Hi all - I am relatively new to BCR-seq analysis and am trying to run airrflow with a custom protocol that is currently not supported. For this reason, I decided that it would probably be best to assemble the reads myself using pRESTO and then feed the assembled reads as fasta into airrflow (

--mode assembled

). Now I have a question regarding the format and what type of annotations airrflow can parse. My protocol contains UMIs; however, with very low complexity which led to grouping of too diverse sequences likely due to errors in the UMI sequence. I followed the pRESTO recommendations and additionally performed a clustering within each UMI sequence to obtain more reliable UMI clusters. I then built consensus sequences for each and subsequently performed deduplication and filtering for sequences with at least 2 supporting reads as also performed by airrflow. My fasta headers look now like this:

>CTGAAT_1|PRCONS=IGHMR|PRFREQ=1.0|CONSCOUNT=4|DUPCOUNT=1

where CTGAAT_1 was the UMI + first cluster consensus sequences and the counts reflect read counts from UMI/clustering collapsing (CONSCOUNT) and counts from deduplication (DUPCOUNT). Is this a format that airrflow could detect and use downstream or do I need to modify this somehow to preserve the count information? I would appreciate any help or insights (also if there is something off with my strategy). Thanks a lot in advance!

Hi @Sara Hetzel yes that should work like that and the annotations will be parsed into the AIRR table results, the first string will be used as a sequence ID. Let us know if this doesn't work as expected.

Great, thank you!