- phenixbb - phenix-online.org

Re: [phenixbb] Refining sucrose
by Yuri 01 Jun '11

01 Jun '11

All coot needs to run its real-space refinement is a .cif file as Nigel pointed out. You can also get this file from other ligand library websites (proDrg, hic-cup, etc...). The only catch here is. The name of your sucrose atoms must match, exactly, those in the file. Once you have such file go to the the file drop-down menu and click on Import CIF dictionary in coot. That should enable to rotate the rotatable bonds of your ligand -- Yuri Pompeu

5 5

heavy atoms
by ChenTiantian 01 Jun '11

01 Jun '11

Hi, I am dealing with a complex structure which the ligand contains a atom Iodine, I guess it's because Iodine is heavy atom, the electron density of Iodine turns red after 5 cycles refinement. Anybody coped with similar issue? How to let phenix know that the Iodine atom is part of the ligand? Best, Tiantian -- Drug Discovery and Design Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences Address: Room 101, 646 Songtao Road, Zhangjiang Hi-Tech Park, Pudong New Area, Shanghai, 201203, P.R. China

2 1

Phenix MTZ files,
by Saba Ferdous 154-FAS/BSBI/F05 01 Jun '11

01 Jun '11

Dear Phenix users I am new to Phenix..probably my question might be childish to you. so with appologies, I have predicted protein structure by using modeller, i want to refine my structure but i am unable to know that what is MTZ files n CIF files because i have only my PDB. And i want to fix beta deviations and outliers, bad angles and bad bonds in my structure. kindly help me and suggest me that what can i do to refine my structure. Best Regards -- Saba Ferdous Research Scholar (M. Phil) National Center for Bioinformatics Quaid-e-Azam University, Islamabad Pakistan

2 1

ligand cif file
by ChenTiantian 01 Jun '11

01 Jun '11

Hi there, Now I use phenix.elbow to produce the cif file of the ligand instead of using Prodrg. but I find a problem with this cif file, the C=N bond is not right, it becomes C-N in the cif file, so it does not fit the electron density, what should I do to modify the cif file to chage C-N to C=N ? Best Regards, Tiantian -- Drug Discovery and Design Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences Address: Room 101, 646 Songtao Road, Zhangjiang Hi-Tech Park, Pudong New Area, Shanghai, 201203, P.R. China

3 3

mr_rosetta test failure
by Laurie Betts 31 May '11

31 May '11

We installed latest phenix and latest rosetta last week. I ran test script: phenix_regression.wizards.test_command_line_rosetta_quick_tests and got the test failing at the prerefine because of missing shared libraries, as shown below in the FAQ (italics). I tried again in both bash & csh after setting or exporting the environment variables shown below, but got the same exact error. If I do printenv the shell sees the new env var. What should I try next? THanks so much Laurie Betts & CD Poon University of North Carolina *Why does mr_rosetta bomb and say "error while loading shared libraries: libdevel.so: cannot open shared object file: No such file or directory ? This may indicate that somewhere your system is defining the shared libraries that Rosetta needs, and these are for a place that is not where Rosetta expects them to be. You can try to ignore the previous definitions this way: If you are using the bash or sh shells: export PHENIX_TRUST_OTHER_ENV=1 or csh (C-shell): setenv PHENIX_TRUST_OTHER_ENV 1 in the script where you run mr_rosetta, or before you run it from the command line. *

