I submit the peptide fragment of 30 aa (pdb is unavailable) in robetta and ITASSER for structure prediction, I got the data from ITASSER but not from robetta (still in line) from over 30 days. But the results from itasser do not validate in SAVES. What should I do now? Any suggestions?
Maybe this can help,
http://www.academicjournals.org/AJb/abstracts/abs2011/25Jul/Butt%20et%20al.htm
Since you are an academic user, you could bypass the Robetta server and download the Rosetta software to run the modeling yourself. However, this requires significant computational resources, some knowledge of computers and command-line interface, and has a steep learning curve.
@saurabh phyre2 for more than 30 aa.........although i try it with more that 30 aa but the problem remain same.
@Amit somebody told me that modelar is not easy to handle, basically I am not expert in bioinformatics.........is there any easy method?
@David although I am not expert in bioinfo but I must follow your suggestion. Is there any easy suggestion from you?
If you have templates for comparative modeling then Modeller will actually be easier to handle because there are GUIs available.
You can download Rosetta at http://www.rosettacommons.org/. The Forums are well patrolled by the development community and can answer any questions you might have.
There are various tutorials available if you search with Google.
For shorter sequence Ab intio method is always preferable. and best server is I-TASSER. and as far as i know homology modeling is not possible in shorter sequence because you will not get appropriate template for it.
In homology modelling THE critical steps are 1) identifying a proper template and 2) aligning the target and the template sequences. The quality of an alignment of two HMMs as done by HHsearch is among the best ones possible at the moment. Submitting a sequence, this algorithms will identify homologous proteins, create a HMM and compare it to its internal HMM database. Using the web interface
http://toolkit.lmb.uni-muenchen.de/hhpred
you can use this program and prepare all data needed for modeller.
However, to me its unclear how good the quality of a 3D model for a "floppy" peptide will be. Do you expect a well defined structure or rather one induced by the interaction partner? Can cysteine-bridges stabilize this peptide?
@Rainer basically my target is a calcium binding domain (a peptide fraction of a large protein, ) and i want to dock this with another protein (calmodulin) because i get result in my wet lab for those association.
You might want to take into account the larger protein system, during computational modeling, if it at all limits the confomation of your 30 aa peptide.
Also, if you have any other experimental structural data, you can use these to narrow the conformational search space and/or provide confidence to the final model. An entire field is devoted to this: hybrid methods modeling.
http://zhanglab.ccmb.med.umich.edu/QUARK/
limit
1. >20 aa
2. institutional mail required (no login needed)
try avogadro. though made initially for small chemicals. but has an option to add aa seq.
should therefore be good for small peptides. after making sequence go to optimizing
http://avogadro.openmolecules.net/wiki/Main_Page
Dear friends, I search a lots for template of my peptide fragment but unable to get it. most probably this is the cause of my predicted model (by online servers) not to validate in PROCHEk (SAVES). What should I do now?
@Soni Shaikh, by which mehtod you have modelled the peptides? and in PROCHEK how much outliers you are getting? you can use looping for outliers.
I use ITASSER, prochek value about 83%. best model exist mostly in coil form.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools