I ran submit_NGK.py with the following results - any suggestions on what I did wrong?
/home/nkato2/Rosetta/main/source/bin/loopmodel.linuxgccrelease -database /home/nkato2/Rosetta/main/database/ -in:file:fullatom -loops:loop_file /home/nkato2/NGK/h3.loops -loops:remodel perturb_kic -loops:refine refine_kic -in:file:native 3I2C07162013_0001.pdb -in:file:s 3I2C07162013_0001.pdb -out:prefix 3I2C07162013_0001_committed_1_noob -out:pdb_gz -nstruct 5 -out:path noob_3I2C07162013 -mute core.io.pdb.file_data -ex1 -ex2 -extrachi_cutoff 0 -loops:outer_cycles 5 -loops:kic_bump_overlap_factor 0.36 -legacy_kic false -kic_min_after_repack true -corrections:score:use_bicubic_interpolation false -loops:kic_rama2b -loops:kic_omega_sampling -allow_omega_move -loops:ramp_fa_rep -loops:ramp_rama
the tail end of my results look like:
protocols.looprelax: === Getting Statistics
protocols::checkpoint: Deleting checkpoints of Loopbuild
protocols.loop_build.LoopBuildMover: Initial kinematic closure failed. Not outputting.
protocols.jd2.FileSystemJobDistributor: Too many retries (max_retry_job = 10)
protocols.jd2.JobDistributor: 3I2C07162013_0001_committed_1_noob3I2C07162013_0001_0005 reported failure and will retry
protocols.jd2.JobDistributor: no more batches to process...
protocols.jd2.JobDistributor: 5 jobs considered, 24 jobs attempted in 66591 seconds
if you want to see all my results, see the attached gzip file (it's called nk2.gz_.txt - to look at it rename the file to nk2.txt.gz and gunzip it)
It looks like there's issues with trying to close the loops you're modeling.
Check to make sure that the loops you're modeling are the ones you want to be. You could try increasing the size of the loops (try increasing by one residue at a time). Larger loops mean more flexibility, and thus more chance of success. Also, take a look at you input PDBs. If there's issues with the regions where the closed loops would be sitting (e.g. steric clashes), that can make them difficult to properly close.
It could also be that you just have a tough modeling job. A low rate of success may be the best you can do with it. It looks like you are getting output structures, though not as many as you asked for. Do the ones you get seem reasonable? If so, you can just keep repeating runs (or increasing the -nstruct number) until you get the number of structures you want.