2 1

Re: [phenixbb] strategy suggestion ???
by Michael Thompson 31 May '11

31 May '11

I once used the following strategy to search for ligands when the apo structure had already been determined. I was using Refmac at the time, but I don't see why you couldn't do it with phenix.refine. After you process your data, if your crystals with and without the ligands are isomorphous, then you don't necessarily need to do molecular replacement to get phases. After you process the data from your crystals with the ligands, you can take the mtz file and start refining with the pdb file for your apo structure. You can start with a rigid body refinement to account for any small shifts or rotations that may have taken place when you added the ligand. Then after that refine according to your protocol of choice. The main place you can run into trouble is if you have a polar space group. If you do the rigid body refinement and get a horrible R-factor, the there's a good chance you have indexed the two data sets in opposite orientations. This can be corrected by reindexing, I think the phenix tool for this is xmanip but someone can correct me if I'm wrong. Good luck, Mike ----- Original Message ----- From: "ChenTiantian" <chentiantian2010(a)gmail.com> To: "PHENIX user mailing list" <phenixbb(a)phenix-online.org> Sent: Monday, May 30, 2011 6:31:51 PM GMT -08:00 US/Canada Pacific Subject: [phenixbb] strategy suggestion ??? Hi, I just started learning crystallography months ago, and the software I used most is XDS, HKL2000, phenix and coot. Now I am working on a series of datasets-----the same target protein with various ligands. The following is my refinement workflow: 1, Processing data with XDS (or HKL2000 for some of the datasets) 2,Molecule replacement with phaser. 3,Check and modify the structure residue by residue with COOT, 4,phenix.refine with the following command line: phenix.refine protein.1.mtz protein.pdb \ strategy=individual_sites+individual_sites_real_space+individual_adp+tls \ simulated_annealing=true simulated_annealing.start_temperature=1000 \ simulated_annealing.cool_rate=10 main.number_of_macro_cycles=5 \ ordered_solvent=true refinement.input.xray_data.labels="F,SIGF" \ output.prefix=myprotein 5,Fit the ligand and redo step 3&4; stop refining when the Rwork and Rfree looks reasonable. 6,Chage the strategy line to strategy=individual_adp adp.individual.isotropic=all \ to remove the ANISOU lines in pdb file. Any suggestion for this workflow ? or how do you always deal with the similar case ? cause I'm new about this field,maybe I did something stupid. Ps, the value of Rwork and Rfree are too close, like, Final: r_work = 0.1741 r_free = 0.1817 bonds = 0.036 angles = 1.818 Is that reasonable ? -- Drug Discovery and Design Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences Address: Room 101, 646 Songtao Road, Zhangjiang Hi-Tech Park, Pudong New Area, Shanghai, 201203, P.R. China _______________________________________________ phenixbb mailing list phenixbb(a)phenix-online.org http://phenix-online.org/mailman/listinfo/phenixbb -- Michael C. Thompson Graduate Student Biochemistry & Molecular Biology Division Department of Chemistry & Biochemistry University of California, Los Angeles miket(a)chem.ucla.edu

1 0

strategy suggestion ???
by ChenTiantian 31 May '11

31 May '11

Hi, I just started learning crystallography months ago, and the software I used most is XDS, HKL2000, phenix and coot. Now I am working on a series of datasets-----the same target protein with various ligands. The following is my refinement workflow: 1, Processing data with XDS (or HKL2000 for some of the datasets) 2,Molecule replacement with phaser. 3,Check and modify the structure residue by residue with COOT, 4,phenix.refine with the following command line: phenix.refine protein.1.mtz protein.pdb \ strategy=individual_sites+individual_sites_real_space+individual_adp+tls \ simulated_annealing=true simulated_annealing.start_temperature=1000 \ simulated_annealing.cool_rate=10 main.number_of_macro_cycles=5 \ ordered_solvent=true refinement.input.xray_data.labels="F,SIGF" \ output.prefix=myprotein 5,Fit the ligand and redo step 3&4; stop refining when the Rwork and Rfree looks reasonable. 6,Chage the strategy line to strategy=individual_adp adp.individual.isotropic=all \ to remove the ANISOU lines in pdb file. Any suggestion for this workflow ? or how do you always deal with the similar case ? cause I'm new about this field,maybe I did something stupid. Ps, the value of Rwork and Rfree are too close, like, Final: r_work = 0.1741 r_free = 0.1817 bonds = 0.036 angles = 1.818 Is that reasonable ? -- Drug Discovery and Design Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences Address: Room 101, 646 Songtao Road, Zhangjiang Hi-Tech Park, Pudong New Area, Shanghai, 201203, P.R. China

4 3

Inconsistent R-factors in PDB file
by Schubert, Carsten [PRDUS] 30 May '11

30 May '11

Hi, Not sure if this has been asked before but recently I noticed phenix seems to report inconsistent R-values after each refinement. The final values reported in the screen output after refinement match the values reported in the 'Quick Facts' section, but are not matched by the values in the PDB deposition section. Attached is an example: ... REMARK ******************** REFINEMENT SUMMARY: QUICK FACTS ******************* REMARK Start: r_work = 0.1907 r_free = 0.2151 bonds = 0.004 angles = 0.727 REMARK Final: r_work = 0.1838 r_free = 0.2141 bonds = 0.005 angles = 0.832 REMARK ************************************************************************ ... REMARK IF THIS FILE IS FOR PDB DEPOSITION: REMOVE ALL FROM THIS LINE UP. REMARK 3 REMARK 3 REFINEMENT. REMARK 3 PROGRAM : PHENIX (phenix.refine: 1.7.1_743) REMARK 3 AUTHORS : Adams,Afonine,Chen,Davis,Echols,Gopal, REMARK 3 : Grosse-Kunstleve,Headd,Hung,Immormino,Ioerger,McCoy, REMARK 3 : McKee,Moriarty,Pai,Read,Richardson,Richardson,Romo, REMARK 3 : Sacchettini,Sauter,Smith,Storoni,Terwilliger,Zwart REMARK 3 REMARK 3 REFINEMENT TARGET : ML REMARK 3 REMARK 3 DATA USED IN REFINEMENT. REMARK 3 RESOLUTION RANGE HIGH (ANGSTROMS) : 1.950 REMARK 3 RESOLUTION RANGE LOW (ANGSTROMS) : 37.113 REMARK 3 MIN(FOBS/SIGMA_FOBS) : 1.33 REMARK 3 COMPLETENESS FOR RANGE (%) : 92.21 REMARK 3 NUMBER OF REFLECTIONS : 29410 REMARK 3 REMARK 3 FIT TO DATA USED IN REFINEMENT. REMARK 3 R VALUE (WORKING + TEST SET) : 0.1894 REMARK 3 R VALUE (WORKING SET) : 0.1877 REMARK 3 FREE R VALUE : 0.2122 REMARK 3 FREE R VALUE TEST SET SIZE (%) : 6.74 REMARK 3 FREE R VALUE TEST SET COUNT : 1982 For regular work I really don't care as long as the values are consistent and tell me if the refinement reached convergence, but for a deposition it matters a bit more. So which values should I report for a paper? Cheers, Carsten

2 1

Iterative build composite omit hanging
by José Trincão 30 May '11

30 May '11

Hello, I tried to run an iterative build composite omit for my protein - P1, 1350 residues/monomer, 4 molecules in the A.U.. After about 1.5 weeks running almost all was done except one omit region which was stuck. I had two processes running indefinitely (reduce and python) with no output. I tried to rerun that region alone but it gets stuck also with reduce producing no output. What should I do to get the complete map (I am going to merge all the other regions together, but would like to know why this particular one is problematic). Thanks! Jose José Trincão, PhD CQFB@FCT-UNL 2829-516 Caparica, Portugal "It's very hard to make predictions... especially about the future" - Niels Bohr

2 1

isotropic
by ChenTiantian 30 May '11

30 May '11

-- Drug Discovery and Design Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences Address: Room 101, 646 Songtao Road, Zhangjiang Hi-Tech Park, Pudong New Area, Shanghai, 201203, P.R. China

1 0

sigmaA-weigted 2FoFc maps by only mtz file
by zhangh1＠umbc.edu 28 May '11

28 May '11

Hi, If I only provide a mtz file with Fo, Fc and phase data, and do NOT have the pdb file, can any phenix utility calculate the sigmaA-weigted 2FoFc maps? Thanks! Hailaing

