- modeller_usage - salilab.org

Multicore Jobs on QB3 Cluster Nodes
by Albion Baucom 08 Dec '08

08 Dec '08

How would I go about passing arguments to sge_qsub_job to get slave node jobs to use multiple cores on the QB3 cluster? I am familiar with using a SGE script to start parallel jobs on QB3 nodes using qsub. What I would like to do is use Modeller to start multiple jobs on nodes using multiple cores. For instance in a qsub script I might pass the following arguments #$ -pe smp 4 #$ -R yes Which uses 4 cores after using the queue to reserve that slave node. I am currently using the following Modeller directive to submit multiple jobs to the cluster j = sge_qsub_job(maxslave=jobparas.maxslave,options =jobparas.optionline,modeller_path=jobparas.modeller_path) where jobparas = jobparameters(maxslave=5, arch='opt64', queue='long.q', name='htpg', modeller_path="/diva1/home/modeller/SVN/bin/ modpy.sh \ /diva1/home/modeller/SVN/bin/ modslave.py") My question, is how do I pass the arguments "-pe smp 4" and "-R yes" for instance? Looking at the Modeller documentation, its not clear how those parameters could be passed as arguments to sge_qsub_job in this instance. Perhaps there is a better way to issue the parallelization options to Modeller? Thanks Albion

2 1

Re: [modeller_usage] parallel modeller
by Mauricio Carrillo Tripp 08 Dec '08

08 Dec '08

I ran into the same problem. I searched the wiki but couldn't find a solution. From the description provided in the thread is not obvious how to proceed. Is there a working script that could be used as a starting point to run parallel jobs in a cluster using PBS? I managed to submit jobs using independent modeller.py scripts differing in the start and end model parameter only: in file modeller01.py start=1 end=1 in file modeller02.py start=2 end=2 each one of these run in a different node (jobs submitted with qsub), but the two resulting models 1s58.B99990001.pdb and 1s58.B99990002.pdb are exactly the same, which brings me to my second question: Is there a way to specify the initial random seed used by each modeller script? Thanks! Starr Hazard wrote: > The references to parallelization seem to point rather strongly to the SGE scheduler... Not at all - the 'job' class is simply a bag of 'slave' objects. There is no requirement that you use any particular resource management system. For example, local_slave starts up a slave on the local machine (ideal if you have a multi-core machine). ssh_slave starts up a slave on a machine accessible by ssh, ideal if you have a cluster set up to allow passwordless ssh (or rsh) to individual nodes. The only slave classes which use SGE are sge_pe_slave and sge_qsub_slave. I wrote those because we happen to have an SGE cluster. But there's no reason why you couldn't write your own slaves to use PBS mechanisms. Can any of the commands eg sge_qsub_job(options, maxslave, seq=(), modeller_path=None, host=None) work with PBS scheduler? Well, obviously not sge_qsub_job, as the name would suggest, since that is a convenience class for SGE. Just use the regular job base class. If you then have a traditional ssh-to-any-node setup, all you then need to do is loop over the nodes in your machine file and make an ssh_slave for each one. Alternatively, it would be pretty simple to write a class that used the PBS TM mechanism via something like mpiexec. I suggest you put the result into the Modeller wiki, so that other PBS users can use or modify it. -- 0 | Mauricio Carrillo Tripp, PhD / | Department of Molecular Biology, TPC6 0 | The Scripps Research Institute \ | 10550 North Torrey Pines Road 0 | La Jolla, California 92037 / | trippm(a)scripps.edu 0 | http://www.scripps.edu/~trippm ** Aut tace aut loquere meliora silentio **

2 2

Re: [modeller_usage] R: R: question for a (putative?) disulfide bridge
by Modeller Caretaker 02 Dec '08

