You are here

Rosetta Antibody Design: Error with foldtree: problem with the fold_tree:biggest_label != num_jump 2 1

3 posts / 0 new
Last post
Rosetta Antibody Design: Error with foldtree: problem with the fold_tree:biggest_label != num_jump 2 1
#1

Hi all,

I'm attempting to use Rosetta Antibody Design (RAbD) v2019.35.60890 to design antibodies for a protein whose structure is a trimer.
BACKGROUND: I can successfully run the RAbD protocol on the ful trimer structure from the PDB (6VXX) with the following command:

antibody_designer.default.linuxgccrelease \
    -antibody:design:graft_design_cdrs L1 L2 L3 H1 H2 H3 -antibody:design:seq_design_cdrs L1 L2 L3 H1 H2 H3 \
    -antibody:light_chain kappa -score:mc_optimize_dG \
    -antibody:design:use_epitope_constraints \
    -antibody:design:epitope 545G 546L 547T 548G 549T 550G 551V 552L 553T 554E 555S 556N 557K 558K \
    -run:constant_seed -run:jran 21212121 -out:pdb_gz True \
    -antibody:design:do_dock -nstruct 16000 \
    -antibody:design:outer_cycle_rounds 10

I have attempted to run it on a single monomer of the structure, which produces an error:

core.kinematics.FoldTree: [ FATAL ] problem with the fold_tree:biggest_label != num_jump 2 1
core.kinematics.FoldTree: [ FATAL ] FOLD_TREE EDGE 1 72 -1 EDGE 72 348 -1 EDGE 455 456 -2 C N EDGE 456 349 -1 EDGE 456 579 -1 EDGE 579 580 2 EDGE 580 800 -1

From the limited information I can find on the FoldTree, I know this error indicates a problem with the monomer structure that is somehow not present in the full trimer structure, but I haven't found any information on the biggest_label issue and what that means for narrowing down the problem.

To create the monomer structure that fails, I took the original trimer structure (that runs without issue) and used PyMol to delete all non-protein structures (including glycans and ions), then removed the atoms corresponding to the redundant monomers (e.g., retained monomer chain A and removed identical chains B and C).

While there are "sequence" gaps in the monomer chain, the gaps are not present in the structure, e.g., the sequence may jump from K344 to R353, but the K344 and R353 residues are connected in the structure, so I don't think this is a broken loop issue.

QUESTION:
What does the fold_tree:biggest_label error indicate? Any insights you might have on how to resolve this issue would be greatly appreciated!

Thanks,

Carla

Category: 
Post Situation: 
Wed, 2020-05-20 08:55
cm21

Strip the ters from the PDB.  When you do this in PyMol, sometimes pymol will put TERs in the middle of the chain.  Rosetta reads this at two separate chains instead of chain break, adding jumps between those chains.  I've been burned by things like this before. So first, I would try to remove the TERs from the output PDB.  I would also try to use the score_jd2 app to see if its a problem specifically with RAbD or just loading the structure.  Also make sure the antigen is not designated as chain L and H, and after you split the trimer, renumber the antibody complex again with PyIgClassify.

Hope this helps.  If you still have issues, post the full output log, might give some more context.

 

-Jared

Wed, 2020-05-20 09:34
jadolfbr

Hi Jared,

Thanks so much for your reply!

I have confirmed that there are not TERs in the pdb file, score_jd2 runs to completion on the structure, and the antigen is chain B.

I have re-run the structure through PyIgClassify, though according to the output from that the numbering on the structure has not changed at all. I have submitted the resulting structure, and I will post further details if I get the error again.

Thanks,

Carla

Wed, 2020-05-20 11:22
cm21