2 6

SA omit map for ligand
by Tjaard Pijning 27 May '11

27 May '11

Hi, I am trying to make an SA-omit Fo-Fc map for a sugar ligand (chain B) bound to protein (structure solved by MR), without doing any rebuilding of protein or ligand – just calculate an Fo-Fc map that I can load into PyMol. So far I used the AutoBuild – Create omit map module with these settings : - input files : model (protein+ligand+waters), data (.mtz), restraints for sugar ligand (.cif) - omit map type : simulated annealing - omit region : omit selection - omit map options : keep input waters, omit chain list : omit_chain_list chainB However, the ligand gets modified even though I use “omit_chain_list chainB”. What are the parameters to set to keep the model (protein, ligand, waters) from being rebuilt ? Furthermore, what is the procedure to load the SA Fo-Fc omit map into pymol ? Phenix version is 1.6.4-486. Thanks in advance, Tjaard Ing. Tjaard Pijning Research Assistant Protein Crystallography Group Nijenborgh 7 9747 AG Groningen The Netherlands (31)(0)50 363 43 85

4 3

get_cc_mtz_pdb results
by Jerome Nwachukwu 27 May '11

27 May '11

Hi: I am using get_cc_mtz_pdb to compare some structural models and found differences in overall CC around ~0.0056 and local CC around 0.0033. How much difference in local and overall CC values between two models is generally considered significant? Thank you, -Jerome Jerome C. Nwachukwu, PhD Dept. of Cancer Biology The Scripps Research Institute-Florida 130 Scripps Way, 2C1 Jupiter, FL 33458

2 2

a new PDB clustering tool available in phenix from nightly build dev-756
by Francois Berenger 27 May '11

27 May '11

Dear Phenix users, Starting from the nightly build dev-756, a new clustering tool called DURANDAL is included in Phenix. It performs exact clustering of proteins in PDB format, fast. It is primarily aimed at people working in protein structure prediction and having to deal with a lot of protein models. The correspondence used to compute CARMSD for a pair of proteins is 1 to 1. If two proteins A and B are to be optimally superposed, both PDB files (A.pdb and B.pdb) must have the same numbers of alpha carbons. The first alpha carbon in A.pdb will be matched to the first alpha carbon in B.pdb, and so on. Clustering up to 20,000 PDB files should work on most machines. Workstations with a lot of memory will be able to cluster even more files. Examples: --------- * to perform exact clustering using a fixed cutoff value of 3.0 Angstroms: durandal.cluster_pdbs -i [file_containing_one_PDB_filename_per_line]\ -d 3.0 -o out -s -v * there is also another tool included, to compute CARMSD of many files versus a reference one (the first in the list is considered the reference): durandal.rank_pdbs [file_containing_one_PDB_filename_per_line] If you want to know more about how the software is working, you can have a look at http://bioinformatics.oxfordjournals.org/content/27/7/939 Don't hesitate to ask if you encounter any problem while using the software. Best regards, Francois Berenger.

1 0

ADP refinement
by Christian Roth 26 May '11

26 May '11

Dear all, I am sorry for this very stupid question, but I probably overlooked the relevant text in the documentation. I requested the refinement of a heavy atom (chloride) with anisotropic ADP. In the log than isotropic refined ADP were set for 0 atoms. I guess I have to define all other chains as isotropic. Is that right. I have a lot of chains, it would be possible but a pain (more than 40 chains). Is there a something like refine anisotropic chain x and all others isotropic? Thanks and best Regards Christian

2 1

Quick PHENIX install question
by Yuri 26 May '11

26 May '11

Hello PHENIXers, I just got a computer with Red Hat workstation 5 running 64-bit architecture. Which build should I install? Thank you -- Yuri Pompeu

