ligand docking with rosetta_scripts.linuxgccrelease

Member Site Forums Rosetta 3 Rosetta 3 – General ligand docking with rosetta_scripts.linuxgccrelease

  • This topic has 3 replies, 2 voices, and was last updated 9 years ago by Anonymous.
Viewing 1 reply thread
  • Author
    • #2149

        I installed the latest rosetta common (rosetta_src_2015.05.57576_bundle.tgz) and tried the following demo with error:

        setting ligand_soft_rep weight hack_elec to 0.42

        ERROR: unrecognized score_type type hack_elec
        ERROR:: Exit from: src/core/scoring/ line: 477

        Error: ERROR: Exception caught by JobDistributor while trying to get pose from job ‘1HXW_input_0001’

        [ERROR] EXCN_utility_exit has been thrown from: src/core/scoring/ line: 477
        ERROR: unrecognized score_type type hack_elec

        This script worked on rosetta 3.5. Could anyone explain the change ?

      • #10849

          The score term that was “hack_elec” in Rosetta3.5 and before was renamed to “fa_elec” for the weekly releases.

        • #10850

            Thank you, rmoretti. Yes, I noticed the change. I tried to run the same input files from my test. I got ~1200 total energy on rosetta 3.5, but ~750 on rosetta 2015.
            (1) Is this change a simply name replacement (Sorry, I don’t have knowledge behind these terms)? Would you keep the name in next release 3.6? What is the difference between total energies calculated by the two version?
            (2) There are two ligand dock executables, one I am using is rosetta_scripts.linuxgccrelease, I can not find description about it in the current online user guide. The other one is ligand_dock.linuxgccrelease. What is the major difference between them? which one should I use for single ligand, multi-ligand?

            Thank you in advance.


          • #10916

              The difference between hack_elec and fa_elec is just a name change, and one that will be kept going forward.

              I’m guessing the big changes in energy you see are due to the default scorefunction changes between Rosetta 3.5 and the weekly releases (which is what the 2015 releases are.) The default in 3.5 was score12, and the default in the weeklies is talaris2013. Part of that is a weights file (so would not affect usage with an explicit ligand.wts file) but part of that is a change in settings. The ligand.wts file was made in the score12 context, so it’s probably best to add the “-restore_pre_talaris_2013_behavior” flag to runs when you’re using the ligand.wts file extensively.

              An additional consideration is that the default behavior of RosettaScripts is to rescore the output file with the commandline scorefunction after the run is done. This means that the output PDBs have likely been scored by the default score12/talaris2013 weights scorefunction rather than the ligand weights one. To correct for this, you’ll want to add the tag “” to the XML at the same level as the PROTOCOL block (so inside the top-level block but outside everything else), where sfxn_name is the name of the modified scorefunction which you want to rescore the pose with.

              rosetta_scripts is a much more flexible way of specifying Rosetta protocols, not just a ligand docking application. You can do much more with rosetta_scripts than with ligand_dock, although it may be more confusing on how to change things. The RosettaScripts version is the one to prefer for all cases. ligand_dock cannot handle multiple ligand, and even for single ligand, there are new protocol tweaks that can be done with RosettaScripts that will never be back ported to the ligand_dock application. (Keep your eyes out for a DeLuca et al paper in the next few months that will discuss one such improvement.)

              The ligand_dock application at this point is mainly for backward compatibility, and so that people with old scripts and old protocols still have it around. (Although the change in defaults with scorfunctions make that somewhat moot.) You’re right that the documentation definitely needs to be updated.

          Viewing 1 reply thread
          • You must be logged in to reply to this topic.