Issues with chain numbering in disulfide bonds

Hi,

Thank you for your wonderful program. I am having issues with a nanobody structure, keeping the frameworks regions and diversifying the CDR loops. I believe due to the proximity of the main chain because of the existence of a disulfide bond, the numbering scheme is getting confused. Is there a way to force this to be the same as the input PDB? I have attached a sample nanobody and the command I am using. I have also attached a sample output so that you can see the problem.

./scripts/run_inference.py inference.output_prefix=nanobody_binders/nanobody_binders inference.input_pdb=../nanobody.pdb 'contigmap.contigs=[X1-25/7-10/X33-98/12-15/X111-121]' inference.num_designs=1000 denoiser.noise_scale_ca=0 denoiser.noise_scale_frame=0

nanobody_input.pdb.zip
rf_diffusion_output.pdb.zip

Best,
Kiarash.

Start by reproducing the command from the issue with nanobody_alone.pdb and compare the attached nanobody_output.pdb against the input numbering. Read scripts/run_inference.py and trace how the input PDB and contigmap.contigs are handled around disulfide bonds. Done means identifying why numbering changes and establishing whether input numbering can be preserved or clearly documenting the limitation.