2 1

Equations of transformation
by Hena Dutta 26 May '11

26 May '11

Dear Members, Sorry, if my question is out of the field. I want to know the mathematical relations between 2 sets of co-ordinates, one before and one after the rotation. I used COOT to rotate the side chain (containing atoms N1, C2, N3, C4, C5, C6, N6, N7 and C8) of Adenine by 180 degree about the glycosidic bond (C1’-N9). How the initial co-ordinates (Xi, Yi, Zi) and final co-ordinates (Xf, Yf, Zf) are mathematically related? I shall be very grateful if someone can tell me the equations. Regards... Hena

5 4

Refining sucrose
by Priit Eek 23 May '11

23 May '11

Hi, I couldn't find this (probably basic knowledge) in the manual or here, so... I have a problem with refining a sucrose ligand. It is not covalently bound to the protein – just electrostatic interaction. The density isn't the best, but it's still obvious it must be there. LigandFit can find the blob and fits the sucrose in there quite well, but when I run phenix.refine it distorts the geometry of the molecule and complains about it later in MolProbity - Basic Geometry validation in almost every possible way. How can I apply more stringent restraints to the ligand so its geometry wouldn't get so twisted? This might not be the most appropriate list for my second question, but the problem is the same and somebody probably knows: how to manually modify that sucrose in Coot (eg rotate only one ring or a -CH2OH group) or perform real space refinement? Right now I can only rotate/translate the whole sucrose and trying to perform a real space refinement results in an error ("Failed to match (to the dictionary) the following model atom names: SUC C1 C2 C3 C4 C5 C6 O1 O2 O3 O4 O6. That would cause exploding atoms, so the refinement didn't start.") I haven't specified any custom cif libraries as SUC should be present in the built-in library of both Phenix and Coot. I'm using Phenix v1.7.1-743 and Coot v0.6.2-pre-1-rev3450 on OS X 10.6.7. Thank you in advance! Regards, Priit Eek

2 1

Call for articles and short communications for Computational Crystallography Newsletter
by Nigel Moriarty 23 May '11

23 May '11

Dear Colleagues I'm accepting short communications and articles of general interest to crystallographers working on macromolecular systems for the next issue of the Computational Crystallography Newsletter. You can see recent issues at http://www.phenix-online.org/newsletter/ as well as the word template for article submission. The deadline is June 17th but early submission is appreciated. Nigel -- Nigel W. Moriarty, PhD Building 64R0246B, Physical Biosciences Division Lawrence Berkeley National Laboratory Berkeley, CA 94720-8235 Phone : 510-486-5709 Email : NWMoriarty(a)LBL.gov Fax : 510-486-5909 Web : CCI.LBL.gov

1 0

Weird R-free discrepancy.
by Yuri 22 May '11

22 May '11

I was working on a model for at 1.9 A using the previous phenix release (650). I was almost finished r-free was 0.197. I installed the newest release and refined again and my r-free went up to 0.228!!. The values for bsol and ksol look significantly different. I do not understand where the biggest difference is coming from... Below I pasted part of the .log files. Old and new, respectively. =============================== refinement start ============================== ----------structure factors based statistics (before refinement)---------- ----------X-ray data---------- |--(resolution: 1.90 - 28.74 A; n_refl. = 32203)------------------------------| | | | r_work= 0.1747 r_free= 0.1967 ksol= 0.40 Bsol= 39.70 scale= 0.064 | | | | overall anisotropic scale matrix (Cartesian basis; B11,B22,B33,B12,B13,B23):| | (-1.65,-1.65,3.31,-0.00,0.00,0.00); trace/3= -0.00 | | | | maximum likelihood estimate for coordinate error: 0.21 A | | x-ray target function (ml) for work reflections: 2.567485 | =============================== refinement start ============================== ----------structure factors based statistics (before refinement)---------- ----------X-ray data---------- |--(resolution: 1.90 - 28.74 A; n_refl. = 34277)------------------------------| | | | r_work= 0.1911 r_free= 0.2279 ksol= 0.40 Bsol= 59.62 scale= 0.063 | | | | overall anisotropic scale matrix (Cartesian basis; B11,B22,B33,B12,B13,B23):| | (-1.73,-1.73,3.47,0.00,0.00,-0.00); trace/3= 0.00 | | | | maximum likelihood estimate for coordinate error: 0.55 A | | x-ray target function (ml) for work reflections: 2.591014 | |-----------------------------------------------------------------------------| -- Yuri Pompeu

