- phenixbb - phenix-online.org

TLS + anisotropic ADP mixed refinement
by Scott Classen 03 Feb '09

03 Feb '09

I am trying to perform refinement on a model with three chains (two protein and one DNA... plus water, Cl, and a few other solventy things). The B factors for chain A are generally better than for the other two and I am thinking that the resolution (1.8Å) warrants attempting anisotropic individual adp refinement for at least chain A, and then using TLS for the other two. So my strategy is: *individual_sites *individual_adp *tls *occupancies adp.individual.isotropic = "not Chain A" adp.individual.anisotropic = "Chain A" adp.tls = "Chain B" adp.tls = "Chain D" Previously the model had only been refined with individual isotropic adp (no TLS). The starting model has an r_work=0.2079 and r_free=0.2427. With this new aniso/TLS strategy in the first macrocycle when it gets to the initial TLS refinement it completely blows up with r_work and r_free both going up to 60% have I specified my refinement with bad syntax? or is this even possible with phenix.refine? I can submit more of the log file if you're interested. Thanks, Scott ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Scott Classen, Ph.D. SIBYLS Beamline 12.3.1 http://bl1231.als.lbl.gov Advanced Light Source Lawrence Berkeley National Laboratory 1 Cyclotron Rd MS6R2100 Berkeley, CA 94720 C) 510.206.4418 O) 510.495.2697 Beamline) 510.495.2134 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

2 1

can phenix.refine calculate r_work and r_free without doing refinement?
by Scott Classen 02 Feb '09

02 Feb '09

Hi all, I was wondering if phenix can calculate r_work and r_free without refinement, sort of like the model_stats from CNS. Thanks, Scott ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Scott Classen, Ph.D. SIBYLS Beamline 12.3.1 http://bl1231.als.lbl.gov Advanced Light Source Lawrence Berkeley National Laboratory 1 Cyclotron Rd MS6R2100 Berkeley, CA 94720 C) 510.206.4418 O) 510.495.2697 Beamline) 510.495.2134 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

3 2

phenix version 1.4-3: problem with bulk solvent scaling
by Oliv Eidam 31 Jan '09

31 Jan '09

Hi, I am currently refining a good 1.6 Angstroem data set using Phenix. Thereby I noticed a significant difference of R-free (>3%) depending on whether I used phenix version 1.3-rc6 (installed on my Mac laptop) or phenix version 1.4-3 (installed on my linux work station). Rfree differs by more than 3% (!) although I am using the same starting model, the same reflection file and the same refinement protocol (shown at the end of this email with detailed R and R-free factors for the 3 steps). In my opinion, my phenix 1.4-3 installation has a problem in the "bulk solvent modeling and scaling" steps. Indeed, I find in the log file that the "overall anisotropic scale matrix" shows all values set to zero. Analogous the "maximum likelihood estimate for coordinate error". This is true in all steps where an "overall anisotropic scale matrix" is calculated. Here is one example during 1.4-3 refinement: ====================== bulk solvent modeling and scaling ====================== |--(resolution: 1.61 - 29.59 A; n_refl. = 99535)------------------------------| | | | r_work= 0.3419 r_free= 0.3358 ksol= 0.39 Bsol= 19.12 scale= 1.247 | | | | overall anisotropic scale matrix (Cartesian basis; B11,B22,B33,B12,B13,B23):| | (0.00,0.00,0.00,0.00,0.00,0.00); trace/3= 0.00 | | | | maximum likelihood estimate for coordinate error: -0.00 A | | x-ray target function (ml) for work reflections: 5.981207 | |-----------------------------------------------------------------------------| The same example during 1.3-rc6 refinement: ===================== bulk solvent correction and scaling ===================== |-macro cycle = 1-(resolution: 1.61 - 29.59 A; n_refl. = 99537)---------------| | | | r_work= 0.3153 r_free= 0.3064 ksol= 0.40 Bsol= 43.79 scale= 1.087 | | | | overall anisotropic scale matrix (Cartesian basis; B11,B22,B33,B12,B13,B23):| | (-4.01,-5.36,-8.20,-0.00,2.74,0.00); trace/3= -5.86 | | | | maximum likelihood estimate for coordinate error: 0.33 A | | x-ray target function (ml) for work reflections: 5.946583 | |-----------------------------------------------------------------------------| I am asking now: Did something go wrong during my Phenix installation? I already re-installed Phenix (no errors): same result. Do I have to change something in my input files? I should probably mention that "bulk_solvent_and_scale = True" in my input .def files. Any help appreciated. Many thanks, Oliv -------------------------------------------------------------------------------- Refinement protocol and development of R and Rfree during each refinement round: -------------------------------------------------------------------------------- 1)Rigid body refinement 2) Simulated annealing (5000K) 3) Ordered solvent (plus annealing @ 2000K) phenix version 1.3-rc6: 1) R=31.5, Rf=30.6 2) R=23.9, Rf=25.8 3) R=19.8, Rf=22.0 phenix version 1.4-3: 1) R=31.6, Rf=30.6 2) R=27.8, Rf=29.8 3) R=23.7, Rf=25.2 -- Oliv Eidam, Ph.D. Postdoctoral fellow University of California, San Francisco Dept. of Pharmaceutical Chemistry 1700 4th Street, Byers Hall North, Room 501 San Francisco, CA 94158 - Box 2550 Phone: 415-514-4253 Fax : 415-514-4260 Email: eidamo(a)blur.compbio.ucsf.edu

