@Gisela Gabernet Wanted to enquire regarding the output TSV file. I'm analysing bulk BCR-Seq prepared using the NEBNext Immune Profiling kit. I wanted to ask how the isotype column is populated? I'd assumed it would be based on the genes called on the constant chain (c_call) but I see that some of them were populated even if the c_call is not available. How is this decided?

If I recall correctly, the isotype is determined by pRESTO based on the internal c-region sequences hosted at https://bitbucket.org/kleinstein/immcantation/raw/2025594fd9a2a64df4444070171d6fc00c4e78c7/protocols/AbSeq/AbSeq_Human_IG_InternalCRegion.fasta Whereas the c_call is populated by ChangeO using the entire c-region. However, the regions amplified by the NEB kit often do not allow to accurately determine the gene, meaning that the gene call fails.

@Thomas-Wolf Verdonckt Thanks for the info. In that case, which one should I rely on? The isotype column or the c_call column? I'm using the data in another package for downstream analysis & making plots, in which it uses the c_call column to determine isotypes.

Well, I do not think there is a definitive answer. The isotype call is done in the same way as the pRESTO pipeline developed by NEBNext. https://usegalaxy.org/published/workflow?id=51f936d1a4e849fa So if you wish to compare your results with the ones of other papers who used the galaxy pipeline, use isotype. (I do not know how overlapping the results will be, but I think they should be quite similar) If your program requires the c_call (as in, the actual families, genes, and alleles annotated according to IMGT/GENE-DB standards) then you will need to use the c_call column and deal with the many missing values / resulting biases. I am not sure if this issue can be solved in a future release, curious if Gisela has some prospects on this.

I see, I will look into it. Thanks for your opinion @Thomas-Wolf Verdonckt.

Hi @Benjy, exactly as Thomas mentioned, often the sequenced C region is not long enough for IgBLAST to determine a confident c_call, that’s why for many sequences this field is empty. So for this protocol, we align the internal C region sequence as known from the position of the NEBNext primers and determine the isotype this way. So you will definetely have more confident C region annotations, and more sequences with an assigned isotype, using the isotype column.

Hi @Gisela Gabernet thank you for the information! Appreciate it!

Hi @Benjy, as this was confusing for several users, we decided to annotate the C region or C primer information in the c_call column, for the sequences in which the c_call column is NA. This is introduced in v4.2.0 (soon to be released). Hope this is helpful!