2 1

BRU error
by Krystle Williams McLaughlin 20 May '11

20 May '11

Hi, I'm using phenix for refinement and I keep getting an error. My DNA strand has a bromo-uridine in it and phenix.refine keeps crashing on this nucleotide. Here is the error: Number of atoms with unknown nonbonded energy type symbols: 40 "ATOM 2770 P BRU P 5 .*. P " "ATOM 2771 O1P BRU P 5 .*. O " "ATOM 2772 O2P BRU P 5 .*. O " "ATOM 2773 O5* BRU P 5 .*. O " "ATOM 2774 C5* BRU P 5 .*. C " "ATOM 2775 C4* BRU P 5 .*. C " "ATOM 2776 O4* BRU P 5 .*. O " "ATOM 2777 C1* BRU P 5 .*. C " "ATOM 2778 N1 BRU P 5 .*. N " "ATOM 2779 C6 BRU P 5 .*. C " ... (remaining 30 not shown) Here is what I have in my pdb: ATOM 2901 P BRU E 5 33.977 -16.324 24.303 1.00 27.21 P ATOM 2902 O1P BRU E 5 34.027 -15.724 22.919 1.00 28.22 O ATOM 2903 O2P BRU E 5 33.857 -15.340 25.393 1.00 25.06 O ATOM 2904 O5* BRU E 5 32.721 -17.236 24.418 1.00 28.02 O ATOM 2905 C5* BRU E 5 32.487 -18.228 23.495 1.00 28.50 C ATOM 2906 C4* BRU E 5 31.047 -18.347 23.082 1.00 32.59 C ATOM 2907 O4* BRU E 5 30.337 -18.858 24.152 1.00 28.61 O ATOM 2908 C1* BRU E 5 29.369 -17.891 24.535 1.00 29.63 C ATOM 2909 N1 BRU E 5 29.279 -17.855 25.962 1.00 29.72 N ATOM 2910 C6 BRU E 5 30.426 -17.755 26.799 1.00 30.57 C ATOM 2911 C5 BRU E 5 30.251 -17.738 28.232 1.00 30.89 C ATOM 2912 BR BRU E 5 31.869 -17.591 29.321 1.00 57.08 Br ATOM 2913 C4 BRU E 5 28.917 -17.845 28.826 1.00 30.70 C ATOM 2914 O4 BRU E 5 28.768 -17.825 30.033 1.00 29.36 O ATOM 2915 N3 BRU E 5 27.785 -17.945 27.981 1.00 29.36 N ATOM 2916 C2 BRU E 5 27.961 -17.966 26.552 1.00 29.05 C ATOM 2917 O2 BRU E 5 26.994 -18.064 25.848 1.00 35.29 O ATOM 2918 C3* BRU E 5 30.464 -17.006 22.751 1.00 34.58 C ATOM 2919 C2* BRU E 5 29.845 -16.605 23.968 1.00 33.10 C ATOM 2920 O3* BRU E 5 29.526 -17.213 21.768 1.00 37.93 O I have already run a BRU molecule through ReadySet and when I use the cif file generated from this phenix.refine will run through but when the Bromine atom is in the structure has a large negative density peak in the Fo-Fc map. I think it may just be a naming thing. BRU is already in the monomer library so I don't understand why I get an error at first also without the readyset cif file. Any suggestions? Thanks, Krystle

2 1

Phenix 1.7 with Rosetta and PBS/Torque
by Jeremy Mann 20 May '11

20 May '11

Reading the documentation is is unclear with queue system will work with Phenix 1.7. qsub is mentioned but not referred as the qsub from Torque/PBS or qsub from Grid Engine. Which qsub are the developers referring to? -- Jeremy Mann jeremy(a)biochem.uthscsa.edu University of Texas Health Science Center Bioinformatics Core Facility http://www.bioinformatics.uthscsa.edu Phone: (210) 767-3419 ^^^^^^^^ ** NEW OFFICE NUMBER **

