Unexpected insertion positions in IMGT-numbering results

[katyachemistry]

Hi! I am working with OAS data where IMGT-numbering is provided for each sequence, and, as stated in OAS paper (https://onlinelibrary.wiley.com/doi/10.1002/pro.4205), it was done using ANARCI tool. According to IMGT, the only insertion site is between 111 and 112 positions in CDR3. However, in some OAS records there are positions like '33A', '85A', '85B', '85C', '32A', '81A', '81B', '81C' etc. I have tried to number some of such "non-standard" sequences using ANARCI web-tool (https://opig.stats.ox.ac.uk/webapps/sabdab-sabpred/sabpred/anarci/) and got similar results. The only difference was that web-tool positions with letters were smaller by 1: e.g. 22A instead of 23A in OAS.

To reproduce you can try:

for 86A and 86B

DIQMTQSPSSLSASRGDSVTITCRASQNIGHFLNWYQHKPGRAPKLLIYSTNVLQGGVPSRFSGSASGTDTFFTLTISNLQPEDFGSYYCQQSHSTPHSFGGGTKVELK

for 22A

DIVMTQSPLSLPVTPGQPAFISDCRCSKRCQDSNEKNYYDRCQQKPGKTPPLLIYWDSYRAAGVPDRCIGSGSGTDFTLNISSLEADDVGVYYCWQDYHTPWTFGQGTKVEIK

Important note: for now, I have only checked this for kappa-chain unpaired data. Would appreciate any explanation on why insertions in non-CDR3 IMGT regions appear. Thanks.