This is a pair of residue:chain strings, which identify the starting and ending residues and chains to read from a PDB file, when reading the structure to match the sequence. This matches the pair specified in a PIR alignment file header (see Section B.1) and in the model_segment argument to model.read().