3 3

map error?
by Christian Roth 19 May '11

19 May '11

Dear all, I got a error message in the last round of phenix.refine which seesm to be related to map or mtz file creation (error.txt attached). I used the GUI and I unchecked the map creation option in ccp4 or Xplor before I started the run. I am not sure whats going wrong here. Please could you point me to the problem. Best regards Christian

3 2

AutoSol
by Nora Cronin 19 May '11

19 May '11

Whilst I can input SAD SAD data ie two different SAD datasets using Group =1 and Group =2 with the "Wavelength" keyword, how do I input 2 different crystals using the "crystal_info" keyword which does not recognize Group = Inputting .sca unmerged anomalous data how do I input data for 2 different crystal_info keywords with different heavy atoms and different unit cells? thanks Nora Dr N B Cronin X-Ray Laboratory Manager Tel. (+44) 020 7153 5445 Section of Structural Biology Fax. (+44) 020 7153 5457 Chester Beatty Laboratories mobile: (+44) 07787554059 The Institute of Cancer Research 237 Fulham Road London SW3 6JB The Institute of Cancer Research: Royal Cancer Hospital, a charitable Company Limited by Guarantee, Registered in England under Company No. 534147 with its Registered Office at 123 Old Brompton Road, London SW7 3RP. This e-mail message is confidential and for use by the addressee only. If the message is received by anyone other than the addressee, please return the message to the sender by replying to it and then delete the message from your computer and network.

3 3

mr_rosetta on a supercomputer
by Patricia Amara 18 May '11

18 May '11

Hello, I have been trying to run phenix.mr_rosetta on a supercomputer where we are limited to 24h for parallel jobs and we have different queues for job mono. Thus to run we have two files. We do not really understand how the code proceeds with nproc so we just put 6 for now. When we sent it, it created 1000 jobs on MR_ROSETTA_1/GROUP_OF_PLACE_MODEL_1/RUN_1/GROUP_OF_ROSETTA_REBUILD_1/RUN_1/REBUILD_IN_SETS_1 it sent 10 jobs mono, but ran only 5 first than the other five. Then the job running the script hangs. when we looked at the log files in the MR_ROSETTA_1/GROUP_OF_PLACE_MODEL_1/RUN_1/GROUP_OF_ROSETTA_REBUILD_1/RUN_1/REBUILD_IN_SETS_1 directories it filled up only 4 if we re-run only 3 and never the same. do you have any idea what we're doing wrong? thanks in advance for your input Patricia Amara ps; Thus to run we have two files. One to send the script #!/bin/bash #MSUB -e ros1_%I.e #MSUB -o ros1_%I.o #MSUB -n 6 #MSUB -T 86400 cd /scratch/cont003/amara/Rosetta-NadA ./rosetta.run and the script itself: phenix.mr_rosetta \ seq_file=/scratch/cont003/amara/Rosetta-NadA/NADA.seq \ data=/scratch/cont003/amara/Rosetta-NadA/NadA-7-BM16-remote-no_ano-P21212-Rfree_shell.mtz \ search_models=/scratch/cont003/amara/Rosetta-NadA/model_2.pdb \ run_prerefine=True \ number_of_prerefine_models=1000 \ fragment_files = /scratch/cont003/amara/Rosetta-NadA/aat000_03_05.200_v1_3.gz \ fragment_files = /scratch/cont003/amara/Rosetta-NadA/aat000_09_05.200_v1_3.gz \ rescore_mr.relax=False \ rosetta_models=20 \ ncs_copies=3 \ space_group=p21212 \ use_all_plausible_sg=False \ max_wait_time=100 \ nproc=6 \ group_run_command= "/usr/share/lsf/7.0/linux2.6-glibc2.3-x86_64/bin/bsub -n 1 -q mono -W 03:00"

2 1