June 2009 - modeller_usage

resuming a crashed job
by Søren Skovstrup Hansen 02 Jul '09

02 Jul '09

hDear Modellers A very simple question: Is it possible to resume a crashed Modeller job? I was running a more or less standard automodel job, but my computer crashed after building 200 out of 250 models. Hence, is it possible to resume this job and thereby evaluating and ranking all 250 build models via e.g. the DOPE score? BW Søren S Hansen ---------------------------------------------------------------------------- ----- Søren Skovstrup Hansen, Ph.D. student, M.Sc. Biostructural Research Group, Department of Medicinal Chemistry Faculty of Pharmaceutical Sciences, University of Copenhagen DK-2100 Copenhagen, Denmark Phone office (+45) 3533 6108 Mobile (+45) 2090 9866

2 1

Aligning two protein sequences.
by Sehrish Batool 30 Jun '09

30 Jun '09

Hi, As the reliability of predicted structure depends on the reliability of alignment of protein sequences, I want to ask which one tool can provide the best and most reliable alignment of two protein sequences that can be used as input in MODELLER? Thank you. Sehrish Batool

2 1

Re: [modeller_usage] Modelling Query
by Cool Hunk 26 Jun '09

26 Jun '09

Dear Ben, Thank you so much for the instant response, i have understood what you have told, but following is the error messege i get when i use the command "mod9v7 model.py" mod9v7 evaluate_model.py Could not find platform independent libraries <prefix> Could not find platform dependent libraries <exec_prefix> Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>] 'import site' failed; use -v for traceback my second question is about checking the reliability of an alignment in PIR format which we use to generate models... as i have generate this alignment after doing hand threading so i want to check its reliabilty, please tell me if i can do it using following script: # Example for: alignment.append(), alignment.write(), # alignment.check() # Read an alignment, write it out in the 'PAP' format, and # check the alignment of the N-1 structures as well as the # alignment of the N-th sequence with each of the N-1 structures. from modeller import * log.level(output=1, notes=1, warnings=1, errors=1, memory=0) env = environ() env.io.atom_files_directory = '../atom_files' aln = alignment(env) aln.append(file='ompA-1mal-align.ali ', align_codes='all') aln.write(file='ompA-1mal.pap', alignment_format='PAP') aln.write(file='ompA-1mal.fasta', alignment_format='FASTA') aln.check() as i have run this script i got following error, first using the "mod9v7 align.py" the same error messege i got for "mod9v7 model.py" and the remaining error messeges are about the alignment format which i could not understand mod9v7 align.py Could not find platform independent libraries <prefix> Could not find platform dependent libraries <exec_prefix> Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>] 'import site' failed; use -v for traceback Traceback (most recent call last): File "align.py", line 18, in ? aln.check() File "/usr/lib/modeller9v7/modlib/modeller/alignment.py", line 198, in check self.check_structure_structure(io=io) File "/usr/lib/modeller9v7/modlib/modeller/alignment.py", line 207, in check_structure_structure return f(self.modpt, io.modpt, self.env.libs.modpt, eqvdst) IOError: readlinef__E> Error encountered on file read: Is a directory thanking in advance Muhammad Noon --- On Thu, 6/25/09, Modeller Caretaker <modeller-care(a)salilab.org> wrote: From: Modeller Caretaker <modeller-care(a)salilab.org> Subject: Re: [modeller_usage] Modelling Query To: "Cool Hunk" <fascinatingheart2000(a)yahoo.com> Cc: modeller_usage(a)salilab.org Date: Thursday, June 25, 2009, 9:34 AM Cool Hunk wrote: > I am a new Modeller user, i am having some problems running Modeller, > while i run the script using command "*mod9v7 model.py*" it doesnt run > completely and generates errors What errors? We can only help if you tell us what the errors are. > but when i run the script using command > *"Python model.py" *it runs completely and generate all the pdb files of > models but not the log file for it even though i have included the > function "*log.verbose()"* in the script, and the same is the case for > model evaluation script and energies calculation script..... Python doesn't generate log files - the output comes out on standard output (i.e. your screen, if you didn't direct it somewhere else). To get something very similar to running mod9v7, use something like python model.py > model.log > my other question is about the model evaluation, that do i have to run > the evaluation script for each model individually and the compare the > final dope scores or the scripts takes all the models at once and give > me the final good model??? You can do it either way. You can evaluate a single model with the script at http://salilab.org/modeller/9v7/manual/node242.html Alternatively you can have automodel assess every model it builds automatically by setting assess_methods; an example is at http://salilab.org/modeller/9v7/manual/node20.html Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage

4 5

Any version without python embeded ?
by Stéphane Téletchéa 26 Jun '09

26 Jun '09

Although i am well aware of the *nightmare* it can be to follow all GPL programs, is there any considerations to produce modeller revisions *without* python embedded ? Except for python 3.0, there are only little differences between python 2.4 and above, so i'm not sure this will cause a lot of problems. Just a question, though (and a volonteer for beta testing if needed). Cheers, Stéphane

2 1

Modelling Query
by Cool Hunk 25 Jun '09

25 Jun '09

Dear Bin, I am a new Modeller user, i am having some problems running Modeller, while i run the script using command "mod9v7 model.py" it doesnt run completely and generates errors, but when i run the script using command "Python model.py" it runs completely and generate all the pdb files of models but not the log file for it even though i have included the function "log.verbose()" in the script, and the same is the case for model evaluation script and energies calculation script..... i dont know where i am making the mistake??? my other question is about the model evaluation, that do i have to run the evaluation script for each model individually and the compare the final dope scores or the scripts takes all the models at once and give me the final good model??? please help me understanding the scripts and the generated files to have good view of a good model. Thanking in advance. Muhammad Noon

2 1

Homology Modelling homodimer - alignment
by Christopher Mohr 25 Jun '09

25 Jun '09

Dear Modellers, first of all I'd like to thank you for helping with my last question. Now I have another question. I have to build a model by homology modelling for a protein(only sequence available) which builds a homodimer. The template-structure is a homodimer as well. I used the following script to align the sequence and the structure. from modeller import * env = environ() aln = alignment(env) mdl = model(env, file='y', model_segment=('FIRST:A','LAST:B')) aln.append_model(mdl, align_codes='y', atom_files='y.pdb') aln.append(file='x.ali', align_codes='x') aln.align2d() aln.write(file='xy.ali', alignment_format='PIR') aln.write(file='xy.pap', alignment_format='PAP') In the resulting alignment there is a separation ("/") between the two chains of the template sequence. But in the sequence of the target there isn't a separation of course. Do i have to put a separation in the sequence of the target in the alignment file or in the input file for the alignment script? Or whats the best way to handle this problem ? If I put the separation in the resulting alignment the alignment of the chains wouldn't be good. Does the x.ali file of my target have to look like this: >P1:x .... XXXX / XXXX ? Is it sufficient to use the following script afterwards to build the model out of the alignment ? from modeller import * from modeller.automodel import * env = environ() a = automodel(env, alnfile='xy.ali', knowns='y', sequence='x', assess_methods=(assess.DOPE, assess.GA341)) a.starting_model = 1 a.ending_model = 5 a.make() Thank you very much Christopher

