- modeller_usage - salilab.org

modeller 9.2: spaghetti effect with multiple templates
by Raik Gruenberg 10 Aug '07

10 Aug '07

Hi modellers, already a while ago I have noticed that modeller tends to produce what I would call spaghetti proteins (containing knots and other fancy structures) when it is supplied with multiple templates that are far from the target (35 % or less) and also quite diverse among each other. However, Modeller 9 seems to suffer more from this effect than the previous versions: One of our test cases uses 10 templates with between 86% and 31% sequence similarity to the target. I know that this is not the optimal strategy (especially after reading this paper: http://bioinformatics.oxfordjournals.org/cgi/content/abstract/btm262v1 ) but modeller 8.2 still performed quite well and all 10 models had good topology (close to the 86% template) and secondary structure. However, I now re-run our test with modeller 9.1, using the identical alignment (T-Coffee) and the same input file (below) and the result looks much worse. Most among the 10 generated models show a messed up (swaped) topology, many have knots and the different models are generally more diverse. So to what extend differs the multiple-template treatment of Modeller 9 from previous versions? Do I need to use different parameters for multi-template modeling in the new version? Or is the general default modeling protocol different now? So here is our (old-style) input file: INCLUDE SET OUTPUT_CONTROL = 1 1 1 1 1 SET ALNFILE = 'final.pir_aln' SET KNOWNS = '1K2H_A' '1M31_A' '1J55_A' '1KSO_A' '1MQ1_A' '1MWN_A' '1MHO_' '1NSH_A' '1K8U_A' '1DT7_A' SET SEQUENCE = 'target' # code of the target SET ATOM_FILES_DIRECTORY = '../templates/modeller/' SET STARTING_MODEL= 1 # index of the first model SET ENDING_MODEL = 10 # index of the last model CALL ROUTINE = 'model' I am attaching a typical result of modeller 8 and one of modeller 9 for comparison. Thanks in advance for any comments! Best regards, Raik

2 4

making a patch work
by Starr Hazard 10 Aug '07

10 Aug '07

Folks, I have created a RESI entry in top_heav.lib for a ligand of interest to me. When run without attempting to patch anything the RESI is accepted just fine and models are made with the appropriate ligand in an appropriate location. In reality the ligand forms a covalent bond with the protein. This is a Schiff Base linkage. My RESI is an aldehyde. The covalent bond is to the NZ of a lysine. I am deleting hydrogens from the LYS and the aldehyde oxygen from from the ligand to make the bond of interest to me. I created a PRES entry based on the disulfide example. PRES ELAZ +1.0000 ATOM 1:CE CT2 -0.10000 ATOM 1:NZ NN3 1.0000 ! trying to make a protonated Schiff base ATOM 2:C15 CN2 -0.08000 ATOM 2:C14 CT1 -0.10000 DELE ATOM 1:HZ3 DELE ATOM 1:HZ2 DELE ATOM 2:O32 BOND 1:NZ 2:C15 ANGL 1:CE 1:NZ 2:C15 1:NZ 2:C15 2:C14 DIHE 1:CD 1:CE 1:NZ 2:C15 1:NZ 2:C15 2:C14 2:C13 1:CE 1:NZ 2:C15 2:C14 IC 1:CD 1:CE 1:NZ 2:C15 0.0000 0.0000 92.5000 0.0000 0.0000 IC 1:CE 1:NZ 2:C15 2:C14 0.0000 0.0000 183.0000 0.0000 0.0000 IC 1:NZ 2:C15 2:C14 2:C13 0.0000 0.0000 177.0000 0.0000 0.0000 When I perform a test run of the code patch.py from the on-line manual (where I have added "env.io.hetatm=true" to read the ligand). The program appears to run but the log file shows numerous entries of this sort r_stere_607W > Cannot find parms in parm file Then remarks indicate that it has "added" some parms; so it gets past some instances where explicit information has not been supplied. It gets to a particular dihedral involving an atom that is supposed to be deleted by the patch ( see line above- 'DELE ATOM 2:O32' is the line in top.heav.lib; this is the aldehyde oxygen) and stops. The final line reads simply: FATAL error: assertion nbndpf < mbndpf failed in r_stereo.F90, line 608 At this point that error means absolutely nothing to me. The entries in the parm.lib file are often redundant for many entries. Are the columns described somewhere? Do I need to have explicit entries in the parm file for an atom that I hope to delete? I have not deciphered how to designate bond order in either the RESI entry or the PRES entry lines. How does one specify bond order? Starr

2 2

regrding superpose.py
by vidhya sankar 10 Aug '07

10 Aug '07

I have got a RMS and DRMS value of after super position 2.0482 and 1.5631 respectively.is it reliable value? .for 53.85% sequence identity i am expecting your reply --------------------------------- Shape Yahoo! in your own image. Join our Network Research Panel today!

2 1

regarding DOPE score
by vidhya sankar 10 Aug '07

10 Aug '07

i have got a DOPE score of -41275.816406 is it reliable value ? .how can i come to conclusions for my 3Dmodel .how can i analyse DOPE SCORE --------------------------------- Luggage? GPS? Comic books? Check out fitting gifts for grads at Yahoo! Search.

