modeller_usage July 2012

modeller_usage@salilab.org

12 participants
15 discussions

Problem with model-segment.py (same model obtained)
by Lipi Thukral 16 Jul '12

16 Jul '12

Dear Modeller users, I would like to model only a segment of my protein, precisely res 76 - 94. I am using model-segment.py script for this purpose. I have looked at the existing example on web (http://salilab.org/modeller/9v1/manual/node22.html). I created alignment file with the same protein. I have no problems running the script but the model obtained is exactly the same as the template which is expected but even the residue range 76-94 is exactly the same. I am pasting my script here, please let me know where am I going wrong? ------------------------------------------------------------------------------ from modeller import * from modeller.automodel import * # Load the automodel class log.verbose() class MyModel(automodel): def select_atoms(self): # Select residues 1 and 2 (PDB numbering) return selection(self.residue_range('76:', '94:')) 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 # Be sure to use 'MyModel' rather than 'automodel' here! a = MyModel(env, alnfile = '5tmg-5tmg.ali', # alignment filename knowns = '5tmgA', # codes of the templates sequence = '5tmg') # code of the target a.starting_model= 1 # index of the first model a.ending_model = 3 # index of the last model # (determines how many models to calculate) a.make() # do homology modeling ------------------------------------------------------------------------------ Many Thanks, Lipi

2 1

Question about compressing Uniprot90
by Dimitrije Jevremovic 11 Jul '12

11 Jul '12

I am trying to compress the Uniprot90 FASTA file database into a binary file using convert() function as provided in examples for the MODELLER. However, my compressed file ends up being even larger than the uncompressed original one. What may be the problem here? Here is the Python code which I a using with MODELLER v9.7. from modeller import * log.verbose() env = environ() sdb = sequence_db(env) sdb.convert(seq_database_file='/gne/research/data/bioinfo/modeller/db/uniprot90', seq_database_format='FASTA', chains_list='ALL', minmax_db_seq_len=[30, 4000], clean_sequences=True, outfile='/gne/research/data/bioinfo/modeller/db/uniprot90_pruned.bin') Kind regards and thank you, Dimitrije Jevremovic -- ----------------- Dimitrije Jevremovic intern at Bioinformatics & Computational Biology Department

2 1

Is the function - forms.upper_bound(..) - supposed to restrain precisely?
by Adhikari, Badri (MU-Student) 11 Jul '12

11 Jul '12

Hello modeller users and experts, I am trying to impose distance restrains in structure prediction python script as explained in the example at http://salilab.org/modeller/manual/node27.html. Following the example I used the forms.gaussian(group, feature, mean, stdev) function. But my predicted model does exactly satisfy the distance restrain. I can observe that adding the distance restrain, HAS added some distance restrain but it looks like modeller is not giving 100% weightage to the distance restrain. I also tried forms.upper_bound(group, feature, mean, stdev) as explained at http://salilab.org/modeller/9v5/manual/node97.html. With this function, I verified that modeller gives more weightage to the distance restrain, but STILL does not precisely satisfy the restrain. Is this what these functions are supposed to do? OR is there any other technique to instruct modeller to absolutely satisfy the distance restrain? Case I: Restrain used: rsr.add(forms.gaussian(group = physical.xy_distance, feature = features.distance(at['CA:2'], at['CA:4']), mean = 7.0, stdev = 0.1)); Distance between CA:2 and CA:4 in the final model: 8.5 Case II: Restrain used: rsr.add(forms.upper_bound(group = physical.xy_distance, feature = features.distance(at['CA:2'], at['CA:4']), mean = 7.0, stdev = 0.1)); Distance between CA:2 and CA:4 in the final model: 8.2 Sincerely, Badri Adhikari CS Graduate Student, University of Missouri-Columbia, Columbia, Missouri BAP54(a)mail.missouri.edu<mailto:BAP54@mail.missouri.edu>

2 1

Re: [modeller_usage] modeller_usage Digest, Vol 11, Issue 73
by Ashish Runthala 08 Jul '12

08 Jul '12

Dear Modellers, As per the ongoing discussions between thomas gaillard and Ben Webb, the two cases i.e. employing HETATM MSE as MET in template(s) leads to a model different than the case, when they are taken as HETATM only. Surprisingly, MODERLLER consideration is wrong in 1 case, as i realized. One template for a CASP9 target was an enzyme. So its structural mutual orientation of domains was more accurate when MSE was considered as compared to the case, when such HETATM MSE records were forced as MET. Therefore, while modelling a target and also knowing the available hetatom environment, consideration of MSE record makes sense. Whereas on the other hand, modelling an unknown target, the MODELLER's forceful mapping of MSE to MET is acceptable. In the former case, CG-SD-CE bond length, angle and topology is disturbed when it is considered as MET. Therefore, the distance map based model accuracy was found higher in first case and the same as inferior in the latter case. Therefore, I believe the best solution should be to put an option in Model-multiple or any such modelling script as env_io_hetatm_default_statement=True if such HETATM records need to be considered as standard amino acids or, env_io_hetatm_default_statement=False, if such HETATM records need to be considered as such with no alteration of their bond lengths, angles and inter-atomic orientation/switchings. This would help us to improve the accuracy of model's Hydrogen bonding network for sure. Regards, Ashish Ashish Runthala, Lecturer, Structural Biology Cell, Biological Sciences Group, BITS, Pilani Rajasthan, INDIA Message: 1 Date: Mon, 25 Jun 2012 15:41:56 -0700 From: Modeller Caretaker <modeller-care(a)salilab.org> To: Thomas Gaillard <thomas.gaillard(a)polytechnique.edu> Cc: modeller_usage(a)salilab.org Subject: Re: [modeller_usage] HETATM MSE records Message-ID: <4FE8E934.7020909(a)salilab.org> Content-Type: text/plain; charset=UTF-8; format=flowed On 6/25/12 10:33 AM, Thomas Gaillard wrote: > I have noticed that some HETATM records are read by Modeller and > automatically handled as corresponding standard residues (for example > MSE -> MET) Actually MSE -> MET is the only such translation done. > I would like to know if you could suggest a way to disable this > behaviour and really ignore all HETATM records. Unfortunately there isn't any way to do that - it is hardcoded into the PDB reader. Why do you want to treat MSE as HETATMs anyway? There may be a better way to achieve your goal. Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage Oh yes, I'd forgotten about the less common MEX and ABU cases. Yes, those are also mapped, to CYS. > Because when MODELLER is integrated with other programs, some do not > make the same choice of automatically treating HETATM MSE as ATOM MET > and this can lead to confusion. I think this behaviour should not be > hardcoded. Makes sense. I'll add an option to turn this off in the next release. > Additionnally, I think your mapping of MSE SE atom to MET SD atom is slightly > incorrect in the eyes of the PDB standard: Thanks for catching that - I will fix.

1 0

(no subject)
by harpreet singh 05 Jul '12

05 Jul '12

http://examwings.com/ppuds.html?jkaa=wepcukk

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

modeller_usage July 2012