- modeller_usage - salilab.org

[clustering] does MODELLER provide tools to cluster structures ?
by Benjamin SCHWARZ 08 Nov '11

08 Nov '11

I am looking for tools that would allow me to cluster a set of superposed structures such as those obtained for instance through an iterative salign run. Does MODELLER provide anything to perform (structures) clustering ?

2 3

Fwd: Fwd: automodel.max_molpdf recommended values and usage format
by Andrew Voronkov 08 Nov '11

08 Nov '11

Something was wrong with that script, I have taken model-outputs.py modified it and it worked fine. Best regards, Andrew -------- Пересылаемое сообщение -------- 08.11.2011, 16:41, "Andrew Voronkov" <drugdesign(a)yandex.ru>: Does it have name molpdf? I can't find DOPE or GA341 words in the log file. I have also got an error while ranking the solutions. Like this? But I have both DOPE and GA341, which of them is this? >> Summary of successfully produced models: Filename molpdf ---------------------------------------- chimera.B99990001.pdb 1580.34619 chimera.B99990002.pdb 1593.52393 chimera.B99990003.pdb 1693.15637 chimera.B99990004.pdb 1743.10718 chimera.B99990005.pdb 1586.12122 chimera.B99990006.pdb 1652.49170 chimera.B99990007.pdb 1477.14709 chimera.B99990008.pdb 1681.59668 chimera.B99990009.pdb 1625.87500 chimera.B99990010.pdb 1632.77686 chimera.B99990011.pdb 1689.23584 chimera.B99990012.pdb 1557.99890 chimera.B99990013.pdb 1622.84631 chimera.B99990014.pdb 1805.83228 chimera.B99990015.pdb 1525.45911 Also I have got an error while trying to select best solutions based on DOPE score (script is below): ok_models.sort(lambda a,b:cmp(a[key]), b[key])) KeyError: 'DOPE score' Here is the script, which was used: from modeller import * from modeller.automodel import * import sys log.verbose() # request verbose output env = environ() # create a new MODELLER environment to build this model in # directories for input atom files env.io.atom_files_directory = ['.'] # Read in HETATM records from template PDBs env.io.hetatm = True a = automodel(env, alnfile = '2rf53mhj.pir', # alignment filename knowns = ('3mhj', '2rf5'), # codes of the templates sequence = 'chimera') # code of the target assess_methods=(assess.DOPE, assess.GA341) # assess each loop with DOPE a.starting_model= 1 # index of the first model a.ending_model = 15 # index of the last model # (determines how many models to calculate) # Thorough MD optimization: a.md_level = refine.slow a.make() # do the actual homology modeling # Get a list of all successfully built models from a.outputs ok_models = [x for x in a.outputs if x['failure'] is None] # Rank the models by DOPE score key = 'DOPE score' if sys.version_info[:2] == (2,3): # Python 2.3's sort doesn't have a 'key' argument ok_models.sort(lambda a,b: cmp(a[key], b[key])) else: ok_models.sort(key=lambda a: a[key]) # Get top model m = ok_models[0] print("Top model: %s (DOPE score %.3f)" % (m['name'], m[key])) Best regards, Andrew 07.11.2011, 19:51, "Modeller Caretaker" <modeller-care(a)salilab.org>: > On 11/7/11 5:21 AM, Andrew Voronkov wrote: >> Dear Ben, very simple question . I have used DOPE assessment, but in >> which file can I find the assessment results for the models? It's not >> in resulted PDB, there is only objective function.Where can I find >> those DOPE and GA341 values? > In the log file. They are not put in the PDB file. > > Ben Webb, Modeller Caretaker > > -- > modeller-care(a)salilab.org http://www.salilab.org/modeller/ > Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage -------- Завершение пересылаемого сообщения --------

1 0

Fwd: automodel.max_molpdf recommended values and usage format
by Andrew Voronkov 07 Nov '11