3 6

joint xn refinement question
by Walter Novak 30 Jan '09

30 Jan '09

Hi, I am new to the list, but I didn't see this issue in the list archive. I am doing a joint refinement using x-ray and neutron data (each set is ~2.0 Å). I added H positions using reduce and have been switching to Ds as the density suggests for exchangeable positions. My problem is that I cannot seem to get the H (or D) positions to refine into the neutron density. Perhaps I need an additional command, so I am attaching my refinement script. I can provide a few pictures from coot to demonstrate if helpful. Thanks! Wally phenix.refine prot_joint4_r8_001.pdb \ neutron_data.file_name=prot_best_mcH_D2O_2_data.mtz \ xray_data.file_name=prot_best_mcH_D2O_2_data.mtz \ output.prefix=prot_joint4_r9 \ strategy=rigid_body+individual_sites+individual_adp \ simulated_annealing=False \ input.xray_data.labels=F-obs input.neutron_data.labels=F-obs-neutron \ input.xray_data.r_free_flags.label=R-free-flags \ input.xray_data.r_free_flags.file_name=prot_best_mcH_D2O_2_data.mtz \ input.xray_data.r_free_flags.test_flag_value=1 \ input.neutron_data.r_free_flags.label=R-free-flags-neutron \ input.neutron_data.r_free_flags.file_name=prot_best_mcH_D2O_2_data.mtz \ input.neutron_data.r_free_flags.test_flag_value=1 \ main.number_of_macro_cycles=5 \ restraints_edits.params \ refinement.hydrogens.refine=individual \ find_and_add_hydrogens=True \ switch_to_isotropic_high_res_limit=2.5 \ optimize_wxc=true \ optimize_wxu=true \ --overwrite Walter R.P. Novak, Ph.D. Postdoctoral Fellow Rosenstiel Basic Medical Research Center Brandeis University 415 South St. MS 029 Waltham, MA 02454-9110 Phone: (781) 736-4944 Fax: (781) 736-2405

2 1

Site-wide defaults for phenix.refine parameters
by Stephen Graham 29 Jan '09

29 Jan '09

Hi all, Apologies if I've missed this in the documentation, but is there any way to set 'site-wide' defaults for, say, phenix.refine? That is, I'd like to modify the default parameters for phenix.refine so that occupancy refinement is turned off by default, the output.serial_format is different, geo files aren't created, always use F/SIGF not I/SIGI, etc. Cheers, Stephen P.S. I know I can change this by hand in my .def file and then propagate the .def file through subsequent rounds of refinement for a given project. It would just be nice to have my default personal preference parameters set (~/.phenix?) so I don't need to change them each time I start a new project... -- 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

