Scoring functions: talaris2013 incompatible versions

Member Site › Forums › Rosetta 3 › Rosetta 3 – Applications › Scoring functions: talaris2013 incompatible versions

Viewing 2 reply threads
  • Author
    Posts
    • October 13, 2020 at 1:18 pm #3599
      Anonymous

         I’m tryintg to sort the generated sequences from the .spec output with the following command and I use the flag to activate talaris:

        ~/tools/analysis/apps/gen_pepspec_pwm.py 1cka_spec.spec 3 0.1 binding-prot_score pepspec_background.binding-prot-0.1.pwm -corrections::restore_talaris_behavior

        But I get error:

        ERROR: talaris2013(.wts) requested, but -corrections::restore_talaris_behavior not set to true. This leads to a garbage scorefunction.  Exiting.

        I read some solutions advices to the similar problems from the forum. If I get it right , for the score functions I should use either talaris or REF , as the flag doesn’t help in this case, I need to change all the REF to talaris2013 score fucntion script equals. But I have no idea how which files I should replace nor how I should edit those files.

        Could anyone help me with this?

        Thanks in advance.

      • October 13, 2020 at 5:35 pm #15550
        Anonymous

          The issue you’re running up against is that the ~/tools/analysis/apps/gen_pepspec_pwm.py script isn’t a Rosetta C++ application itself, so it’s not necessarily going to obey C++ application options when passed on the command line. (Indeed, I don’t see anything scorefunction dependent in the script.)

          I don’t know what version of the gen_pepspec_pwm.py script you’re using, but the version that I’m seeing in the current versions of Rosetta is just a file manipulation script. It won’t launch Rosetta applications by default, and it shouldn’t be giving you that error message.

          Which command are you actually running when you get that error message about talaris? It’s not necessarily the gen_pepspec_pwm.py script itself which needs the -corrections::restore_talaris_behavior flag — it’s probably the command you’re passing the generated .spec file to.

          • October 14, 2020 at 10:49 am #15555
            Anonymous

              This is the complete error message,                                                           

              [FILE]: src/core/scoring/ScoreFunctionFactory.cc                                         

              [LINE]: 157                       

              [START_MESSAGE]                       

              [ ERROR ] UtilityExitException

              ERROR: talaris2013(.wts) requested, but -corrections::restore_talaris_behavior not set to true. This leads to a garbage scorefunction. Exiting.                                                                     

              The c++ file may be ScoreFunctionFactory.cc .         

              As for the command – There is a default sorting command, that needs the term “binding -prot_score” to be passed , in my understanding.

               

            • October 14, 2020 at 1:16 pm #15557
              Anonymous

                For the command, what did you type into the command prompt which gave you that error message? Did it come directly after pressing enter on the gen_pepspec_pwm.py command above, or did that program finish and you entered in another command? (If so, what was that second command?)

                 

              • October 15, 2020 at 7:48 am #15560
                Anonymous

                  I just pressed enter after the command above, nothing more.

                  This is the protocol which I followed   https://www.rosettacommons.org/demos/latest/public/peptide_specificity/README .

              • October 14, 2020 at 12:42 pm #15556
                Anonymous

                  As addition to the above question: 

                  I don’t know which command is used backend in this case. Could you advice on how to pass the correction::restore_talariis_behavior through the neccessary command. Or which C++ file I should edit, to restore talaris?

                  Thank you

              • Author
                Posts
              Viewing 2 reply threads
              • You must be logged in to reply to this topic.