07 Nov '11

Dear Ben, very simple question . I have used DOPE assessment, but in which file can I find the assessment results for the models? It's not in resulted PDB, there is only objective function.Where can I find those DOPE and GA341 values? I have used the next Python script for the assessment: # Loop refinement of an existing model from modeller import * from modeller.automodel import * log.verbose() env = environ() # directories for input atom files env.io.atom_files_directory = ['.'] # Read in HETATM records from template PDBs env.io.hetatm = True # Create a new class based on 'dopehr_loopmodel' so that we can redefine # select_loop_atoms (necessary) class MyLoop(dopehr_loopmodel): # This routine picks the residues to be refined by loop modeling def select_loop_atoms(self): # One loop from residue 19 to 28 inclusive return selection(self.residue_range('92:', '111:')) m = MyLoop(env, inimodel='3mhj.pdb', # initial model of the target sequence='3mhj' # code of the target ) assess_methods=(assess.DOPE, assess.GA341) # assess each loop with DOPE m.loop.starting_model= 1 # index of the first loop model m.loop.ending_model = 20 # index of the last loop model m.loop.md_level = refine.slow_large # loop refinement method m.make() Should those scores be available as molpdf values in .log file? Best regards, Andrew 01.11.2011, 22:55, "Modeller Caretaker" <modeller-care(a)salilab.org>: > On 11/01/2011 10:28 AM, Andrew Voronkov wrote: >> how on your opinion modeller objective function correlates with >> quality of the models. For example I have objective function value >> 1730 for one modelk and -750 which as I understand reflects free >> energy. Does it mean that I shouldn t consider the first one at >> all? > The Modeller objective function is simply a measure of how many > restraints were violated in building your model. While the score could > be thought of roughly as an "energy" (not free energy) this is certainly > not a rigorous measure of model quality. Use an assessment function such > as normalized DOPE for this purpose. >> Let's say I want to make cutoff -100 for the objective function. In >> which format it should be done? >> >> automodel.max_molpdf -100 > Build your models, then assess them and pick the best ones. max_molpdf > is designed as a filter to throw away the very worst models - it doesn't > act on the final PDF, but the score of intermediate (partially > optimized) models - the reasoning being to not waste time on trying to > optimize a model that is very bad. Thus max_molpdf is not really what > you want here. > > Ben Webb, Modeller Caretaker > > -- > modeller-care(a)salilab.org http://www.salilab.org/modeller/ > Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage -------- Завершение пересылаемого сообщения --------

2 1

parallelization with randomize.dihedrals?
by R K Belew 04 Nov '11

04 Nov '11

hi Ben et al, just when i thought i had my parallel version working nicely... i tried doing a run that had only one difference from previous: a.rand_method = randomize.dihedrals vs previous: a.rand_method = randomize.xyz but this throws an error: > File "seed_align_pp.py", line 93, in doAllAlign > for results in j.yield_tasks_unordered(): > File "/usr/lib/python2.6/dist-packages/modeller/parallel/job.py", line 145, in yield_tasks_unordered > for task in self._finish_all_tasks(): > File "/usr/lib/python2.6/dist-packages/modeller/parallel/job.py", line 163, in _finish_all_tasks > task = self._process_event(obj, s) > File "/usr/lib/python2.6/dist-packages/modeller/parallel/job.py", line 179, in _process_event > task = obj.task_results() > File "/usr/lib/python2.6/dist-packages/modeller/parallel/slave.py", line 61, in task_results > r = self.get_data(allow_heartbeat=True) > File "/usr/lib/python2.6/dist-packages/modeller/parallel/communicator.py", line 84, in get_data > (cmdtype, obj) = self._recv() > File "/usr/lib/python2.6/dist-packages/modeller/parallel/communicator.py", line 125, in _recv > raise RemoteError(obj.exc, self) > modeller.parallel.communicator.RemoteError: <type 'NameError'>: global name 'alignment' is not defined from <Slave on localhost> any guesses as to why? thanks again for your help. Rik