02 Dec '08

Anna Marabotti wrote: > I'm attaching here the pdb file of the model in which I would like to > create the disulfide bridge, and the > scripts I used to do it. My knowledge of Python language is very very > poor and so I'm doing a trial&error > procedure, but at present I obtain only errors... If I understand > well, I should only provide the PDB file to Modeller, > since I don't see any request of > alignment. The sequence then should be read from the PDB file, right? Correct. > The PDB file I have attached is the one that I obtained using > mutate_model: residue 55 was previously a Gly, > and now is a Cys in both subunits.So, the number of the residue > should be correct. Well, your first problem is that there is no residue "55" in your input PDB. There is, however, a residue "55:A". The chain ID must be specified if it is present in the PDB. > I tried to model a disulfide bridge between Cys55 of the subunit A > and Cys180 of the subunit B, and I tried: > > 1) model-patch_1.py (see attachment): here the error message: ... > KeyError: 'No such residue: 55:' OK, here it's because you didn't specify the chain IDs. > 2) model-patch_2.py (see attachment): here the error message: ... > KeyError: 'No such residue: 180:B' See http://salilab.org/modeller/9v5/manual/node403.html. By default, complete_pdb (and Modeller in general) renumbers residues starting from 1, and chains starting from A. This deals with PDBs containing duplicated residue numbers, and in the general case of comparative modeling, where there are gaps and multiple templates, the residue numbers in the target cannot be deduced from those in the templates anyway. So in your model, there is no residue 180:B; your residues are labeled 1:A through 359:A and 360:B through 718:B. So in your case you have two options: 1. Modify your script to a) not call special_patches and b) write out the resultant PDB file, e.g. # Read the sequence: code = '1R3ACYS55CYS55' mdl = complete_pdb(env, code) mdl.write(file='test.pdb') By inspecting test.pdb you can then figure out what the residue numbers should be for your disulfide bridge in the model. You can then use your original model-patch_2.py script with these corrected residue numbers. 2. Specify transfer_res_num=True when you call complete_pdb, to have Modeller use the original PDB file residue numbering. By the way, since all your script is doing is calculating the energy (not optimizing the model) you won't see the disulfide bridge pulling those two Cys residues together, although you may see some restraint violations in the log file if they're quite far apart. Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage

1 0

manual restaints in muultiple template modeling?
by plmallip＠mail.uh.edu 02 Dec '08

02 Dec '08

Hi, I want to build a model using two templates. One of the templates have >70% sequence homology while the other has <35% sequence homology. To the extent of my knowledge, automodel gives more weight to the template with more sequence homology. Is there any way I can manually assign more weight to the template with less sequence homology, at some of the regions? If there is, where can I get the details, as to how to do it? I appreciate any kind of suggestions/hints in this regard. Thanks & regards, Latha, Graduate student.

2 1

Atom has GB/SA radius 0.0000.
by Joaquim Rui Rodrigues 02 Dec '08

02 Dec '08

Dear Modellers, I made some changes to the script examples/automodel/loop.py to do loop refinement with dope_loopmodel in the presence of ligands. While this used to work well with modeller 9v4, with 9v5 I am getting the following error: e_gbsa__234E> Atom 8137 (type 112) has GB/SA radius 0.0000. Radii must all be greater than zero. Please edit the library file (e.g. modlib/solv.lib) to set a correct radius for this atom type. The atom is an MG ion. From solv.lib: # RSOL GBSA ASP ASPW SCKS OONS JRF MG 0.000 0.117 0.000 0.000 0.000 0.000 0.000 ! Magnesium Ion I suppose I should change the value of RSOL. Is there any recomended value? I foresee more problems for other atom types as solv.lib seems to have radii data only for protein atom types... Many thanks, Rui Rodrigues -- Webmail ESTG de Leiria (http://webmail.estg.ipleiria.pt)

2 1

Re: [modeller_usage] Transfer_xyz or other methods?
by Modeller Caretaker 02 Dec '08

