Le 23/06/2016 à 06:24, Sushil Mishra a écrit :
> Dear all: > > I am quite new to MODELLER and I would like to know if it is possible > to to models a protein having with few trans-membrane (TM) regions. > Are there such an example or tutorial available? I tried to search > archive but, unfortunately, there was nothing that can help me. Is it > possible to assign a few TM regions (in the form of restrain or > something like that) which would be modeled in a plane. Any > suggestions/inputs related to modelling enzyme having TM regions will > be appreciated. My system has very poor similarity (< 20 %) in PDB, > and also no similarity between TM regions of query and TM regions of > proteins in PDB. > > Best, > SK
Dear Sushil,
Membrane protein prediction is the most difficult modelling part you'll have to address. If possible, decompose your protein into different domains and assemble them adterwards.
For the transmembrane region, since mostly no templates are available, I would recommend using ROSETTA for this purpose and MODELLER to assemble the results, but not MODELLER alone, see for instance: https://www.rosettacommons.org/manuals/archive/rosetta3.5_user_guide/db/d38/...
You could also have a look at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4556676/ and http://opm.phar.umich.edu/ and certainly: http://www.dsimb.inserm.fr/dsimb_tools/anvil/ http://www.dsimb.inserm.fr/dsimb_tools/maiden/
Best,
Stéphane