2 1

Build a model with multiple models
by junior 02 Nov '11

02 Nov '11

Hi, How do I build a PDB file containing more than 10 models? Thank you. Sam Kauba

3 2

Re: [modeller_usage] modeller_usage Digest, Vol 10, Issue 136
by Ashish Runthala 01 Nov '11

01 Nov '11

Dear Sir, No need to manual compilation of pooling of models in single PDB file. You can use linux with a simple bash script. cat *.pdb > Single.pdb Then you may end MODEL and ENDMDL manually or you can also write a for loop to automatically append these strings on ends of every single PDB file called. Ashish Runthala, Lecturer, Structural Biology Cell, Biological Sciences Group, BITS, Pilani Rajasthan, INDIA ----- Original Message ----- From: "modeller usage-request" <modeller_usage-request(a)salilab.org> To: "modeller usage" <modeller_usage(a)salilab.org> Sent: Wednesday, November 2, 2011 12:57:24 AM GMT +05:30 Chennai, Kolkata, Mumbai, New Delhi Subject: modeller_usage Digest, Vol 10, Issue 136 Send modeller_usage mailing list submissions to modeller_usage(a)salilab.org To subscribe or unsubscribe via the World Wide Web, visit https://salilab.org/mailman/listinfo/modeller_usage or, via email, send a message with subject or body 'help' to modeller_usage-request(a)salilab.org You can reach the person managing the list at modeller_usage-owner(a)salilab.org When replying, please edit your Subject line so it is more specific than "Re: Contents of modeller_usage digest..." Today's Topics: 1. automodel.max_molpdf recommended values and usage format (Andrew Voronkov) 2. Re: Build a model with multiple models (Modeller Caretaker) 3. Re: automodel.max_molpdf recommended values and usage format (Modeller Caretaker) 4. flexible ligand parametrization using nonstd_restraints(aln) (Andrew Voronkov) ---------------------------------------------------------------------- Message: 1 Date: Tue, 01 Nov 2011 21:28:48 +0400 From: Andrew Voronkov <drugdesign(a)yandex.ru> Subject: [modeller_usage] automodel.max_molpdf recommended values and usage format To: modeller_usage(a)salilab.org Message-ID: <175671320168528(a)web127.yandex.ru> Content-Type: text/plain Dear Modeller users, how on your opinion modeller objective function correlates with quality of the models. For example I have objective function value 1730 for one modelk and -750 which as I understand reflects free energy. Does it mean that I shouldn t consider the first one at all? Let's say I want to make cutoff -100 for the objective function. In which format it should be done? automodel.max_molpdf -100 is that correct? Sincerely yours, Andrey ------------------------------ Message: 2 Date: Tue, 01 Nov 2011 11:47:21 -0700 From: Modeller Caretaker <modeller-care(a)salilab.org> Subject: Re: [modeller_usage] Build a model with multiple models To: junior <sam.kauba(a)gmail.com> Cc: modeller_usage(a)salilab.org Message-ID: <4EB03EB9.80204(a)salilab.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed On 10/31/2011 07:16 AM, junior wrote: > How do I build a PDB file containing more than 10 models? If you really mean you want to make a single PDB file containing multiple models (which the PDB format allows, with a MODEL record at the start of each model and an ENDMDL record at the end) then you can't do that with MODELLER, for any number of models (10 or otherwise). The models produced by MODELLER are each in their own PDB file. If you simply want to make more than 10 models, each in a *different* PDB file, then go with Mensur's suggestions. If you really want a single PDB file containing multiple models, you will have to first produce multiple PDB files, then stick them together manually, inserting MODEL and ENDMDL records, either with a text editor or with some simple scripting. Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage ------------------------------ Message: 3 Date: Tue, 01 Nov 2011 11:55:27 -0700 From: Modeller Caretaker <modeller-care(a)salilab.org> Subject: Re: [modeller_usage] automodel.max_molpdf recommended values and usage format To: Andrew Voronkov <drugdesign(a)yandex.ru> Cc: modeller_usage(a)salilab.org Message-ID: <4EB0409F.8050901(a)salilab.