Lessons from an Antibody Amplicon-Seq Project

I thought the Antibody Amplicon-Seq project was very interesting (as part of the broader study in Mutsvunguma et al. 2019, also described in the earlier pre-print), so that I thought there was value in re-emphasizing some points here:

• From this project, I learned about the antibody 72A1 (which I believe was purified from the HB-168 hybridoma cell line for this study).  72A1 consists of 2 previously described heavy-chain sequences and 2 previously described light-chain sequences, so this provides a very nice positive control.
• The 72A2 L2 sequence did not appear in any MiSeq reads, but it could clearly be amplified with Sanger sequencing with a different primer pair (this is what is meant by the "specific primers for the lambda light chain").  However, importantly, the trace file had to be visually inspected.  For example, the automated EMBOSS-merged sequence incorrectly predicted a pre-mature stop codon.  Upon visual inspection of the Sanger trace, we could tell that the public 72A1 L2 sequence was correct (with the right trimming and editing of this sequence).  So, if you use Sanger sequence (or even next-generation sequencing) and you get a potentially pathogenic result (such as a frameshift), it is extremely important you visually inspect your data!
• It is not safe to assume the automated sequence / analysis is always right (and I felt a certain obligation to bring this up, in terms of the broader implications for genetic testing / counseling)
• In addition to sequences being completely unable to be amplified, there was a non-coding sequence that was usually the sequence with the highest frequency in the light-chain samples.  This is important for other people to realize if they use a similar primer set.

Finally, I want to congratulate my colleagues / collaborators for all of their hard work.


Change Log:

10/15/2019 - public post date
10/19/2019 - minor change, shortening final paragraph