View on GitHub


Refinement of protein-protein complex model structures driven by interface repacking


Refinement of protein-protein complex model structures driven by interface repacking

0. Remark

The GalaxyRefineComplex distribution version supports only Linux 64-bit OS and binary files compiled with serial/parallel option.

1. Installation

  1. Download the GalaxyRefineComplex program
    • Download a copy of GalaxyRefineComplex
    • * Or, clone the GalaxyRefineComplex git repository
    • git clone
  2. Unzip and place the downloaded files
    • unzip
    • mv GalaxyRefineComplex $GALAXY_HOME
      (example: GALAXY_HOME=/applic/GalaxyRefineComplex)
  3. Execute
  4. Check the downloaded files
    • There should exist:
    • bin: directory for executables
      There should be build_initial_model, local_optimize, generate_model, and GalaxyRefineComplex
    • data: directory for data files
    • examples: directory for example files
  5. Set environment variable
      (example: export GALAXY_HOME=/applic/GalaxyRefineComplex)
  6. Find a proper GalaxyRefineComplex executable
    • We supply two different GalaxyRefineComplex executables having compiled on different environments
    • CentOS 6.6: bin/GalaxyRefineComplex.centos66
    • Ubuntu 16.04 LTS: bin/GalaxyRefineComplex.ubuntu1604 * We have compiled Python scripts using PyInstaller, and it uses built-in Linux C libraries. Since, each computing environments may have different C libraries, so you need to find a working version with your environment. If both of them are not working, then please ask us with your computing environment.
  7. Set MPI environment (optional)
    • GalaxyRefineComplex supports MPICH, so if you have installed MPICH, you can run it in parallel.
    • GalaxyRefineComplex has been tested with MPICH hydra (
    • You can use parallel version of GalaxyRefineComplex by setting EXEC_MPI environment
    • (in BASH) export EXEC_MPI=$PATH_TO_THE_MPIEXEC (example: export GALAXY_HOME=/opt/mpi/mpich/bin/mpiexec) * You can set the number of CPU to use by setting NSLOTS environment
    • (in BASH) export NSLOTS=8
    • Since it will use CPUs up to the number of simulation trajectories (default=16 for each protocol), so we recommend you to set it one of the divisor (i.e., 4, 8, 16) to make it efficient.

2. How to use GalaxyRefineComplex

  1. Prepare the input protein structure in PDB format
    • You have to provide a protein complex model in the PDB file format for complex refinement.
    • Protein chains must have different chain IDs and be separated by “TER” line.
  2. Run GalaxyRefineComplex
    • Usage: $GALAXY_HOME/bin/GalaxyRefineComplex [-h] [-p INPUT PDB File] [-t TITLE] [-f Sequence] [-s Number of samplings] [-o Number of outputs] [–sym] [–protocol1] [–protocol2]
    • Input arguments and options:
    • -p/–pdb: The input PDB file (mendatory)
    • -t/–title: The title for a GalaxyRefineComplex job
    • -f/–seq: The sequence for the PDB file in FASTA format
    • -s/–n_sample: The number of sampling trajectories (default: 32=16x2)
    • -o/–n_output: The number of refined outputs (default: 10=5x2)
    • –sym: use symmetric refinement (default: False)
    • –protocol1: use the protocol1 only (default: False)
    • –protocol2: use the protocol2 only (default: False)
  3. Output of GalaxyRefineComplex
    • The GalaxyRefineComplex generates a working directory, which is named by the job title (default: GalaxyRefineComplex)
    • In the working directory, the following three directories will be generated:
    • init: A working directory for cleaning up the input PDB file after initial side-chain optimization is performed.
    • refine: A working directory for refining protein structure model with SC perturbations and MD relaxations.
    • model: The output of GalaxyRefineComplex, refined protein model structures in PDB format will be placed.
      * The final refined model will be ${JOB Title}/model/model_[1-10].pdb

3. Release log

4. References

5. Contact