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed On 11/01/2011 10:28 AM, Andrew Voronkov wrote: > how on your opinion modeller objective function correlates with > quality of the models. For example I have objective function value > 1730 for one modelk and -750 which as I understand reflects free > energy. Does it mean that I shouldn t consider the first one at > all? The Modeller objective function is simply a measure of how many restraints were violated in building your model. While the score could be thought of roughly as an "energy" (not free energy) this is certainly not a rigorous measure of model quality. Use an assessment function such as normalized DOPE for this purpose. > Let's say I want to make cutoff -100 for the objective function. In > which format it should be done? > > automodel.max_molpdf -100 Build your models, then assess them and pick the best ones. max_molpdf is designed as a filter to throw away the very worst models - it doesn't act on the final PDF, but the score of intermediate (partially optimized) models - the reasoning being to not waste time on trying to optimize a model that is very bad. Thus max_molpdf is not really what you want here. Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage ------------------------------ Message: 4 Date: Tue, 01 Nov 2011 23:21:23 +0400 From: Andrew Voronkov <drugdesign(a)yandex.ru> Subject: [modeller_usage] flexible ligand parametrization using nonstd_restraints(aln) To: modeller_usage(a)salilab.org Message-ID: <217441320175283(a)web128.yandex.ru> Content-Type: text/plain; charset=utf-8 I have a task to remodel the loop conformation which binds ligand in order to optimize the ligand-protein structure. Structure is X-ray, but ligand position is not known, so I take it from docking assuming that docking gives pose close to correct. So the goal is to model flexibility. I have asked this question couple of days a ago, but now it is more specific. I have used loop modeling with lgiand transfer and it works relatively well, but as I understand as far as this is non conventional residue it is treated as rigid and its interactions with protein are also treated as rigid according to what is written in the manual (listed below). So is it possible to define ligand as flexible using nonst restrains, I need to consult with someone who did. Actually it s kind of usage of Modeller as docking software, I don t knwo whether it is possible to parametrize ligand and optimize its position using simulated annealing and loop rmodeling, so kind of imitating induced fit. Has anyone tried this and question to Modeller Caretaker - is it potentially possible? Best regards, Andrew nonstd restraints(aln) 44 CHAPTER 4. COMPARATIVE MODELING CLASS REFERENCE This routine adds restraints to keep non-standard residues (anything treated as a HETATM or BLK residue, such as ligands or metal ions) in a reasonable conformation. You can override this method if you need to change these restraints. By default, four sets of restraints are built: ? For each residue that has no defined topology (generally BLK residues, used to transfer ligands directly from templates, as described in Section 2.2.1), intra-residue distances are all constrained to their template values. This causes each residue to behave as a rigid body. ? Inter-residue distances are constrained to template values if these are 7?A or less. This has the effect of preserving multiple-HETATM structures such as DNA chains. If the distances are 2.3?A or less they are assumed to be bonds and so are restrained more strongly and also excluded from the nonbonded list. ? Residue-protein atom distances are strongly constrained to template values (and excluded from the nonbonded list) if these are 2.3?A or less. This preserves chemical bonds between ligands and the protein. ? Residue-protein C distances are constrained to template values if these are 10?A or less and are not already bonded by the previous restraints. This preserves the ligand position. - so I get protein-ligand interactions very much restrained and all final models are quite close to the initial coordinates. ------------------------------ _______________________________________________ modeller_usage mailing list modeller_usage(a)salilab.org https://salilab.org/mailman/listinfo/modeller_usage End of modeller_usage Digest, Vol 10, Issue 136 ***********************************************

1 0

flexible ligand parametrization using nonstd_restraints(aln)
by Andrew Voronkov 01 Nov '11