4 5

occupancy refinement
by Nicole Bischoff 29 Jan '09

29 Jan '09

Dear all, in my electron density I can see negative difference density (fo-fc) for the carboxyl groups of some aspartates or glutamates. Assuming that this is a result of partial decarboxylation due to radiation damage, I wanted to refine the occupancies of these carboxyl groups. To this end, I set the occupancy of these atoms to 0.99 in Coot so that they would be picked up by occupancy refinement in phenix.refine (using version 1.4-3). Phenix however just put the value back to one. It looks like the occupancy refinement in phenix is meant for alternative conformations where the sum of the occupancies of the different conformations is 1. How can I make phenix refine partially decarboxylated acidic side chains? It does not seem appropriate to me to make alternative conformations, one with and one without the carboxyl group. That would look funny at my resolution (2.7 Angstrom) and might not fix the problem. Any hints are appreciated. Thank you, Nicole ---- Nicole Bischoff, M. Sc. PhD student in the Gruetter lab Department of Biochemistry University of Zurich Winterthurerstrasse 190 8057 Zurich Switzerland +41 44 635 5556 nbischoff(a)bioc.unizh.ch

3 3

TLS question
by Jeffrey Chao 27 Jan '09

27 Jan '09

hello all- i apologize if this question has previously been posted, but i was not able to track down an answer by google. i am refining a 2.75A structure that has three copies of the protein in the asu with phenix 1.4-3 and using NCS with each chain as an NCS group. after molecular replacement, i set all of the b-factors to the wilson b-factor and then performed four different refinements: grouped b-factor (1 group per residue) : Rwork 22.6 Rfree 27.0 individual b-factor : Rwork 21.4 Rfree 26.6 grouped b + tls ( : Rwork 21.9 Rfree 25.2 individual + tls : Rwork 21.2 Rfree 24.8 i defined the tls groups with the following: adp { tls = chain A tls = chain B tls = chain C } after grouped b-factor the atoms of each residue all have the same b- factor as i expected. after grouped b-factor+tls, however, each atom of a residue has a unique b-factor, which i did not expect. i realize that this confusion is likely caused by me not fully understanding how tls refinement is implemented. if someone could explain this to me, i will be able to sleep better. thank you for your time, jeff Jeffrey Chao, PhD 1300 Morris Park Ave ASB-Golding 601 Bronx NY 10461 jchao(a)aecom.yu.edu 718-430-8597

2 1

rmsd: phenix vs refmac
by Frank von Delft 25 Jan '09

25 Jan '09

Hi I thought that a while ago, I read somewhere that rmsd(bonds) had not been calculated the same way in phenix as in refmac. I believe it was in connection with automated weights, and that rmsd(bonds) always much lower in phenix. I understood too that this had been adjusted since to match refmac. But we're still seeing some much lower rmsd(bond) from phenix: in particular, when we take a refmac model and plug it straight into phenix, the rmsd for cycle zero is already much lower than what it was at the end of refmac. Could someone enlighten...? (Yup, latest version of phenix.) Thanks! phx

5 5

bond and angle rmsds: phenix.refine, Refmac, hydrogens, ...
by Pavel Afonine 25 Jan '09

