RNA-seq based T cell repertoire extraction compared with TCR-seq.

Abstract

The purpose of this study is to evaluate the feasibility of using RNA sequencing data as substrate for the computational extraction of T cell receptor sequences. Data from hundreds of thousands of samples is available as RNA sequencing. However, the use of these data for repertoires has not been contrasted against a gold standard. We conducted a benchmarking analysis, comparing T cell receptor data extracted from RNA sequencing to those obtained from T cell receptor sequencing (as gold standard) of the same tissue samples. The focus was on the extraction of Complementarity-Determining Region 3 (CDR3) sequences. To evaluate the influence of sequencing read lengths, samples were analyzed using both 75 base pair single-end and 150 base pair paired-end sequencing methods. In addition we calculated T cell abundance in these samples to test for any correlation between reads and abundance. The findings reveal a significant, perhaps too great, discrepancy between the ability to extract Complementarity-Determining Region 3 sequences from RNA sequencing data and the results obtained from TCR sequencing. The lack of significant improvement with longer read lengths, combined with the absence of correlation to T cell abundance, emphasize the necessity of using T cell receptor sequencing methodologies.