02 Dec '08

plmallip(a)mail.uh.edu wrote: > I appreciate for taking time and explaining in detail. Its not that I > don't want to use automodel. But I have two questions > > 1. Is modeller going to accept the .ini file (I converted it to .pdb > manually) with undefined coordinates (999 numbers) in a script described at > > http://salilab.org/modeller/manual/node26.html > # Modeling using a provided initial structure file (inifile) Sure - Modeller will accept any model provided it has the same sequence as your target sequence. > 2. If I use automodel, I will again end up with an automated model where > I loose the bacbone Calpha conformation. i.e. my goal to retain backbone > conformation (for which I used transfer_xyz) will not be achieved. You can keep some of the atoms fixed during the refinement; see http://salilab.org/modeller/9v5/manual/node23.html In your case I guess you'd want to select everything but mainchain (or just C alphas) in aligned regions. You could do that with something like the following as your select_atoms() method: def select_atoms(self): aligned_regions = selection(self.residue_range('1:', '10:'), self.residue_range('20:', '30:')) aligned_backbone = aligned_regions.only_mainchain() # or # aligned_backbone = aligned_regions.only_atom_types('CA') return selection(self) - aligned_backbone Alternatively, you could have Modeller determine your insertions automatically with model.get_insertions() - see http://salilab.org/modeller/9v5/manual/node162.html Since this returns a list of all residue ranges that are aligned with gaps in your template, it is essentially the inverse of aligned_regions above, so you could say something like aln = self.read_alignment() aligned_regions = selection(self) - \ selection(self.get_insertions(aln, 0, 100, 0)) Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage

1 0

Residues in disallowed region
by Bo Yang 01 Dec '08

01 Dec '08

Hello, Modeller: The models built from Modeller 8v1 contain some residues (non-Gly and non-Pro) in the disallowed region based on Ramachandran plot. I wonder how this could happen? How to change these residues back to allowed region? I try to use energy minimization such as NAMD and Amber. After all, the minimized models had increased numbers of residues in the disallowed region. Thank you for your advice Bo

2 1

Transfer_xyz or other methods?
by plmallip＠mail.uh.edu 01 Dec '08

01 Dec '08

Hi, I am trying to obtain a model by transfer_xyz method, using a template which have less than 30% homology. I want to copy the coordinates of the template and if there is a gap, then obtain coordinates for those regions using model.build(). But, I end up with a model in pieces i.e. Some regions are displaced and built elsewhere. Even for some of the residues for which there is template structure, I get long linear chains built. May be the .ali file is not being used in getting the model! If transfer.xyz will not work if my the alignement has gaps in between, is there any other way to achieve this? I cannot exclude the gaps because, this will not serve the pupose of getting the template coordinates at equivalent regions where my model should have. The .ali file is >P1;1UW6 structureX:1UW6.pdb: 0 :E:215:E:undefined:undefined:-1.00:-1.00 EFDRADILYNIRQTSRPDVIPTQRD-RPVAVSVSLKFINILEVNEITNEVDVVFWQQTTWSDRTL AWNSSHS--PDQVSVPISSLWVPDLAAYN-AISKPEVLTPQLARVVSDGEVLYMPSIRQRFS CDVSGVDTESG-ATCRIKIGSWTHHSREISVDPTTE---------NSDDSEYFSQYSRFEIL DVTQKKNSVTYSCCPEA-YEDVEVSLNFR* >P1;NRmiss_210testb sequence:NRmiss_210testb:0 :E:215 :E:::-1.00:-1.00 SVMEDTLLSVLFETYNPKVRPAQTVGDKVTVRVGLTLTNLLILNEKIEEMTTNVFLNLAWTDYRL QWDPAAYEGIKKLRIPSSDVWQPDIVLMNNNDGSFEITLHVNVLVQHTGAVSWQPSAIYRSS CTIKVMYFPFDWQNCTMVFKSYTYDTSEVTLQHALDAKGEREVKEIVINKDAFTENGQWSIE HKPSRKNWR--SDDP--SYEDVTFYLIIQ* .py file is from modeller import * from modeller.automodel import * log.verbose() env = environ() env.libs.topology.read(file='$(LIB)/top_heav.lib') env.libs.parameters.read(file='$(LIB)/par.lib') aln = alignment(env, file='NRmiss_210anti_1UW6.ali', align_codes=('1UW6')) aln.malign3d(fit=False) aln.append(file='NRmiss_210anti_1UW6.ali', align_codes= 'NRmiss_210testb') md1=model(env) md1.generate_topology(aln['NRmiss_210testb']) md1.transfer_xyz(aln) md1.build(initialize_xyz=False, build_method='INTERNAL_COORDINATES') md1.write(file='NRmiss_210testb.ini') Any suggestions in this regard is appreciated. Thanks & regards, Latha, Graduate student.

