- modeller_usage - salilab.org

Using a Modeller model as input to repeat the process
by Zoe Katsimitsoulia 15 Mar '06

15 Mar '06

Hello all, I am trying to model a protein which has many structures solved in a particular state, and only 1 structure solved in my desired state. I was wondering if it is possible to do the following: 1. Generate a model with Modeller based on the larger number of solved structures (yes, obviously) 2. Run modeller again, this time only with the 1 crystal structure available in my state as a template, but ALSO somehow incorporate my model from step 1 above, without using the model as a second template, because the states are too dissimilar I feel. The reason I do not want to generate a structure solely using the 1 crystal structure is because there is very low sequence identity (just under 30). So, effectively i'm assuming step 1 should give me an accurate enough model of my protein in state A, and i want to use this model plus the only one available structure to get it to state B. Does my theory make sense and is it possible? Any help is much appreciated. Thanks, -Zoe -- Zoe Katsimitsoulia Division of Mathematical Biology National Institute for Medical Research The Ridgeway, Mill Hill, London NW7 1AA zkatsim(a)nimr.mrc.ac.uk +44 (0)20 8816 2507(phone) +44 (0)20 8816 2460 (fax)

1 0

hydrogen atom name
by Zhiqiang Ye 14 Mar '06

14 Mar '06

Dear Sir, I have question about the hydrogen atom name after building the all-atom model. After building the all-atom model, the hydrogen name "HA" means hydrogen bonded to C-alpha, but GLY has two HA, one is 1HA, the other is 2HA. Which one is correponding to the sidechain in other residue types? I checked several GLY in my model, I find always it is 2HA that is corresponding to the side chain of other residue type. Is this the standard criteria, or just modeller's choice? Thank you very much! -- Zhiqiang Ye PHD candidate

2 2

Report of final energy
by 서주현 14 Mar '06

14 Mar '06

2 1

Dope Vs PROSA2003
by jo hanna 14 Mar '06

14 Mar '06

Hi If am trying to model loops within my protein model that seem to correspond to regions of high energy in both DOPE and PROSA analysis. why should it be that aftre modelling the loop with a lower DOPE energy shoud actaully increase in the region of loop modelling with reference to the original model and PROSA analysis? Any advice much appreciated Do people generally prefer DOPE or use PROSA and DOPE to complement each other? Thanks Jo

1 0

any .log indication of .md_level results?
by Mark Daniels 13 Mar '06

13 Mar '06

Hello all- Is there any way in which one can generate output from, or verify the running of, the .md_level routine in Modeller8v2? I do not see any indication in the .log that a molecular dynamics simulated- annealing routine was applied to any model generated when I specify " automodel.md_level = refine.very_fast ". thanks in advance- Mark

2 1

File name control problem
by 서주현 13 Mar '06

13 Mar '06

2 1

Sequence consensus
by Cassiano 13 Mar '06

13 Mar '06

Hi everybody! I have a question about sequence alignment (i known that it isn't the subject of this forum, but i'll be very glad for any help). From a alignment, i intend to extract the consensus sequence differently. I keep only the aa that are really conserved (and probably structurally important). The others, I substituted by the aa G (for example: GGGDGGGRKGGGHEGG, where the aa D,R,K,H and E are very conserved in the protein's alignment). My aim is to modeler the consensus folding to the classes of protein in my study, emphasizing only the functionally important aa. I don't known if I really have a useful structural information or not. If not, i'll be very glad for any advice. Thanks, Cassiano Carromeu.

1 0

model.mutate problem
by Zhiqiang Ye 12 Mar '06

12 Mar '06

Hi, all I have a homodimer protein structure, it has two chains: A and B, both of whcih start from 2. I want to get the structure of the mutant K2S. So I modified the provided example: mutate_model.py to do this job, the mainly modified part is: (I want to mutate both chains) mdl.pick_atoms(aln, selection_segment=("2:A", "2:A"), selection_status='INITIALIZE', selection_search='SEGMENT', selection_from='ALL') mdl.pick_atoms(aln, selection_segment=("2:B", "2:B"), selection_status='ADD', selection_search='SEGMENT', selection_from='ALL') But in the output model, the two chains are connected as one single chain, and the residue number has changed start from 1, but I want the mutant keep the original chain ID and number. How can I get this kind of results? Thank you very much! Regards! -- Zhiqiang Ye PHD candidate

2 2

Loop
by BiplabBose＠bits-pilani.ac.in 11 Mar '06

11 Mar '06

Hi I have modeled a dimeric protein using Modeller. Now I want to discard a small loop (~8-9 amino acids) of this initial model and want to model that region using some other PDB file. Is it possible by Modeller? I hope, yes. But how? Expecting your help With regards Biplab BITS, Pilani

1 0

using BLAST to find templates for modelling
by Nick Fusseder 09 Mar '06

09 Mar '06

Hi all, I have been dwelling over the following problems and simply do not know how to continue. Any help would be greatly appreciated. My scenario: I am running a BLAST search against PDB sequences in order to find suitable templates to model my target protein. I have all PDB structures locally installed and the atom_files variable pointing to this directory in my .py scripts. I basically have 2 questions: 1. From what I understand the modlib/CHAINS_all.seq is used to store the sequences for the PDB structures. Now if I want to use a novel PDB structure as a template which is not included in this list, will I have to manually create the chain sequence in PIR format (with commands/make_chains.py) and append it to the modlib/CHAINS_all.seq file? ALso is there a way to create a CHAIN_all.seq file from all PDB structure files in one go? 2. Since I am already running a BLAST search to find my template structures, I won't be using search.py. In order to create the .ali alignment file which is neccessary for the computation of the final structure prediction of my target sequence, i will need to run malign.py. The problem is, malign takes a .seg file as input. I have not yet found an automated method to generate such a .seg file from a target sequence and template structures found by BLAST. Will I have to create the .seg file manually? Maybe I am just missing the point and the solution to all my problems is in fact really easy. I guess my main problem boils down to which files can be created by modeller for me and which files need to be generated or edited by hand. I would greatly appreciate any help. Regards, nick -- Echte DSL-Flatrate dauerhaft f�r 0,- Euro*! "Feel free" mit GMX DSL! http://www.gmx.net/de/go/dsl

2 1