i am looking for the functionality of "-favor_native_residue" (add a energy penality favoring/disfavoring the native residue) from rosetta2.3 in rosetta3.x. I found a command line flag in 3.4 "-favor_native_res". fixbb.linuxgccrelease accepts the flag without complaining, but I have the feeling he does not used it properly: e.g. I set an quite high value of -4 REU, but the mutations look roughly like w/o the flag set. Can it be that the parser reads it but the functionality is not implemented in fixbb? Greping in 3.4 for "favor_native_res" I find it mostly in the context of enzdes. Any idea how to get this functionality into fixbb design, one options seems to be RosettaScripts. What's about PyRosetta does anybody know how to directly feed favor_native into the PackRotamer mover?
Thanks a lot!
You are right that favor_native_res is an enzyme design option; it is not implemented for fixbb (so you are correct that it does not do anything). Rosetta doesn't object to valid but irrelevant options because the initialization function can't know which options will and won't be used later in the code (well, not without ultralaborious bookkeeping).
Apparently most developers use RosettaScripts with PackRotamersMover (that's all fixbb is) plus FavorNativeResidue mover (http://www.rosettacommons.org/manuals/archive/rosetta3.4_user_guide/Move...).
It might not be hard to put FavorNativeResidue mover in as an option for fixbb. (It has to be as a Mover because of the way constraints/Pose/JD2 interact). I'll look into this; if it's straightforward I'll offer you a patched fixbb.cc.
All of the existing functionality for FavorNativeResidue that I can find is either not code that can be dropped-in to fixbb (it needs to modify the ScoreFunction and Pose, which is nontrivial with fixbb because the Pose doesn't exist until JD2 runs), or won't run at all with fixbb because it's dependent on the Parser. I can't find any examples of FavorNativeResidue in use in our demos but I think it probably works. If you can't figure the script out let me know and I'll take a stab at it.
Thanks for confirming. I'll try using scripts. I think this will do.
I wrote a demo for the script. It will release with 3.5, but here it is ahead of time.