2 1

dope batch processing
by jinlian 01 Dec '08

01 Dec '08

Hi all I want to use the 'DOPE' to do the filtering the docking structure of thousands of proteins, and I want to do batch possessing, but the script can only calculate one structure: from modeller import * from modeller.scripts import complete_pdb env = environ() env.libs.topology.read(file='$(LIB)/top_heav.lib') env.libs.parameters.read(file='$(LIB)/par.lib') # Read a model previously generated by Modeller's automodel class mdl = complete_pdb(env, '1acb.pdb') # Select all atoms in the first chain atmsel = selection(mdl.chains[0]) score = atmsel.assess_dope() I am not familiar with python, Could you tell me how to do the batch processing in the above script? Thank you very much!

2 1

dope batch processing
by Mauricio Carrillo Tripp 28 Nov '08

28 Nov '08

This is not a python solution, perhaps there is one, but what I did was to write a bash shell script to modify the python script changing the name of the model consecutively (might have to be modified for other shells): # DOPE for c in 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15; do sed 's/NAME/Model_'${c}'/g' model-DOPE.py > tmp.py; mod9v4 tmp.py ; echo "Model "${c}": ";grep 'DOPE' tmp.log >> Models_DOPE.dat; done In the python script (model-DOPE.py in my case) replace the line mdl = complete_pdb(env, '1acb.pdb') with mdl = complete_pdb(env, 'NAME.pdb') your previously built models should be named sequentially Model_01 to Model_15 (or the last one in your case), or change the regexp in the sed command above according to your file naming. At the end you'll have a list of DOPE values in the file Models_DOPE.dat. Hope this helps... Message: 1 > Date: Thu, 27 Nov 2008 10:03:30 +0800 > From: "jinlian" <jinlian05(a)lzu.cn> > Subject: [modeller_usage] dope batch processing > To: <modeller_usage(a)salilab.org> > Message-ID: <427750030.20199(a)lzu.cn> > Content-Type: text/plain; charset="us-ascii" > > Hi all > > I want to use the 'DOPE' to do the filtering the docking structure of > thousands of proteins, and I want to do batch possessing, but the script > can > only calculate one structure: > > > from modeller import * > from modeller.scripts import complete_pdb > > env = environ() > env.libs.topology.read(file='$(LIB)/top_heav.lib') > env.libs.parameters.read(file='$(LIB)/par.lib') > > # Read a model previously generated by Modeller's automodel class > mdl = complete_pdb(env, '1acb.pdb') > > # Select all atoms in the first chain > atmsel = selection(mdl.chains[0]) > > score = atmsel.assess_dope() > > I am not familiar with python, Could you tell me how to do the batch > processing in the above script? > Thank you very much! > -- 0 | Mauricio Carrillo Tripp, PhD / | Department of Molecular Biology, TPC6 0 | The Scripps Research Institute \ | 10550 North Torrey Pines Road 0 | La Jolla, California 92037 / | trippm(a)scripps.edu 0 | http://www.scripps.edu/~trippm ** Aut tace aut loquere meliora silentio **

1 0