Dear all,

I have come across examples of people filtering out immunoglobulin genes during RNA-seq analysis, such as in this paper :

https://www.nature.com/articles/s41467-019-11373-9

However, I am interested in the immunology of the condition I am studying and hence in the expression of these genes.

Is there a reason why they should be removed?

If yes should these genes, alongside non-coding genes, pseudogenes etc. be filtered before or after normalisation & differential expression analysis with Deseq?

Thank you for the thoughts.

I was the senior bioinformatician on the paper you cite. We did not filter Ig genes but instead filtered Ig gene segments, which are the building blocks for B cell receptors. We filtered the receptor segments because

• they are not part of the regular transcriptome, indeed they are not true "genes" according to most definitions, and
• we were not studying B cells so B cell receptors should not be expressed in our experiment

RNA-seq allows you to interrogate many different classes of RNA. It is up to you to choose which classes of RNA are biologically relevant for your study.

We used the limma package to conduct differential expression analyses. We filtered before the normalization and differential expression analysis, as we always do. Filtering after the differential expression analysis would be pointless.