You are here

 The error of InterfaceAnalyzer

4 posts / 0 new
Last post
 The error of InterfaceAnalyzer

I design the catalytic pocket of the enzyme and obtain new PDB.

But  when I use InterfaceAnalyzer, it occurs error.

The PDB has 3 chain , A (protein), Z(cofactor, 582) and X (ligand, 583).

How can I solve it?


core.import_pose.import_pose: (0) File 'protein_LIG.pdb' automatically determined to be of type PDB
core.conformation.Conformation: (0) Found disulfide between residues 162 204
core.conformation.Conformation: (0) current variant for 162 CYS
core.conformation.Conformation: (0) current variant for 204 CYS
core.conformation.Conformation: (0) current variant for 162 CYD
core.conformation.Conformation: (0) current variant for 204 CYD
core.scoring.ScoreFunctionFactory: (0) SCOREFUNCTION: ligandprime
apps.public.analysis.InterfaceAnalyzer: (0) Using interface definition: AZ_X
protocols.analysis.InterfaceAnalyzerMover: (0) Using explicit constructor
protocols.analysis.InterfaceAnalyzerMover: (0) Using interface constructor
core.kinematics.FoldTree: (0) [ ERROR ] FoldTree::reorder( 583 ) failed, new/old edge_list_ size mismatch
core.kinematics.FoldTree: (0) [ ERROR ] old_edge_list.size() 5  new_edge_list.size()3
core.kinematics.FoldTree: (0) [ ERROR ] FOLD_TREE  EDGE 1 581 -1  EDGE 1 583 1  EDGE 582 582 -1  EDGE 582 581 2  EDGE 583 583 -1 


And follwing is my command

/rosetta/main/source/bin/InterfaceAnalyzer.mpi.linuxgccrelease -interface AZ_X -compute_packstat -pack_separated -score:weights ligandprime -no_nstruct_label -out:file:score_only -s protein_LIG.pdb -extra_res_fa  MOL.params LIG.params



Post Situation: 
Thu, 2017-12-21 19:04

InterfaceAnalyzer is written for protein-protein interfaces, not protein-ligand interfaces.  

Specifically the error is caused by the fact that the code that rearranges the kinematic connections in the Pose (sets up the FoldTree) is not smart enough to handle single-residue chains (your ligands).  It might be possible to give you a hack to fix this one error, but another such error is likely to occur just down the line anyway.

What calculation that InterfaceAnalyzer supplies did you want to perform?  Most of them can be acheived through some other means.  For example binding energy can be calculated by making two inputs: one with the ligand in the pocket, one with it out of the pocket (100 angstroms away or something, but still in the PDB file). Repack both with fixbb and subtract one energy from the other to get the binding energy.

Thu, 2017-12-21 20:00

Thanks a lot!

...not protein-ligand interfaces?
I just do it as the reference (Methods Mol Biol. 2016 ; 1414: 47–62. doi:10.1007/978-1-4939-3569-7_4.)
but it was  used for protein-ligand interfaces

<Add mover_name="favor_native"/>
<Add mover_name="low_res_dock"/>
<Add mover_name="designinterface"/> (PackRotamersMover)
<Add mover_name="high_res_dock"/>
<Add mover_name="add_scores"/>

First, the result was  filtered by  the following metrics:

req total_score value < -1010 
req if_X_fa_rep value < 1.0
req ligand_is_touching_X value > 0.5
output sortmin interface_delta_X # binding energy

And then the result was filtered with using InterfaceAnalyzer calculate additional metrics

req packstat value > 0.55  
req sc_value value > 0.45 
req delta_unsatHbonds value < 1.5
req dG_separated/dSASAx100 value < -0.5 
output sortmin dG_separated# binding energy


I check the rosetta manual  “This application will not work on protein-ligand interfaces.”

The reference is  unexpected!


In addition, I do a test: write the protein and the cofactor in one pdb, then relax, design and use InterfaceAnalyzer 
but it shows Failed: No atoms defined for molecule 2 and -interface A_X  havd not yet been used(log as following).

Subsequently,I just do anothor test: there are only the  protein and the ligand (not cofactor).
And it also shows -interface A_X  havd not yet been used, but the caculation does not stop.


1.How could I solve the problem of  single-residue chains if I still using  InterfaceAnalyzer?

2.Would the warming -interface A_X  havd not yet been used influence the result?

3.And what modularity of rosetta can analyse the similar  properties with packstat,sc_value and dG_separated/dSASAx100>

Thank you!

protocols.analysis.InterfaceAnalyzerMover: (0) Calculating dSASA
protocols.analysis.InterfaceAnalyzerMover: (0) Calculating per-res data
core.pack.task: (0) Packer task: initialize from command line() 
core.pack.interaction_graph.interaction_graph_factory: (0) Instantiating DensePDInteractionGraph
core.pack.pack_rotamers: (0) built 8547 rotamers at 582 positions.
core.pack.pack_rotamers: (0) IG: 7420308 bytes
protocols.analysis.InterfaceAnalyzerMover: (0) included_nres: 583
core.scoring.ScoreFunctionFactory: (0) SCOREFUNCTION: ligandprime
protocols.analysis.InterfaceAnalyzerMover: (0) Computing interface packstats.
protocols.analysis.InterfaceAnalyzerMover: (0) Total: 583 PackScoresTotal: 583
protocols.analysis.InterfaceAnalyzerMover: (0) Computing delta unsat polar residues... (0) Database file opened: scoring/score_functions/sc/sc_radii.lib
protocols.analysis.InterfaceAnalyzerMover: (0) Computing Shape Complementarity Score...
protocols.analysis.InterfaceAnalyzerMover: (0) Upstream chain(s) numbers: 1, 
protocols.analysis.InterfaceAnalyzerMover: (0) Downstream chain(s) numbers: 1, (0) Failed: No atoms defined for molecule 2
protocols.jd2.JobDistributor: (0) test reported success in 16 seconds
protocols.jd2.JobDistributor: (0) no more batches to process... 
protocols.jd2.JobDistributor: (0) 1 jobs considered, 1 jobs attempted in 16 seconds
protocols.jd2.JobDistributor: (0) WARNING: The following options have been set, but have not yet been used:
    -interface A_X

Sat, 2017-12-23 08:26

I didn't know Rocco had a paper using InterfaceAnalzyer on ligands.  I guess something in the code changed between then and now.  I added the note to the documentation in response to your question, so that the next person wouldn't have to ask; we'll remove it after Rocco gets a look at this.  

Sat, 2017-12-23 15:56