- modeller_usage - salilab.org

How does Modeller handle multiple structures?
by Zoe Katsimitsoulia 10 Feb '06

10 Feb '06

Hello everyone, I am very new to modelling and working a target sequence that has just less than 27% sequence identity with another 10 proteins that have crystal structures. My target and templates are all myosin proteins, but from 2 different classes, since there are no crystal structures available for my target in its same class. I have 2 concerns/queries: 1. How will modeller handle the 10 template sequences (which themselves show a significant amount of variation at certain key positions). More specifically, a particular area of my seq alignment may read like this: template 1 xxxxxGxxxxxxxx template 2 xxxxxTxxxxxxxx template 3 xxxxxGxxxxxxxx template 4 xxxxxGxxxxxxxx template 5 xxxxxGxxxxxxxx template 6 xxxxxGxxxxxxxx template 7 xxxxxGxxxxxxxx target xxxxxTxxxxxxxx In simple terms, would modeller in this case use the coord of T in the template 2 sequence because it matches the target at that location, or will it use G because it is predominant in the majority of the structures? 2. The 10 template structures I am using are themselves in different states, for example, partially closed conformation, open, transition, inhibitor bound, etc. Is this beneficial to building my model or more detrimental? I am assuming the former based on the logic that the more info I am giving to modeller, the better my model should be. However, perhaps there is something about the way modeller works which I have not grasped that means its actually doing more harm adding the structures in different states? Thanks very much in advance, -Zoe

3 2

Protocol of modeling a membrane protein
by Qing Zhang 08 Feb '06

08 Feb '06

Dear Modeller users: I just started using Modeller (version 8v1) and would like to find a good protocol of modeling membrane proteins that contain helices, beta-sheets, loops, and long N and C-terminal tails. By reading Modeller manual and some papers, I have a rough plan like this: 1. Find the closest sequences in PDB using Blast [done]. 2. Do sequence alignment with the best PDB structure using ClustalW [done with a score of only 15] 3. Build 10 models in Modeller with 3 repeats of very thorough VTFM optimization and MD optimization until max_molpdf <= 1E6 [done] 4. Pick the best model based on the quality of helices and beta-sheets only since loops have not been very optimized [I do not know much about verifying protein structure quality; it seems WHAT_CHECK is a good program to use; any suggestions?] 5. Optimize loops in Modeller for the above best model and obtain several loop models 6. Pick the best loop model with quality check 7. Molecular dynamics simulations? Any suggestions on this? Thanks in advance. Sincerely, Qing ====================================== Qing Zhang, Ph.D. Research Associate Department of Molecular Biology, MB-5 The Scripps Research Institute 10550 North Torrey Pines Road La Jolla, CA 92037-1000 Tel: (858) 784-2333 Fax: (858) 784-2860 Cell: (917) 509-3182 Email: qzhang(a)scripps.edu Website: www.qingzhang.info ======================================

1 0

Add loops to a protein
by Mingfeng Yang 07 Feb '06

07 Feb '06

Dear Modeller users, I have a protein, its full sequence and most part of its structure have been known, except 4 segments of loops (6-15 residues long). Can I use modeller package to add the loops to the known core of the protein? How? Where should I begin from? I know that I definitely should delve into the tutorial or manual first, try different things and then ask question here. However I just heard this modeling package minutes ago, and just registered, downloaded, and installed it. I am in a rush to get the last step of my current project done. Hope somebody can give me some hints. Thank you very very much! Cheers, Mingfeng Yang

2 1

(no subject)
by Anna Kaczynska 07 Feb '06

07 Feb '06

Hi, I am trying to find out what I should apply in order to receive a protein structure with the lowest objective function of protein structure. Please, can you let me know is you will have any suggestions. Thank you Anna -- Anna Kaczynska (917)834-2291 (cell)

1 0

dna modelling
by rathankar rao 06 Feb '06

06 Feb '06

hi all i need an important information on modelling. i have tried modelling protein sequences but now that i would like to know whether dna could also be modelled using modeller and if yes how can it be done. are there any new files which i have to download or import it from the package itself. plz reply. with regards rathankar --------------------------------- Relax. Yahoo! Mail virus scanning helps detect nasty viruses!

