You are here

homology membrane modeling demo in rosetta 3.5

2 posts / 0 new
Last post
homology membrane modeling demo in rosetta 3.5

Hi everyone,

I know this is a simple question but I'm having a lot of trouble and I haven't found the answer. I'm using rosetta 3.5 and I'm attempting to model a ion channel that has homology with a crystal structure found on pdb. My goal is to do homology modeling of several membrane passes then loop/abinitio modeling on about 100 amino acids in the c-terminus. I'm attempting to follow the instructions given in the rosetta_demos/membrane_homology_readme, however, that file does not specify what rosetta executable in the rosetta_source/bin file to use. I attempted to make all the input files and used all the flags recommended. I have tried minirosetta and loop modeling with no success, my model always comes out as a straight alpha-helix with coil-coil at the ends. My protein of interest is supposed to have two membrane passes with a membrane dip in-between the two, and an unknown modeling of the c-terminus of 100 amino acids.

I am using pdb 2r9r as a template
./../rosetta_source/bin/loopmodel.default.linuxgccrelease \
-database ../../rosetta_database \
-in:file:fasta kv74225.fasta \
-run:protocol looprelax \
-loops:frag_sizes 9 3 1 \
-loops:frag_files kv74225_09_05.200_v1_3 kv74225_03_05.200_v1_3 none \
-loops:input_pdb out.pdb \
-loops:loop_file kv74225.loopfile \
-in:file:spanfile kv7-4.span \
-out:file:fullatom \
-out:file:silent_struct_type binary \
-out:file:silent kv7-4-homology_${1}.out \
-nstruct 10 \
-loops:remodel quick_ccd \
-loops:random_loop \
-loops:idealize_before_loop_close \
-loops::relax fastrelax \
-fail_on_bad_hbond false \

I've also run this using the loop_modeling executable.

Please let me know if you need any more information

Thanks again,


Post Situation: 
Mon, 2013-09-09 10:26

Looking at the demo, the author *must* have meant the minirosetta application, because that is the only application that uses the "run:protocol" command line flag. (I tweaked the documentation to reflect this).

The bigger problem, of course, is that you already tried that!

It sounds like the problem is the "modeled" parts of your protein aren't moving. Is it that NONE of the protein is moving, or the WRONG PARTS of the protein are moving? I suspect the problem is the loops file is not suggesting the correct residues (this is consistent with the loop_modeling executable failing in the same fashion). Double check that your loops file is written in residue numbers (from 1), not in PDB numbering.

Sun, 2013-09-22 11:45