Author: Jared Adolf-Bryfogle (

PI: Roland Dunbrack


This application matches each CDR of an antibody to North/Dunbrack CDR clusters based on the lowest dihedral distance to each 'center' member of the cluster. Identifications with distances will be echoed in the output PDB file and printed to the screen.


Adolf-Bryfogle J, Xu Q, North B, Lehmann A, Roland L. Dunbrack Jr, PyIgClassify: a database of antibody CDR structural classifications , Nucleic Acids Research 2014

North B, Lehmann A, Dunbrack R, A new clustering of antibody CDR loop conformations (2011). JMB 406(2): 228-256.

RosettaAntibody Framework.


Antibodies need to be numbered in any numbering scheme accepted by Rosetta. These include:

  • Chothia_Scheme
  • Enhanced_Chothia_Scheme
  • AHO_Scheme
  • IMGT_Scheme

Kabat_Scheme is also accepted, but not fully supported due to H1 numbering conventions. Use Kabat_Scheme with caution.

Antibodies can be renumbered to the AHO_Scheme through PyIgClassify

Running the Application

App: identify_cdr_clusters


Typical Rosetta input/output options are followed. MPI is supported, pass the -l option to identify a list of PDBs. Use -no_output to not have pdb files echoed. Make sure to pass -numbering_scheme with your antibody's numbering scheme if it is not renumbered in Chothia.

Example Command Line:

identify_cdr_clusters.macosclangrelease -numbering_scheme AHO_Scheme -ignore_unrecognized_res -no_output -s 2J88.pdb

See Also