ERROR: Couldn't find a freaking tree!

[mandar5335]

Hi,

I am running rna_denovo with ~ 360 nt length RNA. I have provided RNA fragments derived from PDB along with RNA sequence and secondary structure. PDB fragments are renumbered to match corresponding indices in sequence.

ERROR: Couldn't find a freaking tree!
ERROR:: Exit from: src/protocols/rna/denovo/RNA_DeNovoPoseInitializer.cc line: 584
protocols.jd2.JobDistributor: (0) [ ERROR ] 

[ERROR] Exception caught by JobDistributor for job S_000010

I am getting tree error, I asusme this is related to fragments, but as numbering is correct, could not pin point the issue.

Best,
Mandar

[roccomoretti]

I don't think it's based on the fragments -- this error is happening in the FoldTree initialization, which depends on the base pairing setup (e.g. as defined by -rna::denovo::secstruct), and doesn't look to have a fragment dependency. (At least not from my quick glance at the code.)

One thing to try is to add the -rna:denovo:new_fold_tree_initializer option, which should use a different FoldTree initializer, which may circumvent the issue causing the error.

[mandar5335]

Thanks, the -new_fold_tree_initializer true -bps_moves false option worked. Could the pseudoknot basepairs in one of the PDBs be the reason?

[everyday847]

It's conceivable. In general, when you specify your secondary structure, you don't need to give any base pairs that correspond to your template PDBs (because they are just held fixed anyway). If you have base pairs between residues in your template and residues that are freely sampled, that can sometimes cause issues.

[mandar5335]

I tested few options, it is not related to pseudoknot basepairing. FARFAR2 successfully runs but template PDBs are displaced from their expected position.

Is there any way to avoid moving of template PDBs during FARFAR2 sampling? -s option should be using templates as fixed chunks.

[everyday847]

In what way are the template PDBs displaced from their expected position? If you want templates to be fixed in space relative to each other, then they have to be supplied as a single PDB.