Mistake in AIRR epitope database annotation

[grst]

In the chapter it is suggested to convert CDR3 into junction sequences like this:

Warning: many databases follow diverging definition of the CDR3 region. While the cellranger output contains leading ‘C’ and ending ‘F’ and ‘W’, many databases do not. We therefore need to add them to the database entries, in order to obtain correct matches.

cov_abdab["IR_VJ_1_junction_aa"] = "C" + cov_abdab["IR_VJ_1_junction_aa"] + "F"
cov_abdab["IR_VDJ_1_junction_aa"] = "C" + cov_abdab["IR_VDJ_1_junction_aa"] + "W"

However, I don't think this is correct. Both VJ and VDJ sequences can be ending with either F/W depending on the J gene.

Similar problem with IEDB is currently being discussed in scverse/scirpy#469.

ping @drEast .