Doubts on the usage of PDB templates

[alexandrefassio]

Hi!!!

First, I'd like to get more information about the purpose of using a PDB template. I tried to generate the structures for paired Abs without templates, and your method recovered the structure with very good precision (congratulations on this work btw).

I understood from the paper templates help in guiding the prediction when we have a parent PDB file and we want to predict the structure of the mutated protein sequence. I tried to do so without the template, and the prediction for the mutated Ab was still very good. So, I'm not sure when the template option would be helpful.

Anyway...I tried to use this option and defined the following code.

igfold = IgFoldRunner()
igfold.fold(
    pred_pdb, # Output PDB file
    sequences=sequences, # Antibody sequences
    do_refine=True, # Refine the antibody structure with PyRosetta
    do_renum=False, # Renumber predicted antibody structure (Chothia)
    template_pdb="pdb_template.pdb",
    fasta_file="pdb.fasta"
)

In my case, sequences are H and L mutated chains. Initially, I only set the PDB template file. However, the code complains there is no FASTA file to recover a sequence. After debugging, I found out I should provide the option fasta_file as well. I assumed it to be the original sequence. That's because any different sequence raises an error due to a mismatch with the PDB sequence.

After defining the above code, IgFold runs. However, the final structure only contains one residue. So I believe I'm not doing this correctly. So could you help me on setting up IgFold to use templates?

thanks!!!