Proper Read Filtering Method to Adequately Analyze Whole-Transcriptome Sequencing and RNA Based Immune Repertoire Sequencing Data for Tumor Milieu Research

SIMPLE SUMMARY: The recent advancement in high-throughput sequencing has become indispensable for immune-genomics and profiling the T- and B-cell receptor repertoires. Immune repertoire sequencing (IR-seq) and whole transcriptome sequencing (WTS) can be implemented to investigate and quantitatively characterize the complex pattern of the CDR3 region. We conducted T-cell diversity analysis result comparisons of these sequencing methods and suggest an intuitive approach to discriminate reliable TCR sequences and clonotype patterns from capturing errors. Although bulk-RNA sequencing is commonly used for cancer analysis, we confirmed capturing highly enriched TCR transcripts with IR-seq is more reliable for accurate immune repertoire discovery, and singleton read filtering criteria should be applied to capture true clonotypes from error-prone sequencing data. The use of such well-established data and analytical methodologies can broaden understanding of antigen specificity in immunity and enabling efficient therapeutic antibody finding. ABSTRACT: Analysis of the T-cell receptor (TCR) repertoire is essential to characterize the extensive collections of T-cell populations with recognizing antigens in cancer research, and whole transcriptome sequencing (WTS) and immune repertoire sequencing (IR-seq) are commonly used for this measure. To date, no standard read filtering method for IR measurement has been presented. We assessed the diversity of the TCR repertoire results from the paired WTS and IR-seq data of 31 multiple myeloma (MM) patients. To invent an adequate read filtering strategy for IR analysis, we conducted comparisons with WTS results. First, our analyses for determining an optimal threshold for selecting clonotypes showed that the clonotypes supported by a single read largely affected the shared clonotypes and manifested distinct patterns of mapping qualities, unlike clonotypes with multiple reads. Second, although IR-seq could reflect a wider TCR region with a higher capture rate than WTS, an adequate comparison with the removal of unwanted bias from potential sequencing errors was possible only after applying our read filtering strategy. As a result, we suggest that TCR repertoire analysis be carried out through IR-seq to produce reliable and accurate results, along with the removal of single-read clonotypes, to conduct immune research in cancer using high-throughput sequencing.