25 Jan '09

For all who asked about observed differences in bond rmsd reported by phenix.refine and Refmac, H atoms included in statistics calculation and more... 1) First, I wanted to check bond rmsd reported by phenix.refine. I made up a PDB file containing just one residue (SER, no H atoms) and I wrote a crystal-clear and simple Python script that computes bond rmsd for this residue. Here are the numbers: Bond length rmsd: Script: 0.148 phenix.pdbtools: 0.148 phenix.refine: 0.148 Refmac_5.5.0070: 0.152 Both programs are using the Monomer Library. Since both of these very different ways of computing bond rmsd: this simple script and phenix.refine, give the identical values, I tend to think that the PHENIX implementation is actually correct. I have no idea why Refmac gives a different value and I think it's now up to Refmac developers (or any other enthusiast) to find it out or show that I'm wrong in my calculations. It well could also be me who's incorrectly using Refmac - today was my second time I tried it (the first was 3 or 4 years ago) -:) As for the bond angles rmsd, I didn't bother to include it into the script, but for the rest of tools in question, the numbers are: phenix.pdbtools: 16.628 phenix.refine: 16.628 Refmac_5.5.0070: 16.628 so all results seem to be in-sync. 2) Some people reported that the geometry statistics for the exact same structure with and without hydrogens seems to be reported different in phenix.refine. To find out, I did the same test as above, where I used the same one residue model with all hydrogens added. Unfortunately, I failed to make Refmac accept a PDB file containing H atoms so I gave up after 20 minutes of struggle. Here are the numbers: Bond length rmsd: Script: 0.155 phenix.pdbtools: 0.155 phenix.refine: 0.155 So, it seems like the hydrogens are included... I have to fix it. Well at least those who are doing refinement using neutron data or both X-ray and neutron data, were not concerned, since in that case I presume the H and D atoms should be included. Anyway, I'm fixing it right now, I'm making a corresponding regression test (to make sure the fix stays there forever) and the fixed version will be available in the next release. By the way, the new tool, phenix.model_vs_data is and was smart about this: it automatically excludes H and D atoms if X-ray data is used. And this proves it: phenix.model_vs_data: 0.148 which is the exact same number as in "1)", where H were not present. Thanks to all who raised this issue (Frank von Delft, Jianghai Zhu, Karolina Michalska and others)! Pavel. PS> All files used and the script are here: http://cci.lbl.gov/~afonine/rmsd_issue/ Use "phenix.python run.py" to run the script.

1 0

IAS causes phenix.refine to die
by Anna Gardberg 24 Jan '09

24 Jan '09

Dear phenix list, My 0.75 A refinement has some problems in its higher resolution bins, so I thought I might try adding in IAS scatterers. However, the phenix.refine (version 2008_09_13_0905) crashed when I set "ias = True". I hope someone here can help me figure out the problem. I include snippets of output below, as the entire error message was very lengthy. Am I misusing the command, is there a problem with my system setup, or what? Regards, Anna Gardberg ----------Building IAS model---------- >>> Adding IAS.......... Total number of covalent bonds = 895 show_stack(1): /home/a1g/downloads/cci_apps_sources/mmtbx/ias/__init__.py(438) _ _init__ show_stack(2): /home/a1g/downloads/cci_apps_sources/mmtbx/ias/__init__.py(371) s et_peaks show_stack(3): /home/a1g/downloads/cci_apps_sources/mmtbx/ias/__init__.py(616) _ _init__ show_stack(4): /home/a1g/downloads/cci_apps_sources/mmtbx/model.py(494) add_ias show_stack(5): /home/a1g/downloads/cci_apps_sources/phenix/phenix/refinement/str ategies.py(409) refinement_machine show_stack(6): /home/a1g/downloads/cci_apps_sources/phenix/phenix/refinement/dri ver.py(1082) run <snip pages of output> /home/a1g/downloads/cci_apps_build/lib/cctbx_maptbx_ext.so(boost::python::obje cts::make_holder<4>::apply<boost::python::objects::value_holder<cctbx::maptbx::o ne_gaussian_peak_approximation>, boost::mpl::vector4<scitbx::af::const_ref<doubl e, scitbx::af::trivial_accessor> const&, scitbx::af::const_ref<double, scitbx::a f::trivial_accessor> const&, bool const&, bool const&> >::execute(_object*, scit bx::af::const_ref<double, scitbx::af::trivial_accessor> const&, scitbx::af::cons t_ref<double, scitbx::af::trivial_accessor> const&, bool const&, bool const&)+0x 75c) [0x2a9d79598c] /home/a1g/downloads/cci_apps_build/lib/cctbx_maptbx_ext.so(cctbx::maptbx::find _gaussian_parameters::find_gaussian_parameters(scitbx::af::const_ref<double, sci tbx::af::trivial_accessor> const&, scitbx::af::const_ref<double, scitbx::af::tri vial_accessor> const&, double const&, double const&, double)+0x3fe) [0x2a9d7827f e] /lib64/tls/libm.so.6 [0x32ee506263] /lib64/tls/libc.so.6 [0x32ee22e300] Floating-point error (Python and libc call stacks above) [1]+ Done gedit newscale-ias.eff

3 2