- modeller_usage - salilab.org

(no subject)
by Michael Accardi 16 May '10

16 May '10

1 0

Re: [modeller_usage] modeller_usage Digest, Vol 9, Issue 74
by Ashish Runthala 11 May '10

11 May '10

As Webb has told. Follow that and you may use these to get good models. 1. Generate maximum possible model number. 2. Try to look for multiple templates for your target sequence. 3. Try to see the alignment file, if your model lacks the conserved sequence information somewhere resulting in complete loss of fold, locally at some place. 4. Try to assess models with all the means provided and then select the best model. Don't rely on any single assessment test. 5. Refinement is not of any usage if you are trapped in local minima earlier in the Potential Energy Surface topological curve. Good Luck, Ashish ----- Original Message ----- From: "modeller usage-request" <modeller_usage-request(a)salilab.org> To: "modeller usage" <modeller_usage(a)salilab.org> Sent: Tuesday, May 11, 2010 7:37:36 AM GMT +05:30 Chennai, Kolkata, Mumbai, New Delhi Subject: modeller_usage Digest, Vol 9, Issue 74 Send modeller_usage mailing list submissions to modeller_usage(a)salilab.org To subscribe or unsubscribe via the World Wide Web, visit https://salilab.org/mailman/listinfo/modeller_usage or, via email, send a message with subject or body 'help' to modeller_usage-request(a)salilab.org You can reach the person managing the list at modeller_usage-owner(a)salilab.org When replying, please edit your Subject line so it is more specific than "Re: Contents of modeller_usage digest..." Today's Topics: 1. Re: How to optimize a Homology model. (Modeller Caretaker) 2. How to generate Homology models using a computer farm (Caio Veloso) 3. Re: How to generate Homology models using a computer farm (Modeller Caretaker) 4. how to extract the part of structure from a complex ? (Yue) ---------------------------------------------------------------------- Message: 1 Date: Mon, 10 May 2010 07:46:44 -0700 From: Modeller Caretaker <modeller-care(a)salilab.org> Subject: Re: [modeller_usage] How to optimize a Homology model. To: Harish Reddy <harisredy(a)gmail.com> Cc: modeller_usage(a)salilab.org Message-ID: <4BE81C54.9020405(a)salilab.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed On 5/7/10 7:42 PM, Harish Reddy wrote: > I need some help with optimization. I am confused what to do. Your best option is to build more models, assess them, and choose the best one (or change your alignment/template if none of the models are acceptable). You would be better off doing this with Python scripts, as the old Modeller TOP scripts are no longer documented or supported. > On the first instance i used the scrip one to build the models. As i was > not satisfied with the structural quality of the models, i used the > second script below where i included a refinement command. Even then i > couldn't get good models. So, i tried to a optimization script (third) Actually, all your scripts do the same thing - they call the regular modeling procedure. You have changed the amount of refinement and the optimization parameters slightly, but this does not usually have a large influence on the output models. > But that > script just kept on failing. So, I am not sure what the problem is.. I can't tell what the problem is either! What do you mean by "the script just kept on failing" ? Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage ------------------------------ Message: 2 Date: Mon, 10 May 2010 17:19:16 -0300 From: Caio Veloso <caju(a)uai.com.br> Subject: [modeller_usage] How to generate Homology models using a computer farm To: modeller_usage(a)salilab.org Message-ID: <4BE86A44.5000807(a)uai.com.br> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Dear All, In order to build several models of a target protein, we are running the same Python script on different computers at the same time. In this case, the script is the same and the proteins used as models are also the same. At the first attempt we expected to build different models and then assess them, and choose the best one or change your alignment/template if none of the models are acceptable. However when these models were evaluated, we discovered that MODELLER has built the same models, in the same order on all computers. I want know whether exist some parameters that could be adjusted at the Python script, (a kind of random number generator ?) in order to assure that different sets of models will be generate on each computer? Regards, Caio Veloso ------------------------------ Message: 3 Date: Mon, 10 May 2010 14:29:25 -0700 From: Modeller Caretaker <modeller-care(a)salilab.org> Subject: Re: [modeller_usage] How to generate Homology models using a computer farm To: Caio Veloso <caju(a)uai.com.br> Cc: modeller_usage(a)salilab.org Message-ID: <4BE87AB5.7090104(a)salilab.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed On 5/10/10 1:19 PM, Caio Veloso wrote: > In order to build several models of a target protein, we are running > the same Python script on different computers at the same time. ... > we discovered that MODELLER has built the same models, in the same > order on all computers. Indeed, because Modeller is deterministic, and gives reproducable results. > I want know whether exist some parameters that could be adjusted at > the Python script, (a kind of random number generator ?) in order to > assure that different sets of models will be generate on each > computer? You need to set rand_seed to a different value on each machine - something like -2 - (number of machine) should do the trick: http://salilab.org/modeller/9v8/manual/node110.html Ben Webb, Modeller Caretaker -- modeller-care(a)salilab.org http://www.salilab.org/modeller/ Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage ------------------------------ Message: 4 Date: Tue, 11 May 2010 09:54:33 +0800 (CST) From: Yue <coolmoon(a)yahoo.cn> Subject: [modeller_usage] how to extract the part of structure from a complex ? To: modeller_usage(a)salilab.org Message-ID: <11138.10742.qm(a)web92402.mail.cnh.yahoo.com> Content-Type: text/plain; charset="iso-8859-1" Hi, everybody ?I have a? question, how to extract a ligand-part fragment of protein (mainly is catalyzed domain) in PDB format?? e.g. in advaced example(), how? to get the 1emd_bs from 1emd.pdb? ?I tried to text editor to edit the part, but i failed. ?Thanks for any advise! ?????????????????????????????????????????????????????????????????????????????????????????????????????????? YJ ? ?