01 Nov '11

I have a task to remodel the loop conformation which binds ligand in order to optimize the ligand-protein structure. Structure is X-ray, but ligand position is not known, so I take it from docking assuming that docking gives pose close to correct. So the goal is to model flexibility. I have asked this question couple of days a ago, but now it is more specific. I have used loop modeling with lgiand transfer and it works relatively well, but as I understand as far as this is non conventional residue it is treated as rigid and its interactions with protein are also treated as rigid according to what is written in the manual (listed below). So is it possible to define ligand as flexible using nonst restrains, I need to consult with someone who did. Actually it s kind of usage of Modeller as docking software, I don t knwo whether it is possible to parametrize ligand and optimize its position using simulated annealing and loop rmodeling, so kind of imitating induced fit. Has anyone tried this and question to Modeller Caretaker - is it potentially possible? Best regards, Andrew nonstd restraints(aln) 44 CHAPTER 4. COMPARATIVE MODELING CLASS REFERENCE This routine adds restraints to keep non-standard residues (anything treated as a HETATM or BLK residue, such as ligands or metal ions) in a reasonable conformation. You can override this method if you need to change these restraints. By default, four sets of restraints are built: • For each residue that has no defined topology (generally BLK residues, used to transfer ligands directly from templates, as described in Section 2.2.1), intra-residue distances are all constrained to their template values. This causes each residue to behave as a rigid body. • Inter-residue distances are constrained to template values if these are 7°A or less. This has the effect of preserving multiple-HETATM structures such as DNA chains. If the distances are 2.3°A or less they are assumed to be bonds and so are restrained more strongly and also excluded from the nonbonded list. • Residue-protein atom distances are strongly constrained to template values (and excluded from the nonbonded list) if these are 2.3°A or less. This preserves chemical bonds between ligands and the protein. • Residue-protein C distances are constrained to template values if these are 10°A or less and are not already bonded by the previous restraints. This preserves the ligand position. - so I get protein-ligand interactions very much restrained and all final models are quite close to the initial coordinates.

2 1

Homology Modeling of the Loop
by Alexander Predeus 01 Nov '11

01 Nov '11

Hello all, I'm very much a beginner at modeling, and this question might be very trivial. I hope you still would hint me about the answer, for it would help me a lot. I have a large protein (say 1500 residues, let's call it protein A) and its homologue (protein B) of about the same size. Both proteins have published crystal structures. Both proteins have a key structural element, loop L. The loop is poorly resolved in structure A, but well resolved in homologous structure B. Now, I can take just the L loop regions as separate PDB files from both A and B and do homology modeling of A based on B; this works really well. But I also would like the model to consider the surroundings (the rest of the structure) when evaluating the energies. Thus, I'd like to keep the REST of A's crystal structure intact, and only model loop L based on the structure of the same region in protein B. How can I achieve that? Thank you in advance! - Alex Predeus Michigan State University

2 1

parallelization and the master/slaves' env.io.atom_files_directory?
by R K Belew 01 Nov '11

