What issue are you trying to fix? Those dictionaries were separate for a reason.

4 out of 6 entries in the dictionary exist in structures in PDB (a total of 90 structures), and you would fail when you read the sequence from these structures without the extended entries for 3to1 conversion. I don't think the current implementation is providing a separate dictionary for 3to1 conversion for these 6 residues. Also, what's the reason to separate them for 3to1 conversion?

Below is the structure count for the residues in the dictionary

• ASX: 6
• GLX: 5
• SEC: 75
• PYL: 4

Many of the 3-letter codes used in the PDB are not defined in the IUPAC listing, thus a separate listing.

But what is the harm in including those in the extended 3to1 dict?

There are PDB extensions to the core 3-letter protein names which are not IUPAC extensions.

Sorry, I have trouble understanding this. Isn't the protein_letters_3to1_extended supposed to be everything including the residue names defined by IUPAC? The current one does not include the reverse mapping of the additional entries from protein_letters_1to3_extended from IUPACData.py

i.e. The reverse mapping of these residue names is not included in the extended 3to1 dict

{"B": "Asx", "X": "Xaa", "Z": "Glx", "J": "Xle", "U": "Sec", "O": "Pyl"}

For example,

from Bio.PDB.Polypeptide import protein_letters_3to1_extended
protein_letters_3to1_extended['SEC']   #SELENOCYSTEINE (U)

This would raise a KeyError

Moreover, the 20 canonical residue names are indeed in the protein_letters_3to1_extended as expected

from Bio.Data.PDBData import protein_letters_3to1
from Bio.Data.PDBData import protein_letters_3to1_extended

canonical_resnames = set(protein_letters_3to1.keys())
extended_resnames = set(protein_letters_3to1_extended.keys())
canonical_resnames.issubset(extended_resnames)

This would return True, and this is because the 20 canonical residue names are defined in this extended dict through hard coding, but the IUPAC extended ones are not

As Peter mentioned, these dictionaries refer to different sources. The dict in IUPACData refers to IUPAC nomenclature, while PDBData is a collection of data from structures deposited in the wwPDB. For example, Xle is a IUPAC aminoacid, but it does not match any monomer entry in the PDB.

The ones you did find in structures should be added to PDBData. IIRC, I updated the dictionary by parsing the components-pub.cif file made available by wwPDB, but I might have missed some indeed.

Got it! Thank you for the explanations!