1 0

DOPE profile multiple templates
by Marie-Josée Porcheron 11 May '10

11 May '10

Hi all, How to build the DOPE profile for mutliple templates modelling? (in the advanced tutorial, a TvLDH-mult_dope.dat file is present, but not the info on how to obtain it ^^). Thank you very much. Joe _________________________________________________________________ Hotmail : une messagerie performante et gratuite avec une sécurité signée Microsoft https://signup.live.com/signup.aspx?id=60969

2 1

how to extract the part of structure from a complex ?
by Yue 11 May '10

11 May '10

Hi, everybody I have a question, how to extract a ligand-part fragment of protein (mainly is catalyzed domain) in PDB format? e.g. in advaced example(), how to get the 1emd_bs from 1emd.pdb? I tried to text editor to edit the part, but i failed. Thanks for any advise! YJ

2 1

How to generate Homology models using a computer farm
by Caio Veloso 10 May '10

10 May '10

Dear All, In order to build several models of a target protein, we are running the same Python script on different computers at the same time. In this case, the script is the same and the proteins used as models are also the same. At the first attempt we expected to build different models and then assess them, and choose the best one or change your alignment/template if none of the models are acceptable. However when these models were evaluated, we discovered that MODELLER has built the same models, in the same order on all computers. I want know whether exist some parameters that could be adjusted at the Python script, (a kind of random number generator ?) in order to assure that different sets of models will be generate on each computer? Regards, Caio Veloso

2 1

How to optimize a Homology model.
by Harish Reddy 10 May '10

10 May '10

Hi, I need some help with optimization. I am confused what to do. This is what i have done so far. On the first instance i used the scrip one to build the models. As i was not satisfied with the structural quality of the models, i used the second script below where i included a refinement command. Even then i couldn't get good models. So, i tried to a optimization script (third), to see if i can build the models by iterative optimization. But that script just kept on failing. So, I am not sure what the problem is.. *One*: This one does not have the refine command # Homology modelling by the MODELLER TOP routine 'model'. INCLUDE # Include the predefined TOP routines SET OUTPUT_CONTROL = 1 1 1 1 2 # uncomment to produce a large log file SET ALNFILE = '2BCK.pir' # alignment filename SET KNOWNS = '2BCK' # codes of the templates SET SEQUENCE = 'MesoA1' # code of the target # SET ATOM_FILES_DIRECTORY = './:../atom_files' # directories for input atom files SET STARTING_MODEL= 1 # index of the first model SET ENDING_MODEL = 3 # index of the last model # (determines how many models to calculate) SET HETATM_IO = on CALL ROUTINE = 'model' # do homology modelling STOP -------------------------------------------------------------------------------------------------------------------------------------- * * *Two*: This has the refine command INCLUDE # Include the predefined TOP routines SET ALNFILE = '2BCK.pir' # alignment filename SET KNOWNS = '1BCK' # codes of the templates SET SEQUENCE = 'MesoA1' # code of the target SET ATOM_FILES_DIRECTORY = './:../../atom_files' # dirs for input atom files SET STARTING_MODEL = 1 SET ENDING_MODEL = 5 SET DEVIATION = 6.0 SET MD_LEVEL = 'refine_2' #SET OUTPUT_CONTROL = 1 1 1 1 1 SET OUTPUT_CONTROL = 1 0 0 1 0 CALL ROUTINE = 'model' # do homology modelling --------------------------------------------------------------------------------------------------------------------------------------- * * *Third*: INCLUDE SET ALIGNMENT_FILE = '2BCK.pir', KNOWNS='2BCK', SEQUENCE='MesoA1' # Very thorough VTFM optimization: SET LIBRARY_SCHEDULE = 1, MAX_VAR_ITERATIONS = 300 # Very thorough MD optimization: SET MD_LEVEL = 'refine1' # Repeat the whole cycle 3-times and do not stop unless obj.func. > 1E6 SET REPEAT_OPTIMIZATION = 3, MAX_MOLPDF = 1E6 CALL ROUTINE = 'model' --------------------------------------------------------------------------------------------------------------------------------------- I will appreciate if some one can help me. thanks in advance!! Harish