2 1

PSC/NRBSC Workshop on QC/MM Methods and Applications
by Troy Wymore 24 Jun '09

24 Jun '09

****** A workshop that might be of interest to MODELLER users ******** The National Resource for Biomedical Supercomputing at the Pittsburgh Supercomputing Center will host a workshop entitled: "Methods and Applications of Hybrid Quantum Chemical/Molecular Mechanical Simulations of Biomolecular Systems" August 11-14, 2009 Pittsburgh, PA Website: http://shannon.psc.edu/mediawiki/index.php/QCMM_Workshop_2009 The workshop will present computational approaches for the simulation of enzymatic reactions. Lectures will be held on the relevant theoretical background including the applicability of Quantum Chemical (QC) methods for large molecules, re-parameterization of Semiempirical Molecular Orbital (SMO) methods for specific reactions, molecular mechanics (MM) force fields, hybrid QC/MM potential energy functions and simulation techniques on high performance computers. In-depth tutorials will be presented primarily with the use of the freely available pDynamo package developed in the lab of Dr. Martin J. Field and will also include developing SMO parameters for quantitatively accurate enzyme reaction simulations, the issues involved in constructing a QC/MM model and the various methods for generating energy profiles of bio-molecular reactions to determine mechanisms and structure-function relationships. Experience with pDynamo is helpful but not a prerequisite for attendance. The workshop is designed for graduate students, post-doctoral researchers and faculty in computational biophysics and physical organic chemistry who have experience with computational chemistry methods and seek to enhance their capabilities to include the modeling of enzymatic reactions. Participants are encouraged to discuss and work on bio-molecular systems from their own research during the workshop. Space is limited. This workshop is funded by a grant from the NIH National Center for Research Resources. These grants provide tuition and hotel costs, supercomputing time, and workshop materials for US academic participants. All participants are responsible for paying their own travel expenses. Complimentary continental breakfasts and lunches will be provided the days of the workshop. Registration is now open. http://www.nrbsc.org/education/workshops/register.php

1 0

Check out my photos on Facebook
by Salam Pradeep 23 Jun '09

23 Jun '09

I set up a Facebook profile where I can post my pictures, videos and events and I want to add you as a friend so you can see it. First, you need to join Facebook! Once you join, you can also create your own profile. Thanks, Salam To sign up for Facebook, follow the link below: http://www.facebook.com/p.php?i=763737958&k=R5B4XYQXT45M5ELHVCX3XP&r modeller_usage(a)salilab.org was invited to join Facebook by Salam Pradeep. If you do not wish to receive this type of email from Facebook in the future, please click on the link below to unsubscribe. http://www.facebook.com/o.php?k=91d207&u=1788717094&mid=abb3d7G6a9da826G0G8 Facebook's offices are located at 1601 S. California Ave., Palo Alto, CA 94304.

1 0

Superposition Problem
by Christopher Mohr 23 Jun '09

23 Jun '09

Dear Modellers, is there the possibility to create a representation of the superposition of two proteins in MODELLER? I've used the following example script to create the superposition of the two proteins: Example: examples/commands/superpose.py # Example for: selection.superpose() # This will use a given alignment to superpose Calpha atoms of # one structure (2ctx) on the other (1fas). from modeller import * env = environ() env.io.atom_files_directory = ['../atom_files'] mdl = model(env, file='1fas') mdl2 = model(env, file='2ctx') aln = alignment(env, file='toxin.ali', align_codes=('1fas', '2ctx')) atmsel = selection(mdl).only_atom_types('CA') r = atmsel.superpose(mdl2, aln) # We can now use the calculated RMS, DRMS, etc. from the returned 'r' object: rms = r.rms drms = r.drms print "%d equivalent positions" % r.num_equiv_pos mdl2.write(file='2ctx.fit') Do I have to use the outputed *.fit file to create a representation (*.pdb file for example) or is there any other possibility to create a representation to actually see the superpositioned structures of the proteins ? Thanks Christopher

2 1

iterative_structural_align question
by Mike White 18 Jun '09

18 Jun '09

I noticed that 9v7 has a new command that is designed to obtain the best set of parameters (gap_penalties_1d and gap_penalties_3d) for aligning two or more structures. Since the reference (Madhusudhan et al 2009) is not yet available, I had a few questions based on trying the command. (1) Based on the output log, the first parameter of gap_penalties_1d is varied from -150 to 0, and the second is varied from -50 to 0. The default values for gap_penalties_1d in for alignment.salign() are (-900, -50). What is the logic for the choice of ranges for gap_penalties_1d values in iterative_structural_align()? I always assumed that the default values of the various parameters in alignment.salign() represent a good starting point for these values. (2) I assume that the value "final max quality" equivalent to "QUALITY_SCORE(percentage)" in the standard output of alignment.salign(). Is this correct? (2) In the Advanced Tutorial the example script "multiple_template/ salign.py" runs salign multiple times to first get a rough alignment and then refines it by incorporating more information through changes in feature_weights. Is feature_weights fixed at (1, 0, 0, 0, 0, 0) in iterative_structural_align()? If so, is there any way to use iterative_structural_align() in a manner similar to that shown in the tutorial? Thanks!

2 1