2 1

Help!
by jo hanna 06 Feb '06

06 Feb '06

Hello Sorry for the vast number of posts recently, I think I have finally figured out most of problems. My main question, which I am desperate for suggestions to, is how to read in an existing pdb (x-ray crystal structure) and add the missing residue atoms and the loop region, with minimal disruption to the existing sidechains and backbone. I appreciate that when adding sidechains there will need to be some movement of surrrounding sidechains but I am seeing quite a lot of movement all over my model. I guess however this is possibly because some of the sidechains are actually within quite closely packed regions. I have tried to just refine the loop region and the missing sidechains. and then loop model the missing loop. this works fine for the loop region but the missing sidechain atoms are added with no refinement at all and clash with other added atoms and exiisting sidechains. I have also tried the following script: # Homology modelling by the automodel class from modeller.automodel import * # Load the automodel class log.verbose() env = environ() # directories for input atom files env.io.atom_files_directory = './:../atom_files' a = loopmodel(env, alnfile = 'phe-1phe.ali', # alignment filename knowns = '1phe', # codes of the templates sequence = 'Phe') # code of the target a.starting_model= 1 # index of the first model a.ending_model = 1 # index of the last model # (determines how many models to calculate) a.md_level = None # No refinement of model a.loop.starting_model = 1 # First loop model a.loop.ending_model = 50 # Last loop model a.loop.md_level = refine.fast # Loop model refinement level a.make() # do homology modelling which adds the sidechains, optimises them, adds the loop and then optimises that via the loop modelling routine. However this moves all the other atoms also. Could someone please give me some advice, it would be much apprecaited! J

1 0

Help!
by jo hanna 05 Feb '06

05 Feb '06

Hi I have added missing loop residues to my existing pdb file using the refining only part of the model method (I want to keep the rest of the coordinates unchanged) and then loop optimisation method to optimise the added residues. During this process atoms missing from other residues within the protein chain are also added. My question is how is this done? I have not found any reference to this in the manual. After these atoms are added do they undergo any energy minimisation? I would really appreciate any help with this question as I am confused why after energy minimising these residues myself in another program and then reanalysing my model with the evaluate_model.py script, the dope score I obtain is significantly worse than for the original model from modeller. Many thanks J

