# cyclization_type flag in simple_cycpep_predict

6 posts / 0 new
cyclization_type flag in simple_cycpep_predict
#1

Dear Rosetta users or developers,

I'm trying to model small, cyclic peptides with the simple_cycpep_predict routine. While I explicitely asked for a terminal disulfide cyclization (using the flag -cyclic_peptide:cyclization_type terminal_disulfide), the output models are cyclized with a mainchain bond (the default n_to_c_amide_bond).

I tried to turn on or off some other flags (-cyclic_peptide:require_disulfides or -cyclic_peptide:cyclic_permutations) but it didn't change anything. My input file seems to be correctly read by the program (the peptide sequence of the output models is the one I expect), and the N- and C-term are both cysteins. But during the computation, the program applies constraints on N/CA atom of the first residue and CA/C atoms of the last residue.

Below is a typical output with a toy sequence:

protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Building sequence CYS ALA ALA ALA CYS.
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Build successful.
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Setting up terminal bond between residue 1, atom  N   and residue 5, atom  C  .
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Setting up cyclic constraints.
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: The following four atoms define the terminal bond:
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: 1.     Res=5   Atom= CA
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: 2.     Res=5   Atom= C
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: 3.     Res=1   Atom= N
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: 4.     Res=1   Atom= CA
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Finished setting up constraints.
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Randomizing mainchain torsions.
core.kinematics.AtomTree: DOF for this torsion angle could not be found in AtomTree.
core.conformation.Conformation: Unable to set torsion angle in atom_tree: TorsionID 1 BB 1
core.kinematics.AtomTree: No proper DoF can be found for these four atoms: 5-1, 5-2, 5-3, 1-1!
core.kinematics.AtomTree: DOF for this torsion angle could not be found in AtomTree.
core.conformation.Conformation: Unable to set torsion angle in atom_tree: TorsionID 5 BB 2
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Performing GeneralizedKIC closure of loop.
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Setting up terminal bond between residue 1, atom  N   and residue 5, atom  C  .
protocols.relax.FastRelax: ================== Using default script ==================
protocols::checkpoint: Deleting checkpoints of FastRelax
protocols.generalized_kinematic_closure.GeneralizedKIC: Closure unsuccessful.
protocols::checkpoint: Deleting checkpoints of FastRelax
protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication: Closure failed.  Continuing to next job.



To note, I also had to modify slightly the file main/source/src/basic/options/options_rosetta.py prior to compiling as the routine simple_cycpep_predict routine used to complain (ERROR: Option matching -cyclic_peptide:cyclization_type not found in command line top-level context).

I'm using Rosetta 3.8, compiled with clang 5.0.1.

Any help appreciated!

Florent

Post Situation:
Thu, 2018-01-04 09:22
Florent Langenfeld

Seems that the online Rosetta docs is ahead of version 3.8 :-)

The feature is implemented in the latest weekly releases (at least in release 2017.52).

Solved

Mon, 2018-02-05 01:04
Florent Langenfeld

Hi - I'm having problems trying to run this command, too - I am really currently only trying to get a working version but I'm receiving the following error: "ERROR: Assertion residue_numbers_ == residue_numbers failed. ERROR:: Exit from: src/core/io/silent/SilentStruct.cc line: 1168" and then a lot of traceback.

I think my flags etc. are ok, because the equivalent command and flags work fine in v 3.7, and I receive the expected output. The reason I want 3.8 is because I want to use the disulfide cyclization type. Is this a database issue? Thanks for your help! Jennifer

/my/mpi/version/mpirun -np 4 /my/rosetta/bin/simple_cycpep_predict.mpi.linuxgccrelease -database /my/rosetta/database @flags -in:file:native /my_pept_only.pdb -nodelay 2>&1 > test_new_rosetta_run.log

Flags: -nstruct 10 -out:file:silent out.silent -extra_res -cyclic_peptide:MPI_batchsize_by_level 10 -cyclic_peptide:MPI_processes_by_level 1 3 -cyclic_peptide:MPI_output_fraction 1 #-cyclic_peptide:cyclization_type terminal_disulfide #-cyclic_peptide:require_disulfides true -cyclic_peptide:sequence_file seq.txt -score:symmetric_gly_tables true -cyclic_peptide:genkic_closure_attempts 250 -cyclic_peptide:genkic_min_solution_count 1 -cyclic_peptide:use_rama_filter true -cyclic_peptide:rama_cutoff 3.0 -cyclic_peptide:default_rama_sampling_table flat_symm_pro_ramatable -cyclic_peptide:min_genkic_hbonds 2 -cyclic_peptide:min_final_hbonds 2 -mute all -unmute protocols.cyclic_peptide_predict.SimpleCycpepPredictApplication_MPI_summary

Fri, 2018-03-23 05:45
jojenn14

Hi Jennifer,

As I said in my comment, disulfide cyclization did not work for me in v3.8 but did for weekly release 2017.52. You may give it a try. Wasn't the same error message (iirc), though, so I cannot guaranty it will solve your issue.

Cheers,

Florent

Mon, 2018-03-26 01:46
Florent Langenfeld

Hi Thanks Florent - however, the issue is arising here even without the -cyclization_type flags. Unfortuntately, I have no access to install myself, so I can only go on what I have been told, but as far as I understand, this was a freshly downloaded bundle (as of a couple of weeks ago), so should have the latest updates, or am I misunderstanding?

Jennifer

Tue, 2018-03-27 08:29
jojenn14

Hi Jennifer,

Two remarks: First, version 3.8 is the currently latest stable version of Rosetta while I'm talking about a weekly (and possibly unstable) version. Second, your error message seems to complain about residue numbers when writing the silent output file. Have you try another format?

HTH,

Florent

Tue, 2018-03-27 11:44
Florent Langenfeld