2 7

generate Rfree flag for twinned crystals with 3 twin fractions
by Zheng Zhou 19 Sep '10

19 Sep '10

Dear all After reading the informative posts in mailing list, I started to deal with my first twinning dataset. The space group is P3. Xtriage found 3 twin laws. R vs R statistics indicated perfect twinning for two twin laws and one is twin with RPS. I realize that twin-related reflections should be in the same work or free group. Could phenix handle the Rfree flag correctly for me when there are more than one twin fraction identified? By applying one twin law, the Rfree dropped from 37% to 28% using phenix. When I use Refmac, the twin fractions are refined and the Rfree is down to 26%. I copied the Rfree flag from phenix to the ccp4 in this case. When I let ccp4 generate Rfree flag automatically, the Rfree is refined to 26% too. Should I label the Rfree at a higher space group? Thanks for your advice Best, Zheng (Joe) Zhou

2 1

Resolve NCS masks
by Florian Schmitzberger 18 Sep '10

18 Sep '10

Dear All, I would have a question regarding the use of masks in Resolve in Phenix for density modification: Is it possible to output the NCS and/or solvent masks Resolve is using during the Autosol run (similar as the options from the stand alone Resolve version) such as: mask_as_mtz ncs_mask_file mask.mtz protein_mask_file What are the best criteria (other than the final maps of course) within the output from Phenix.resolve to judge if the NCS (and/or solvent) mask was chosen adequately? (NCS_overlap?) I am mostly concerned if the masks chosen do include some of the more peripheral parts of my protein molecule and do not flatten these regions. Secondly, it is possible within Phenix Autosol wizard to supply a pdb file (with dummy atoms) for the mask for Resolve to construct the NCS mask? i.e. similar to set mask_pdb file= dummyatom.pdb Regards, Florian ----------------------------------------------------------- Florian Schmitzberger Biological Chemistry and Molecular Pharmacology Harvard Medical School 250 Longwood Avenue, SGM 130 Boston, MA 02115, US Tel: 001 617 432 5602

2 1

Ligand restraints editing with REEL
by Joseph Brock 18 Sep '10

18 Sep '10

Hi all, I'm working on a 1.5A structure that contains a novel ligand within the active site, which is an Arsenic derivative of Glutathione, using the Phenix GUI, version 1.6.4-486. I have been using the 'Molinspiration' webpage (http://www.molinspiration.com/docu/webme/) to generate a smiles string, which I then feed into ELBOW. Although the structure/chemical restraints of Glutathione should be unchanged by the covalent linkage to arsenic through the cysteinyl sulfur, the resulting restraints file consistently restricts one of the dihedrals to a value that is incompatible with the (quite clear) electron density. This occurs regardless of wether i use AM1 optimization, and is unresponsive to using native Glutathione (which refines into the corresponding density extremely well) as an input geometry file with any of the available 'Use geometry for' options (final geometry, initial geometry, etc). I then tried opening the output file in REEL and manually changing the corresponding dihedrals to those of glutathione, however after running simple optimization, the resulting restraints file has the same problem, and upon choosing the 'transer geometry and restraints' option (or any option) midway through running the 'eLBOW optimization' or 'AM1 optimization' options, Phenix consistently crashes. Any advice on how to transfer the origional geometry of native Glutathione to my new ligand effectively would be greatly appreciated. I'm a little confused as to the abilities of REEL, and how to affect them. Thanks so much in advance. Cheers, -Joe.

2 1

autobuild : MSE's keep coming back
by Lepore, Bryan 17 Sep '10

17 Sep '10

i used "semet=False" the .pdb has MET the .seq has M the output models, however, keep getting MSE's in them, complete with selenium. am i missing something? -bryan

5 6

Re: [phenixbb] Something that came up on the CCP4 list...
by Simon Kolstoe 16 Sep '10

16 Sep '10

Just for the record I've had it make a big difference a number of times as well with both refmac & shelx. I was under the impression that phenix.refine was "cleverly" allowing me to use that trick by just starring the add riding hydrogens in the def file however it seems not - something worth knowing! Simon On 16 Sep 2010, at 20:01, Phil Jeffrey wrote: > I've had it make a big difference in certain circumstances. And in > other circumstances not so much. Definitely worth a try. I do not > use the technique routinely. > > Phil Jeffrey > Princeton > > > > On 9/16/10 11:49 AM, Simon Kolstoe wrote: >> So just to clarify further - if i am having problems with the final >> geometry of a model (eg clashscore as reported by molprobity), is >> it an >> appropriate trick to add hydrogens and then refine in phenix.refine >> using the riding hydrogens option, or does this not really change >> anything? >> >> Simon >> >> On 16 Sep 2010, at 14:26, Pavel Afonine wrote: >> >>> Hi Simon, >>> >>>> Could I just seek some clarification? I was under the impression >>>> that >>>> phenix.refine calculated riding hydrogens by default and took these >>>> into account as a restraint on the geometry? >>> >>> if you want hydrogen atoms to be used in refinement (as "riding >>> model" >>> or explicit) your need to make sure the PDB file that you give to >>> phenix.refine contains H atoms. That is you need to add them using >>> programs like phenix.reduce or phenix.ready_set or using >>> corresponding >>> functionality in the GUI. >>> >>> phenix.refine will not add them internally as part of refinement IF >>> hydrogen atoms are NOT present in input PDB file. >>> >>> Adding hydrogen atoms as 'riding model' does not add any refinable >>> parameters, but the scattering contribution from them is included >>> most >>> of the time. This is why we do not use H atoms silently but require >>> them to be in input PDB file - simply to make sure the reported >>> statistics (such as R-factors) is easily reproducible. >>> >>> All the best! >>> Pavel. >>> _______________________________________________ >>> phenixbb mailing list >>> phenixbb(a)phenix-online.org >>> http://phenix-online.org/mailman/listinfo/phenixbb >> >> _______________________________________________ >> phenixbb mailing list >> phenixbb(a)phenix-online.org >> http://phenix-online.org/mailman/listinfo/phenixbb >