01 Nov '11

i'm separating a script that has worked into separate slave and master modules as described in http://salilab.org/modeller/manual/node435.html. my master script looks like this: > j = job() > j.append(local_slave()) > j.append(local_slave()) > > for afile in alignFiles: > > alnRunDir = '%s%s/' % (dataDir,aname) > if not os.path.isdir(alnRunDir): > print "doAllAlign: Creating directory",alnRunDir > os.makedirs(alnRunDir) > > for runIdx in range(nruns): > j.queue_task(SeedTask(argumentList)) > > for result in j.yield_tasks_unordered(): > print result and my slave task like this: > class SeedTask(task): > """A task to read in a PDB file on the slave, and return the resolution""" > > def run(self,dataDir,afile,aname,runIdx,randMethod,knownLbl,seqLbl,nmodPerRun,atomDeviation ): > env = environ() > env.io.atom_files_directory = [dataDir] > arg4seed = - (int(runIdx) + 2) > env = environ(arg4seed) > > runDir = '%s/%s/run_%02d/' % (dataDir,aname,runIdx) > if not os.path.isdir(runDir): > print "doAllAlign: Creating directory",runDir > os.makedirs(runDir) > > os.chdir(runDir) > > a = automodel(env, alnfile=afile, > knowns=knownLbl, sequence=seqLbl, > assess_methods=(assess.DOPE, assess.GA341)) > > a.starting_model = 1 > # 2do: test task / run ratio? > a.ending_model = nmodPerRun > > a.deviation = atomDeviation > if randMethod=='XYZ': > a.rand_method = randomize.xyz > elif randMethod=='DIHEDRAL': > a.rand_method = randomize.dihedrals > > a.make() i'm stepping thru with a debugger, and it gets to the yield_tasks_unordered() generator in the master. but as i attempt to make the next step it throws an exception that seems to have to do with env.io.atom_files_directory now (in the master/slave version) not getting set correctly? > Running 14 tasks on 2 slaves > /usr/bin/python2.6 /usr/lib/modeller9.10/bin/modslave.py -slave 127.0.1.1:59861:0:BGRESKQY >& seed_align_pp.slave0 > /usr/bin/python2.6 /usr/lib/modeller9.10/bin/modslave.py -slave 127.0.1.1:59861:1:ZCQMRSNQ >& seed_align_pp.slave1 > Identified slave <Slave on localhost> > Identified slave <Slave on localhost> > Traceback (most recent call last): > File "/home/rik/Data/pkg/eclipse/plugins/org.python.pydev.debug_2.2.3.2011100616/pysrc/pydevd.py", line 1267, in <module> > debugger.run(setup['file'], None, None) > File "/home/rik/Data/pkg/eclipse/plugins/org.python.pydev.debug_2.2.3.2011100616/pysrc/pydevd.py", line 1020, in run > pydev_imports.execfile(file, globals, locals) #execute the script > File "/home/rik/Code/eclipse/modeller/src/seed_align_pp.py", line 98, in <module> > doAllAlign() > File "/home/rik/Code/eclipse/modeller/src/seed_align_pp.py", line 91, in doAllAlign > for result in j.yield_tasks_unordered(): > File "/usr/lib/python2.6/dist-packages/modeller/parallel/job.py", line 146, in yield_tasks_unordered > for task in self._finish_all_tasks(): > File "/usr/lib/python2.6/dist-packages/modeller/parallel/job.py", line 164, in _finish_all_tasks > task = self._process_event(obj, s) > File "/usr/lib/python2.6/dist-packages/modeller/parallel/job.py", line 180, in _process_event > task = obj.task_results() > File "/usr/lib/python2.6/dist-packages/modeller/parallel/slave.py", line 61, in task_results > r = self.get_data(allow_heartbeat=True) > File "/usr/lib/python2.6/dist-packages/modeller/parallel/communicator.py", line 89, in get_data > (cmdtype, obj) = self._recv() > File "/usr/lib/python2.6/dist-packages/modeller/parallel/communicator.py", line 130, in _recv > raise RemoteError(obj.exc, self) > modeller.parallel.communicator.RemoteError: <type 'IOError'>: pdbnam_____E> Filename for PDB code not found: 1A04A.pdb > Directories: > Extensions : :.atm:.pdb:.ent:.crd > (Also tried prepending 'pdb', looking for .Z, .gz, .bz2, .7z, .xz, > and trying PDB-style subdirectories - e.g. ab for pdb1abc.ent) > from <Slave on localhost> would you know what i'm doing wrong here? thanks again for your help. Rik

2 1

automodel.max_molpdf recommended values and usage format
by Andrew Voronkov 01 Nov '11

01 Nov '11

Dear Modeller users, how on your opinion modeller objective function correlates with quality of the models. For example I have objective function value 1730 for one modelk and -750 which as I understand reflects free energy. Does it mean that I shouldn t consider the first one at all? Let's say I want to make cutoff -100 for the objective function. In which format it should be done? automodel.max_molpdf -100 is that correct? Sincerely yours, Andrey

2 1