2 1

modelling discarding part of a template
by Anna Marabotti 05 May '10

05 May '10

Hi, sorry for a probable trivial question, but I couldn't find a solution browsing the mailing list by myself. I want to model a protein using two different templates, in particular they differ in the position of a loop. My protein is very similar to template 1, but should have that loop in the same conformation found in template 2. So I'd like to "say" to modeller: please model my protein using template 1 as much as possible, but use only template 2 to model that loop. Please take in mind that both templates cover the entire sequence and show the loop (in different conformations). I tried to do this by replacing the residues that I would like to be discarded (in correspondence of the loop) in template 1 with gaps in the alignment. So for example: target: AAAAAAAAAAAAAAAAA template 1: BBBB-------BBBBBB template 2: ZZZZZZZZZZZZZZZZZ (the part with gaps in template 1 is present in PDB file) But Modeller returned me an error saying: Sequence difference between alignment and PDB: Alignment: BBBBBBBBBB Sequence: BBBBBBBBBBBBBBBBB I understand that Modeller "interpret" my alignment as a gap would present in the structure, and since this is not the case, it returns me an error. The question is: how to correctly send information to Modeller in order to "force" it to follow only a template to model a portion of the sequence, even if the other template is present? It seems to me that there was a way to do it without modifying the PDB file of the first template in order to cancel this loop. Am I correct? Is there a solution? Thank you very much for kind assistance and best regards Anna ____________________________________________________ Anna Marabotti, Ph.D. Laboratory of Bioinformatics and Computational Biology Institute of Food Science, CNR Via Roma, 64 83100 Avellino (Italy) Phone: +39 0825 299651 Fax: +39 0825 781585 Email: anna.marabotti(a)isa.cnr.it Skype account: annam1972 Web page: http://bioinformatica.isa.cnr.it/anna/anna.htm "If you think you are too small to make a difference, try sleeping with a mosquito"

2 2

add restype to and how to select ligand
by Knut J Bjuland 05 May '10

05 May '10

If wish to use acetyl-coa in model which use a polypetid as template. How can I specify to modeller to use the atom in hetatm since I often get other petids if I cut of the amino acid to get better alignment. Knut J _________________________________________________________________ Hotmail: Free, trusted and rich email service. https://signup.live.com/signup.aspx?id=60969

2 1

weighting restraints
by meier＠lmb.uni-muenchen.de 04 May '10

04 May '10

Hi, having multiple templates, I want to weight template based restraints down with respect to other (stereochemical) restraints. Is it then more sensible to weight down the template based restraints (weight < 1, weight=1 for rest) or to weight up the other restraints (weight > 1, weight=1 for rest)? Does it make any difference for the (simulated annealing) optimization scheme (because the "scale" is changed)? Thanks a lot for any hint Armin Meier

2 1

Modeling Circular Permutations
by Spencer Bliven 04 May '10

04 May '10

Hello, I have produced a number of circular permutations for IL-1b. There are several structures available for the wild type protein, so I would like to model how the circular permutation effects the structure. Is MODELLER able to process circular permutations? If so, how would I encode this in my alignment file? It doesn't seem like the PIR format is well suited to this. Thanks in advance for your advice to a new modeller user. -Spencer Bliven

2 1