Chain IDs reduced to first 3 characters

[LeeryScientist]

Version: 0.22.0
The mmCIF and MMTF parsers only keep the first 3 characters of chain IDs.
This makes handling structures like 6n1d impractical, since information is lost
and duplicate chain IDs are produced.

[padix-key]

The reason is that the chain_id annotation array uses the NumPy dtype 'U3', allowing only 3 characters. I was not aware, that 4-character chain IDs were allowed. However, four characters seems the maximum, as the MMTF format decodes the chainNameList and chainIdList fields into arrays of 4-character strings. Therefore, I will create a PR that sets the chain_id dtype to 'U4'.

[padix-key]

pdbx_file = pdbx.PDBxFile.read(rcsb.fetch("6n1d", "pdbx"))
pdbx_structure = pdbx.get_structure(pdbx_file, model=1)
print(np.unique(pdbx_structure.chain_id))

Before:

['A16' 'A23' 'A5S' 'AL0' 'AL1' 'AL2' 'AL3' 'AMR' 'APT' 'AS0' 'AS1' 'AS2'
 'ATH' 'B16' 'B23' 'B5S' 'BAT' 'BL0' 'BL1' 'BL2' 'BL3' 'BMR' 'BPT' 'BS0'
 'BS1' 'BS2' 'BTH']

After fix:

['A16S' 'A23S' 'A5S' 'AL01' 'AL02' 'AL03' 'AL04' 'AL05' 'AL06' 'AL09'
 'AL13' 'AL14' 'AL15' 'AL16' 'AL17' 'AL18' 'AL19' 'AL20' 'AL21' 'AL22'
 'AL23' 'AL24' 'AL25' 'AL27' 'AL28' 'AL29' 'AL30' 'AL31' 'AL32' 'AL33'
 'AL34' 'AL35' 'AL36' 'AMRN' 'APTN' 'AS02' 'AS03' 'AS04' 'AS05' 'AS06'
 'AS07' 'AS08' 'AS09' 'AS10' 'AS11' 'AS12' 'AS13' 'AS14' 'AS15' 'AS16'
 'AS17' 'AS18' 'AS19' 'AS20' 'ATHX' 'B16S' 'B23S' 'B5S' 'BATN' 'BL01'
 'BL02' 'BL03' 'BL04' 'BL05' 'BL06' 'BL09' 'BL13' 'BL14' 'BL15' 'BL16'
 'BL17' 'BL18' 'BL19' 'BL20' 'BL21' 'BL22' 'BL23' 'BL24' 'BL25' 'BL27'
 'BL28' 'BL29' 'BL30' 'BL31' 'BL32' 'BL33' 'BL34' 'BL35' 'BL36' 'BMRN'
 'BPTN' 'BS02' 'BS03' 'BS04' 'BS05' 'BS06' 'BS07' 'BS08' 'BS09' 'BS10'
 'BS11' 'BS12' 'BS13' 'BS14' 'BS15' 'BS16' 'BS17' 'BS18' 'BS19' 'BS20'
 'BTHX']

[padix-key]

The PR is merged now, the issue should be fixed. Thanks for bringing this up.