The scripts and input files that accompany this demo can be found in the demos/public directory of the Rosetta weekly releases.


Setting up the demo

Before running the demo, make sure to change the following variables in your local environment:

PATH_TO_EXE                                         # path to directory with Rosetta binaries
ROSETTA_BINARY                                      # extension of Rosetta binaries, e.g. linuxgccrelease
PATH_TO_DB                                          # path to Rosetta database

Running the demo

IO flags:

-s 1a8d_MinPacked.pdb                               # The starting structure -- must have residues for the segment to be remodeled, but these don't need to have meaningful coordinates
-loops:loop_file 1a8d.loop                          # definition of the loop to be remodeled, file format description at
-out:pdb_gz                                         # compress output structures
-in:file:native 1a8d_MinPacked.pdb                  # native or reference structure for RMSD calculation -- if this flag is not set, an RMSD of 0 will be reported

Number of structures to produce (for demo):

-nstruct 1                                          # number of structures to produce 

Number of structures to produce (for production runs):

-nstruct 500                                        # or 500 independent simulations, depending on the cluster setup.

General kinematic closure loop modeling flags:

-loops:remodel perturb_kic
-loops:refine refine_kic
-run:test_cycles                                    # fast execution, uncomment for testing purposes only
-loops:outer_cycles 5                               # for production runs
-kic_bump_overlap_factor 0.36                       # reduces the threshold for permitted clashes in initial loop closures (default value is 0.4)
-legacy_kic false                                   # remove a slight bias in the initial KIC implementation towards sampling the C-termial part of the remodeled segment more frequently
-kic_min_after_repack true                          # minimize rotamers after repacking (happens every -loops:repack_period iterations after loop closure, default 20)
-corrections:score:use_bicubic_interpolation false  # do not use bicubic interpolation, it has been observed to adversely affect the fraction of sub-Angstrom conformations on the 12-residue benchmark set

Next-generation KIC flags as described in Stein & Kortemme, PLoS ONE, 2013:

-allow_omega_move true

Packing flags:

-extrachi_cutoff 0

Example Rosetta Command Line: (where $ROSETTA3=path-to-Rosetta/main/source)

$> $ROSETTA3/bin/loopmodel.default.linuxgccrelease @input_files/local_remodeling_input/flags

Overall protocol execution (demo):

  1. scripts/ <list_of_native_structures> <output_keyword> (prepacking step)

    This will create a pre-min-packed structure to start the simulation from, as well as a log of the run. The input is a list of starting structures to be processed, as well as a keyword for the output structure and log names. Note that, while standard repacking usually takes a few minutes, with the -min_pack option it can take an hour or more, depending on the size of the structures. This option was developed by Andrew Leaver-Fay and will be discussed in a separate publication (Leaver-Fay et al., Meth Enzym, 2013).

  2. scripts/ <input_list> <output_keyword> (loop remodeling step)

    This will generate structures in which the selected segment is remodeled (1 for the demo, use at least 500 for real life problems, more for longer segments). The input is a list of (pre-packed) structures and a keyword for output naming. The Rosetta syntax for .loop files is explained at

    Note that if a native or reference structure is provided, the backbone RMSD of the remodeled segment (after superimposition of the fixed parts of the structure) will be reported by loopmodel.release and can thus be parsed from the standard output, if that is redirected into a file. The script will perform redirection into a .log file which later is gzipped.

    If submitted via qsub on an SGE cluster system, the script will distribute the number of simulations specified with the -t flag across all structures in the list, i.e., to generate 500 structures each for a list of 10 input structures, use -t 1-5000.

    An NGK trajectory (generating one model) typically take 15-30min for 12-residue lops.

  3. scripts/ <directory with log files> (parsing step)

    This script will extract the scores and RMSDs for each final model from the respective .log files, and optionally generate Rosetta-energy-vs-RMSD plots. Note that relevant RMSDs will only be reported if a native or reference structure is provided to the NGK run via the -in:file:native flag. If no reference structure is available, users should consider clustering of their results to identify the most commonly sampled conformations (see Figure S1 of Stein & Kortemme, PLoS ONE, 2013).


  • Amelie Stein
  • Tanja Kortemme

Rosetta Version

SVN revision 51851 (Dec 2012)


  • Stein A, Kortemme T (2013) Improvements to robotics-inspired conformational sampling in Rosetta. PLoS ONE (submitted)