Dear Victor, Thank you very much for telling us that. Although it is a little strange to me because the same set of the fortran programs and python scripts are used in both the webserver version and standalone version of BALBES (except one fortran program I will mention later), we try to find why that happened.
>From what shows in your email, we see the final solution given by the web server version of BALBES is from the template search model, as1m1, which is an assembly model (built from two input sequences), while the result from the standalone version is from the model, sq1st2m1, which is a single molecule model (built from one input sequence). The former should be better than the latter. But we want to see why the standalone version failed to find that assembly search model. You run our web server on Jun 9 and now you run the standalone BALBES released on June 25. During that time TWO things changed in BALBES (in both web server and standalone versions). (1) We updated the database within BALBES. Your email shows (a)the web server running is via BALBES of Version 1.0.0.May_16_2009, which is version 1.0.0 with the database updated on May 16 2009 (b) your local running is via BALBES of Version 1.0.0.Jun_1_2009, which is version 1.0.0 with the database updated On June 1 2009. (2) We changed a fortran program 'get_structure_DB', which is responsible for building template search models, on Jun 14-15 2009 Both two changes should enhance the performance of BALBES instead of deteriorating it. But you results show the opposite. Could it be possible for you to provide us the files recording the running processes, i,e, the files called "Process_information.txt" under subdirectory ....../results/ for both runs so that we can see what sets of template search models were given in both versions of BALBES and then what went wrong in the standalone version you downloaded ? You can send the files to any of one of us (Garib, Alexei, or me), Thank you very much. Fei 2009/6/29 Victor Alves <vdal...@fmv.utl.pt>: > Hello > > > > After a helpful hand (on a Sunday !!) from Dr. Fei Long I installed BALBES > 1.0.0 in my Ubuntu Linux machine and in order to try it, I ran the same job > I had previously ran in the WWW Server. > > Surprisingly the results were worse in my local BALBES. > > > > Results from the WWW: > > > > BALBES Version is 1.0.0.May_16_2009 > > JOB FOR STARTS AT Tue Jun 9 12:17:31 2009 > > > > ############################################################# > > | Please cite the following paper in any publication | > > | arising from use of BALBES: | > > | F.Long, A.Vagin, P.Young and G.N.Murshudov | > > | "BALBES: a Molecular Replacement Pipeline" | > > | Acta Cryst. D64 125-132(2008) | > > ############################################################# > > You have provide 2 sequences in the input sequence file > > number of amino acids in the sequence 1 is 65 > > number of amino acids in the sequence 2 is 146 > > > > #-----------------------------------------------------------# > > # Structure Factor Data Analysis # > > #-----------------------------------------------------------# > > Information about the structure to be solved > > |------------------------------------------| > > | SPACE GROUP | I 41 2 2 | > > |----------------|-------------------------| > > | CELL LENGTH | 130.050 130.050 70.190 | > > |----------------|-------------------------| > > | CELL ANGLE | 90.000 90.000 90.000 | > > |----------------|-------------------------| > > | RESOLUTIN_MAX | 1.95 | > > |----------------|-------------------------| > > | RESOLUTIN_MIN | 32.79 | > > |----------------|-------------------------| > > | DATA_COMPL | 99.86% | > > |----------------|-------------------------| > > | PSEUDO_TRANS | not detected | > > |----------------|-------------------------| > > | ALPHA_TWIN | not detected | > > |----------------|-------------------------| > > > > ? > > > > The best Q after all model calculations is 0.7060 > > > > SOLUTION SUMMARY > > #-------------------------------------------------------------------------------------------# > > # A structure is suggested by > BALBES # > > # Its probability to be a solution is > 99.0% # > > #-------------------------------------------------------------------------------------------# > > The solution model index is as1m1 > > |-------------------------------------------------------------------------------------------| > > | ITS PDB FILE | > results/refmac_final_result.pdb | > > |-------------------------------------------------------------------------------------------| > > | ITS MTZ FILE | > results/refmac_final_result.mtz | > > |-------------------------------------------------------------------------------------------| > > | R_ini/R_fin | 0.5130/0.3820 | Rfree_ini/Rfree_fin | > 0.4920/0.4070 | > > > > Results from local BALBES: > > > > BALBES Version is 1.0.0.Jun_1_2009 > > JOB FOR STARTS AT Sun Jun 28 22:51:50 2009 > > > > ############################################################# > > | Please cite the following paper in any publication | > > | arising from use of BALBES: | > > | F.Long, A.Vagin, P.Young and G.N.Murshudov | > > | "BALBES: a Molecular Replacement Pipeline" | > > | Acta Cryst. D64 125-132(2008) | > > ############################################################# > > You have provide 2 sequences in the input sequence file > > number of amino acids in the sequence 1 is 65 > > number of amino acids in the sequence 2 is 146 > > > > #-----------------------------------------------------------# > > # Structure Factor Data Analysis # > > #-----------------------------------------------------------# > > Information about the structure to be solved > > |------------------------------------------| > > | SPACE GROUP | I 41 2 2 | > > |----------------|-------------------------| > > | CELL LENGTH | 130.050 130.050 70.190 | > > |----------------|-------------------------| > > | CELL ANGLE | 90.000 90.000 90.000 | > > |----------------|-------------------------| > > | RESOLUTIN_MAX | 1.95 | > > |----------------|-------------------------| > > | RESOLUTIN_MIN | 32.79 | > > |----------------|-------------------------| > > | DATA_COMPL | 99.86% | > > |----------------|-------------------------| > > | PSEUDO_TRANS | not detected | > > |----------------|-------------------------| > > | ALPHA_TWIN | not detected | > > |----------------|-------------------------| > > > > ? > > > > The best Q after all model calculations is 0.4785 > > > > SOLUTION SUMMARY > > #-------------------------------------------------------------------------------------------# > > # A structure is suggested by > BALBES # > > # Its probability to be a solution is > 62.46% # > > #-------------------------------------------------------------------------------------------# > > The solution model index is sq1st2m1 > > |-------------------------------------------------------------------------------------------| > > | ITS PDB FILE | > results/refmac_final_result.pdb | > > |-------------------------------------------------------------------------------------------| > > | ITS MTZ FILE | > results/refmac_final_result.mtz | > > |-------------------------------------------------------------------------------------------| > > | R_ini/R_fin | 0.5520/0.5130 | Rfree_ini/Rfree_fin | > 0.5480/0.5190 | > > |-------------------------------------------------------------------------------------------| > > > > > > As expected the maps in Coot are a lot worse and also the model is much more > incomplete. > > > > Any ideas? > > > > Victor Alves > > ______________________________ > > > > VICTOR DIOGO ALVES, DVM PhD > > Professor Auxiliar > > Faculdade de Medicina Veterinária > > UTL - Lisboa > > __________________________ > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. >
