Single-cell RNA sequencing analyses: interference by the genes that encode the B-cell and T-cell receptors

B and T cells are integral parts of the immune system and are implicated in many diseases, e.g. autoimmunity. Towards understanding the biology of B and T cells and subsets thereof, their transcriptomes can be analyzed using single-cell RNA sequencing. In some studies, the V(D)J transcripts encoding the variable regions of the B- and T-cell antigen receptors have been removed before the analyses. However, a systematic analysis of the effects of including versus excluding these genes is currently lacking. We have investigated the effects of these transcripts on unsupervised clustering and down-stream analyses of single-cell RNA sequencing data from B and T cells. We found that exclusion of the B-/T-cell receptor genes prior to unsupervised clustering resulted in clusters that represented biologically meaningful subsets, such as subsets of memory B and memory T cells. Furthermore, pseudo-time and trajectory inference analyses of early B-lineage cells resulted in a developmental pathway from progenitor to immature B cells. In contrast, when the B-/T-cell receptor genes were not removed, with the PCs used for clustering consisting of up to 70% V-genes, this resulted in some clusters being defined exclusively by V-gene segments. These did not represent biologically meaningful subsets; for instance in the early B-lineage cells, these clusters contained cells representing all developmental stages. Thus, in studies of B and T cells, to derive biologically meaningful results, it is imperative to remove the gene sequences that encode B- and T-cell receptors.

Automated summary

In the study, it was found that removing the gene sequences that encode B- and T-cell receptors before unsupervised clustering of single-cell RNA sequencing data resulted in clusters that represented biologically meaningful subsets. Conversely, if these genes were not removed, the resulting clusters did not represent biologically meaningful subsets.