1 0

HELP!! Refine part of model
by Matt wagner 03 Feb '06

03 Feb '06

I am refining a segment of my model using the refine part of model script ( http://salilab.org/modeller/manual/node28.html) Everything works fine, but it creates so many atom clashes that when I do an energy minimization on the structure, that the backbone of my protein snaps. Is there another way to reduce these atom clahses? I also did a test in which I turned on bond interactions, by setting env.edat.nonbonded_sel_atoms = 1; but there are still too many atom clashes. Below is a copy of my script: # Addition of restraints to the default ones from modeller.automodel import * # Load the automodel class log.verbose() env = environ() # directories for input atom files env.io.atom_files_directory = './:../atom_files' class mymodel(automodel): def special_restraints(self, aln): rsr = self.restraints # Add some restraints from a file: # rsr.append(file='my_rsrs1.rsr') # Restrain the specified CA-CA distance to 10 angstroms (st. dev.=0.1) # Use a harmonic potential and X-Y distance group. rsr.add(atom_ids=('NZ:12', 'NZ:23'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 11.0, 0.01)) rsr.add(atom_ids=('NZ:40', 'NZ:45'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 9.5, 0.01)) rsr.add(atom_ids=('NZ:88', 'NZ:94'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 10.4, 0.01)) rsr.add(atom_ids=('NZ:94', 'NZ:96'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 10.4, 0.01)) rsr.add(atom_ids=('NZ:96', 'NZ:106'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 11.4, 0.01)) rsr.add(atom_ids=('NZ:106', 'NZ:107'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 6.5, 0.01)) rsr.add(atom_ids=('NZ:133', 'NZ:140'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 11.4, 0.01)) rsr.add(atom_ids=('NZ:206', 'NZ:208'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 3.8, 0.01)) rsr.add(atom_ids=('NZ:226', 'NZ:238'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 34.2, 0.01)) rsr.add(atom_ids=('NZ:238', 'NZ:239'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 6.9, 0.01)) rsr.add(atom_ids=('N:1', 'NZ:96'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 11.4, 0.01)) rsr.add(atom_ids=('NZ:23', 'NZ:59'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 7.5, 0.01)) rsr.add(atom_ids=('NZ:94', 'NZ:208'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 4.5, 0.01)) rsr.add(atom_ids=('NZ:96', 'NZ:195'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 10.3, 0.01)) rsr.add(atom_ids=('NZ:96', 'NZ:208'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 7.2, 0.01)) rsr.add(atom_ids=('NZ:96', 'NZ:226'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 11.4, 0.01)) rsr.add(atom_ids=('NZ:118', 'NZ:140'),restraint_parameters=(3, 1, 1, 27, 2, 2, 0, 11.4, 0.01)) rsr.make(aln, restraint_type='ALPHA', residue_ids=('24', '39'), spline_on_site=False) rsr.make(aln, restraint_type='ALPHA', residue_ids=('48', '82'), spline_on_site=False) rsr.make(aln, restraint_type='ALPHA', residue_ids=('100', '129'), spline_on_site=False) rsr.make(aln, restraint_type='ALPHA', residue_ids=('134', '164'), spline_on_site=False) def select_atoms(self): self.pick_atoms(selection_segment=('87:', '113:'), selection_search='segment', pick_atoms_set=1, res_types='all', atom_types='all', selection_from='all', selection_status='initialize') env = environ() # directories for input atom files env.io.atom_files_directory = './:../atom_files' # selected atoms do not feel the neighborhood env.edat.nonbonded_sel_atoms = 2 a = mymodel(env, alnfile = 'TvLDH-mult.ali', knowns = ('1_86','114_243'), sequence = 'TvLDH') # code of the target a.starting_model= 1 # index of the first model a.ending_model = 1 # index of the last model # (determines how many models to calculate) a.make() # do homology modelling

1 0

loop modelling DOPE score
by jo hanna 03 Feb '06

03 Feb '06

Hi I have done some loop modelling on an existing model. I am confused as although I have selected the BLmodel with the lowest modeller objective function and this BLmodel also has a lower dope score than that of the original model, when I plot the *.profile output from DOPE, the loop region I have optimised shows a higher energy than it was original model. I am confused by this, have i missed something? Thanks Jo

1 0

Re: [modeller_usage] helix alignment with gap
by Guillaume LETELLIER 02 Feb '06

02 Feb '06

> As a general argument, just because an alignment has more true matches > does not mean that it is correct. There are penalties associated with > those two gaps you introduced that offset the gain from exact residue > matching. Not knowing exactly which sequence you are modeling, I > submitted P11229 (ACM1_HUMAN) to the SUPERFAMILY server and the > alignment come back with no gaps in that region. Take a look at: > > http://www.supfam.org/SUPERFAMILY/cgi-bin/gene.cgi?genome=up;seqid=P11229 > > and follow the "Align" link. Obviously, the same can be done for any > protein for which you know the SwissProt number by simply inserting it > at the end of the link above. Not to get into prolonged argument here, > I'll just say that I would trust SUPERFAMILY-based alignment over any > human-derived alignment unless there is a very good reason to believe > that the server alignment is wrong. And breaking a long helix to make > few extra positive matches usually doesn't qualify as a good reason, > especially because it forces you to make a compensatory two-residue > gap later in the model. But don't take my word for it, try modeling > the protein with and without those gaps and see which model evaluates > better. > > Hope this helps, > > Mensur Here there are strong argument that kinda end the gap debate ... Thanks for preventing me to lost my time and make huge mistakes. (the gapped models came with objective function score at least double of the 'classical' ones) At least have I learned to play with the alpha constraint of modeller ;o) . Regards Guillaume

1 0