2 1

LABIN labels for Autosol/resolve NCS scores
by Florian Schmitzberger 16 Sep '10

16 Sep '10

Dear All, I would have a number of Phenix related questions: Firstly, I used the anisotropy server ( to anisotropically truncate my mtz file with anomalous signal (SAD data). My question is if the labels (for F+/F- and SIGF+/SIGF- are correctly recognized by Phenix (Autosol). The columns labels are: F_ref for FP; QSIGF_ref for SIGFP; GF(+)pk for F(+); LSIGF(+)pk for SIGF(+), GF(-)pk for F(-); LSIGF(-)pk for SIGF(-). It seems QSIGF and GF(+)pk/LSIGF(+)pk etc. are , but I am uncertain if it recognizes F_ref, because for the LABIN INPUT LABELS it says "NONE" for"F". Phenix does not seem to detect the anistropy ("Not using aniso- corrected data files as the range of aniso b is only 19.52 and 'correct_aniso' is not set); so I suppose I should use the correct_aniso=True keyword. Secondly, what are good values for the evaluation of the NCS mask and NCS averaging Resolve is using in Phenix; NCS overlap values (e.g. in my case 1.1) and NCS_CORR (in my case 0.52) in the "resolve.scores" file. Does an NCS overlap value of greater than 1 make sense? I am working with a (anisotropic) dataset with 4.1/4.5 A resolution, 78 % solvent content, 2-fold NCS; Perhaps a naive question: It is not possible to use B-factor sharpened maps for autobuilding with Phenix Autobuild is it? Thank you for any comments! Regards, Florian ----------------------------------------------------------- Florian Schmitzberger Biological Chemistry and Molecular Pharmacology Harvard Medical School 250 Longwood Avenue, SGM 130 Boston, MA 02115, US Tel: 001 617 432 5602

2 3

f', f'' of Bromide?
by Tri Ngo 16 Sep '10

16 Sep '10

Dear Phenix-BB members, I am trying to solve the MAD data of one bacteria protein. The crystal was soaked with 1M Sodium Bromide. The data were collected at three different wavelength. I intend to use Autosol to solve this structure, but I am not sure which parameter I should put in the field: f' and f''? I think there is a table in the synchrotron but maybe I didn't notice about that. So now what should I do? Thank you very much for your input! Best wishes, -- * TriNgo * Structural Biology Lab School of Medicine - Sungkyunkwan University Phone: 031-299-6150

4 3

Intl Conf on Structural Genomics, Toronto, Canada, May 10-14, 2011
by Thomas C. Terwilliger 16 Sep '10

16 Sep '10

Dear Colleagues, We are pleased to let you know about the upcoming International Conference on Structural Genomics 2011, to be held in Toronto, Canada on May 10-14, 2011. This meeting is the 6th in this series of biennial meetings of the International Structural Genomics Organization. The meeting is designed to serve as a forum to discuss the most recent developments in structural genomics and structural/chemical biology, and their impact on research in biology, medicine and disease. A limited number of student travel fellowships will be available for travel to this meeting. A substantial number of short talks will be selected from poster abstracts. Please see http://www.icsg2011.org for further details. We hope to see you there! Best regards, Cheryl Arrowsmith ICSG 2011 Organizer Ted Baker Stephen Burley Dino Moras Joel Sussman Shigeyuki Yokoyama Tom Terwilliger ISGO Executive Committee

1 0

64-bit Snow Leopard support
by Nathaniel Echols 16 Sep '10

16 Sep '10

Mac users-- The latest nightly version of Phenix (dev-532) is now capable of compiling 64-bit objects on Snow Leopard, which should help those of you working on very large and/or high-resolution structures. You will need to use the source installer for this; it will automatically recognize the machine type as "mac-intel-osx-x86_64" and use the correct sources. There are several things to be aware of before trying this: 1. I had to switch to the development version of wxPython to get the GUI working, and it's both inherently less stable and based on fundamentally different OS libraries, which may cause problems. I've fixed some of the most obvious issues and everything appears to run now, but I expect to find more glitches. (This shouldn't affect command-line users.) 2. We haven't extensively tested 64-bit Phenix on Mac; there may be undiagnosed low-level bugs, even for command-line-only use. This should change soon. 3. For the above reasons, and because source installation is generally difficult to support, we recommend that you continue to use the 32-bit binary installer if you are not experiencing memory problems. (It will figure out what you are trying to do and correct the machine type accordingly.) 4. One side effect of this change is that an installation made on Leopard will not work on Snow Leopard, and vice-versa. I may be able to fix this. Hopefully there will be a binary installer in the near future, if I can ever figure out how to fix NFS on my workstation. -Nat

3 3