- phenixbb - phenix-online.org

MR_ROSETTA
by Wei Shi 19 Nov '11

19 Nov '11

Hi all, I am trying to run phenix.mr_rosetta using the following script for rebuilding the model with Rosetta before MR from phenix website except omitting specifying ncs_copies number. Following is the script I am using: phenix.mr_rosetta \ seq_file = seq.dat \ data=coords1.mtz \ search_models=coords1.pdb \ run_prerefine=True \ number_of_prerefine_models=1000 \ fragment_files = test3.gz \ fragment_files = test9.gz \ rescore_mr.relax=False \ rosetta_models=20 \ space_group=p222 \ use_all_plausible_sg=False \ nproc=4 \ group_run_command=qsub\ When I run the above script, it gives the following error message: child process stderr output: command: 'qsub /home/wei/MR_ROSETTA_14/GROUP_OF_PLACE_MODEL_1/RUN_FILE_1' /bin/sh: qsub: not found Then, I omit the last line in the command “group_run_command=qsub\”. MR_ROSETTA seems to be running. But, it takes forever to run. My questions is: 1. Does anyone happen to know why it tells me “qsub: not found” when I include “group_run_command=qsub” in my script? 2. When I omit “group_run_command=qsub” in the script, mr_rosetta seems to be running. But it just gives me the following message and it takes forever to run. Almost 10h later, no job is finished. I don’t know whether this is normal. If not, any idea about why it’s so slow and any suggestions about how to speed up the run? Thank you so much! Splitting work into 1 jobs and running with 4 processors using sh background=True in /home/wei/MR_ROSETTA_15/GROUP_OF_PLACE_MODEL_1 Starting job 1...Log will be: /home/wei/MR_ROSETTA_15/GROUP_OF_PLACE_MODEL_1/RUN_FILE_1.log

7 9

error: Duplicate selections
by Sebastiano Pasqualato 17 Nov '11

17 Nov '11

Hi all, sorry guys, this might be a silly one, but I can't see where I'm making the mistake. Whenever I turn on TLS refinement, I get this error message. And, as you can see, the Log file doesn't tell me much more. I have generated the TLS groups wit the 'TLS groups selection' tool in phenix. Any hint? Thanks in advance, ciao s -- Sebastiano Pasqualato, PhD Crystallography Unit IFOM-IEO Campus Cogentech - Consortium for Genomic Technologies via Adamello, 16 20139 - Milano Italy tel +39 02 9437 5172 fax +39 02 9437 5990

2 2

Scattering from riding hydrogens?
by Damian Ekiert 16 Nov '11

16 Nov '11

Hello, I have a question regarding two parameters and how they affect riding hydrogens. contribute_to_f_calc = True high_resolution_limit_to_include_scattering_from_h = 1.6 At lower resolution, setting contribute_to_f_calc = false is frequently helpful when using riding hydrogens and has been discussed before. But what is high_resolution_limit_to_include_scattering_from_h = 1.6 doing? If scattering from hydrogens is only included at higher than 1.6A resolution, it would seem like contribute_to_f_calc = True/False would not have any effect at say, 3.5A, but comparing identical refinement runs with contribute_to_f_calc = True or contribute_to_f_calc = False clearly gives different results. Can hydrogens contribute to Fcalc in a way aside from their scattering, or is their contribution accounted for independent of the position of the riding hydrogen (e.g., additional scattering term added to the associated heavy atom)? I think I am misunderstanding what these two parameters are doing and would appreciate some clarification. Thanks, Damian Ekiert

4 5

Multiple TLS groups per molecule
by Yarrow Madrona 15 Nov '11

15 Nov '11

Hi, I would like run TLS and individual isotropic refinement. I want to separate one molecule into four TLS groups. I havn't found documentation for this. Is the number of TLS groups automatically done by phenix refine when you tell it to run TLS refinement on the whole molecule? Thanks for your help -Yarrow -- Yarrow Madrona Graduate Student Molecular Biology and Biochemistry Dept. University of California, Irvine Natural Sciences I, Rm 2403 Irvine, CA 92697

2 3

Small issue with residue PTR
by Terry Lang 15 Nov '11

15 Nov '11

Hey Everyone, I have found a small issue with atom naming for residue PTR. When phenix.ready_set adds hydrogens to the residue, it names the hydrogens connected to the CB atom to be HB3 and HB2. However, the hydrogens in the cif file located in chem_data/mon_lib are labeled HB1 and HB2. Manually changing either the cif file or the atom name in the pdb file solves the problem. Terry -- P. Therese Lang Post Doc Alber Lab, UC Berkeley

2 1

How can I add another chain to my PDB by Coot
by Dialing Pretty 15 Nov '11

15 Nov '11

Dear All, By Phenix refine I got a PDB (Only one chain) and mtz output files. After analysis by Coot I find I need to add a protein fragment so that a large blob will disappear. I know I use Baton to construct the protein fragment. Will you please tell me how can I input the sequence file of the protein fragment so that I can refine the protein fragment by Coot? I am looking forward to getting your reply. Regards. Dialing