2 1

ref: SUPERPOSE.PY
by vidhya sankar 10 Aug '07

10 Aug '07

I have got a RMS and DRMS value of after super position 2.0482 and 1.5631 respectively.is it reliable value? .for 53.85% sequence identity i am expecting your reply --------------------------------- Park yourself in front of a world of choices in alternative vehicles. Visit the Yahoo! Auto Green Center.

1 0

DOPE
by Ester Fusté Domínguez 09 Aug '07

09 Aug '07

How can I see the DOPE score profile? I don't know how to use *assess_dope*command. Thanks, Ester

2 1

code and logic question
by Christian Meesters 09 Aug '07

09 Aug '07

Hi, I have a question concerning code semantics. To me it seems to address the heart of similar questions I've read on the list, too. If a derived class (see http://www.salilab.org/modeller/manual/node27.html for instance) has a method defining additional restraints, how, having no conventional __init__-method of the base class called, do I know in advance that a method is called? I never saw something like a.special_restraints(aln) in the examples. (BTW, where does 'aln' come from? special_restraints() requires this parameter, but it does not get defined in the code.) My scripts follow the example code, but I've also run into problems occasionally, not knowing how I can actually play with the examples sticking to pythonic coding. So, my question(s) can be summarized as: Is there any node in the manual describing the MODELLER-Python interface, particularly a list of methods invoked upon deriving a class? If not, would you accept this as a 'documentation request'? (If this is documented somewhere already, please excuse my ignorance.) Thanks Christian

2 1

Error Messages
by Michael Paterson 08 Aug '07

08 Aug '07

I have encountered a persistent error message when using Modeller for generating initial structures with automatic loop modelling. Please find attached a copy of the alignment file etc The error message is - Traceback (most recent call last): File "valid.py", line 25, in ? a.make() # do homology modeling File "/usr/lib/modeller9v1/modlib/modeller/automodel/loopmodel.py", line 31, i n make automodel.make(self, exit_stage) File "/usr/lib/modeller9v1/modlib/modeller/automodel/automodel.py", line 119, in make self.multiple_models(atmsel) File "/usr/lib/modeller9v1/modlib/modeller/automodel/loopmodel.py", line 48, i n multiple_models self.build_seq(filename, num) File "/usr/lib/modeller9v1/modlib/modeller/automodel/loopmodel.py", line 115, in build_seq atmsel = self._check_select_loop_atoms() File "/usr/lib/modeller9v1/modlib/modeller/automodel/loopmodel.py", line 278, in _check_select_loop_atoms atmsel = self.select_loop_atoms() File "/usr/lib/modeller9v1/modlib/modeller/automodel/loopmodel.py", line 315, in select_loop_atoms raise ModellerError, \ _modeller.error: No loops detected for refinement: you must redefine select_loop _atoms Does this indicate that there are no gaps in the initial model that can be refined, automatically, by loop modelling? Regards Mike Paterson

2 1

Modeller 9v2 release
by Modeller Caretaker 08 Aug '07

08 Aug '07

The new version of Modeller, 9v2, is now available for download! Please see the download page at http://salilab.org/modeller/ for more information. If you have a license key for Modeller 8 or 9v1, there is no need to reregister for Modeller 9v2 - the same license key will work. (It won't do any harm to reregister if you want to, though!) 9v2 is primarily a bugfix release relative to the last public release (9v1). Major user-visible changes include: # Python 2.5 support "out of the box" on Windows and Linux. # Minor changes to the arguments for the alignment.salign(), alignment.compare_structures(), profile.scan(), and restraints.spline() functions. # Switch from CHARMM to PDB naming for nucleic acids. # automodel no longer randomizes starting coordinates for BLK residues by default. # Change in default value of automodel.max_molpdf. See the Modeller manual for a full change log: http://salilab.org/modeller/9v2/manual/node38.html If you encounter bugs in Modeller 9v2, please see http://salilab.org/modeller/9v2/manual/node10.html for information on how to report them. 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

Re: [modeller_usage] loop refinement
by Agnieszka Szarecka 08 Aug '07

08 Aug '07

Dear Modeller users, I'm refining a conformation of a 6-residue long loop and use the script: # Loop refinement of an existing model from modeller.automodel import * log.verbose() env = environ() # directories for input atom files env.io.atom_files_directory = './' # Create a new class based on 'loopmodel' so that we can redefine # select_loop_atoms (necessary) class myloop(loopmodel): # This routine picks the residues to be refined by loop modeling def select_loop_atoms(self): self.pick_atoms(selection_segment=('200:A', '205:A'), selection_status='INITIALIZE') m = myloop(env, inimodel='dimer_Cloop_STR_fakeprot_full.pdb', # initial model of the target sequence='dimer_Cloop_STR_fakeprot_full') # code of the target m.loop.starting_model= 1 # index of the first loop model m.loop.ending_model = 500 # index of the last loop model m.loop.md_level = refine.slow_large # loop refinement method m.make() Problem 1: Is is possible in Modeller to fully optimize the specified loop but also to allow a little bit of "adjustment freedom" to the loop end residues (here 199 and 206)? If so, how should I modify the script? Problem 2: Could anyone advise on the m.loop.md_level? How should I judge what's best for my loop? tons of thanks for help and comments Agnieszka

3 2