September 2010 - modeller_usage

aligning DOPE profiles between model and template
by dmitry lyumkis 30 Sep '10

30 Sep '10

Hi all, I am currently using Modeller to predict a domain structure in a protein of interest. The alignment that was calculated contains gaps in the model, as compared to the template. When I finish modeling, I would like to overlay the DOPE profiles between model and template, but they contain a different number of residues, and therefore graphical overlays do not make sense. I can get around this issue by writing some script to get rid of extra residues or set their DOPE energies to some constant value, but I was more interested in finding out how this is normally done in Modeller? Is there some function that would compare / align DOPE profiles for the calculated models based on the template? Dmitry Lyumkis Automated Molecular Imaging Group, Department of Cell Biology The Scripps Research Institute 10550 North Torrey Pines Road c/o Dmitry Lyumkis, CB129 La Jolla, CA 92037 tel: (858) 784-9208; cell: (408) 772-1792; dlyumkis(a)scripps.edu; http://ami.scripps.edu

2 1

DOPE score for multi chain protein
by Omar Piña 28 Sep '10

28 Sep '10

Hi, I am modelling a tetrameric protein (ion channel). I want to evaluate the models, but I have read that the DOPE assesment is not good for multi chain proteins (http://www.salilab.org/modeller/9v6/manual/node242.html) Then, how is the best way to evaluate them? Each chain should be evaluated independently? Should I take only the best molpdf score? In the case of taking only the best molpdf, are there no way for an optimization with DOPE like the Tutorial says? Thanks, Omar

2 1

dope_loopmodel and TER
by Lionel Ducassou 27 Sep '10

27 Sep '10

Hi, I would like to use the dope_loopmodel but an error occurs during the job with the message : "Atom 3928 (type 77) has GB/SA radius 0.0000". In my pdb file generated by modeller the atom 3928 refers to TER (not really an atom...). May I change the solv.lib file adding a new line for TER atom ? Thanks, Lionel Ducassou

2 1

Using part of a multi protein complex
by Darren Leneghan 14 Sep '10

14 Sep '10

Hi there, I'm trying to generate a model which uses the crystal structure of a homologous protein which is in complex with its interacting partners in a small complex. The crystal structure PDB files for the individual parts of the complex, as well as for the whole complex are available. How would I go about substituting my protein into the model?

2 1

Modeller 9v8 installation on lucid
by Joanne Martin 09 Sep '10

09 Sep '10

Hello I am trying to install Modeller 9v8 on my Ubuntu 10.04 LTS - the Lucid Lynx machine. I have downloaded the Linux 32-bit Debian/Ubunto package and have also received my license. However when i try installing it using the command - *sudo env KEY_MODELLER9v8=xx dpkg -i modeller_9v8-1_i386.deb* (xx being my license number), I get the following output - *(Reading database ... 130902 files and directories currently installed.) Unpacking modeller (from modeller_9v8-1_i386.deb) ... Setting up modeller (9v8-1) .. *nothing more happens?* *I then have to go to the usr/lib/modeller directory and execute the install manually, with the following output- *Installation of MODELLER 9v8 This script will install MODELLER 9v8 into a specified directory for which you have read/write permissions. To accept the default answers indicated in [...], press <Enter> only. ------------------------------------------------------------------------ The currently supported architectures are as follows: 1) Linux x86 PC (e.g. RedHat, SuSe). 2) SUN Inc. Solaris workstation. 3) DEC Inc. Alpha OSF/1 workstation. 4) IBM AIX OS. 5) Itanium 2 box (Linux). 6) x86_64 (Opteron/EM64T) box (Linux). 7) Alternative x86 Linux binary (e.g. for FreeBSD). Select the type of your computer from the list above [1]: *None of the architectures are suitable for my 32 bit Ubunto 10.04 lucid machine??? Can you please tell me how to overcome this, Regards, Joanne

2 1

Re: [modeller_usage] problem of strand/sheet
by Thomas Evangelidis 01 Sep '10

01 Sep '10

In my case although most of the hydrogen bonds imposed by secondary_structure.sheet() have been satisfied, the 3 strands do not adapt a beta sheet conformation. I tried to optimize only the 3 antiparallel beta strands by adapting the examples/automodel/model-segment.py script. After setting physical.values(default=1.0, h_bond=10.0)I get these violations in my log file: # RESTRAINT_GROUP NUM NUMVI NUMVP RMS_1 RMS_2 MOL.PDF S_i ------------------------------------------------------------------------------------------------------ 8 H-bonding potential : 20 0 5 0.103 0.103 31.778 10.000 What type of restraints does secondary_structure.strand() impose to scale them up? The manual says phi and psi binormal restraints only, but in my log file I get this: 11 Mainchain Phi dihedral restraints : 0 0 0 0.000 0.000 0.0000 1.000 12 Mainchain Psi dihedral restraints : 0 0 0 0.000 0.000 0.0000 1.000 I also need some help interpreting the restraint files: 1. What do the different columns (apparently not the first 2) represent in the target.V99990001 file? 2. For what pattern should I search in the target.rsr file with respect to the beta sheet? Surprisingly I cant find any column having a Group value 8 (the 4th column) which corresponds to the H-bonding potential imposed by secondary_structure.sheet().

2 1

R: modelling discarding part of a template
by Thomas Evangelidis 01 Sep '10

01 Sep '10

> You have to either keep the complete template sequence in the PDB file, > or modify the template PDB to remove the residues you don't want. An > alignment that achieves what you want without modifying the PDB is: > > target: AAAAAAAAAAA-------AAAAAA > template 1: BBBB-------BBBBBBBBBBBBB > template 2: ZZZZZZZZZZZ-------ZZZZZZ > > Ben Webb, Modeller Caretaker > > Is it algorithmically difficult to add a function that will select a range of residue in one template and Modeller will exclude them without touching the alignment? When having multiple templates and several unwanted regions this procedure becomes very tedious.

2 2