2 1

twin/pseudo-symmetry question
by Green, Todd 12 Nov '11

12 Nov '11

Hello all- I have a question regarding the possibility of twinning and or psuedo-symmetry. I have a crystal that processes well in R32, ie. less than 6% r-merge at 2.2 angstrom. We collected multiple potential derivatives and were able to get some phasing from one that was good enough to model in the helices. The strange thing is that when looking at the harker sections there were a lot of unexplained strong peaks. And there is a very narrow range of parameters that we could get the sites with Shelx and end up with interpretable maps(i know that this part might not be unusual). My first thought was that there could be twinning. Since there is no twin law for this space group, I thought that it might suggest that there was twinning in a lower symmetry space group and/or that we have a pseudo-R32 crystal. Processing works well in C2 and R3, as expected. In those cases Xtriage suggests that the real symmetry is likely to be R32. In the case of C-2, xtriage suggetss if it is twinned then it is about 45% twinned. That's the background. Fast forward with the help of phenix Autobuild and some manual correction, we have been able to refine the structure with the R32 processed data(with tls) to r_work = 0.1894, r_free = 0.2287, bonds = 0.007, angles = 0.908 at 2.2 angstrom with almost everything accounted for. I think I should be quite comfortable with this. BUT, I'm still wondering about those Harker sections. I've been reading the manuscripts of Dauter regarding twinning and I think I am at this point more curious than anything. But not leaving well enough alone, I went back to the data processed in C2. I can easily do molecular replacement, following refinement at 2.3 angstrom(I need a lot more frames for this sg so the cutoffs are a bit lower but r-merge is ~ the same as R32) with TLS I get r_work = 0.1850, r_free = 0.2249, bonds = 0.007, angles = 0.946. this doesn't seem significant to me. However if I refine with the two possible twin operations, the values drop with one being better than the other. both have the same starting model: r_work = 0.1639 r_free = 0.1880 bonds = 0.006 angles = 0.902 r_work = 0.1582 r_free = 0.1866 bonds = 0.006 angles = 0.891 The twin fraction from the refined pdb header is 0.544 for the better statistical model and the r-values seem too low for 2.3 angstrom data. Is the high twin fraction suggestive of the higher symmetry space group or inconclusive towards that. of interest is that the maps look better to me than the r32 maps. by that i mean that the maps don't have what i would call minor dots of negative density for some(not a lot) of the side chains as we see for some side chains in the R32 case. also some side chains look better in the C2 twin refined case. Lastly for some of the monomers(there are now 6 copies versus 2 in the R32 case), I can add an additional residue or two on the termini which are not seen before. I wonder if this also suggests that the true space group is C2. Are these r-values too low? does the difference in values according to each twin law mean something? In the end, the models are essentially the same in either space group (with the exception of a couple of residues at the termini). So, I don't think I learn much one way or the other. But the curiocity keeps coming back. So is it likely to be R32 or C2 or C2 with a hefty twin fraction? Anyone have an explanation? Thanks in advance- Todd

3 7

peptide link problem solved LINKR in pdb is not recognized but LINK is
by Christian Roth 11 Nov '11

11 Nov '11

Hi, In my pdb from Refmac the keyword LINKR is used but only LINK is interpreted by phenix. I changed this and now it works as it should. Sorry, a careful analysis sometimes help. Regards Christian

1 0

peptide link
by Christian Roth 11 Nov '11

11 Nov '11

Hi, I have an modified aminoacid in my peptide chain (backbone modified). I have a cif file and I refined it previously in Refmac. I put the Link definition in the pdb and Refmac has interpreted the peptide link as I wanted. Now I did a run in phenix to compare the results and the chain is broken I created the SNN cif file with ReadySet Reel and I tried to incorporate the link definition via GUI in the following way apply_cif_link: data_link: TRANS residue_selection_1: chain A and resseq 366 residue_selection_2: chain A and resseq 367 apply_cif_link: data_link: TRANS residue_selection_1: chain A and resseq 367 residue_selection_2: chain A and resseq 368 Model: "" Number of chains: 12 Chain: "A" Number of atoms: 7983 Number of conformers: 2 Conformer: "A" Number of residues, atoms: 496, 7926 Unusual residues: {'SNN': 1} Classifications: {'undetermined': 1, 'peptide': 495} Modifications used: {'COO': 1, 'NH3': 1} Link IDs: {'PTRANS': 37, 'TRANS': 456, None: 2} Not linked: pdbres="ASP A 366 " pdbres="SNN A 367 " Not linked: pdbres="SNN A 367 " pdbres="GLY A 368 " Chain breaks: 1 Unresolved chain link dihedrals: 1 Unresolved non-hydrogen bonds: 1 Unresolved non-hydrogen angles: 1 This fails with an error message that it cannot interpret the links. In my previous pdbs I have also defined the atoms which I wanted to link. Maybe I have to do this also somhow in Phenix. I have tried to extract the Link definition using phenix.link_edits but the resulting .edits file is empty. How can I resolve issue? Thanks Christian

2 1

Rebetta fragment library
by Wei Shi 11 Nov '11

11 Nov '11

Hi everyone, I am trying to generate a fragment library in Rebetta server for my sequence of interest to use with mr_rosetta. I can open the 9-mer and 3-mer fragment results: aat000_09_05.200_v1_3 and aat000_03_05.200_v1_3, but didn't see link to download these two files. Does anyone know how to download these two files? If I do copy and paste, what file type or format is compatible with mr_rosetta? Thank you so much! Best, Wei

2 2