Evaluation of Long-Read RNA Sequencing Procedures for Novel Isoform Identification and Quantification in Human Whole...

Blood flows through the body and reaches all tissues, contributing to homeostasis and physiological functions. Providing information and understanding on how the transcriptome of whole blood behaves in response to physiological or pathological stimuli is critical.

We collected blood from four healthy individuals and performed long-read RNA sequencing (lrRNA-seq) for the precise identification and expression quantification of RNA variants. Moreover, we compared two genome references: the Genome Reference Consortium Human Build 38 (GRCh38) and the Telomere-to-Telomere (T2T) assembly of the CHM13 cell line (T2T-CHM13).

With GRCh38, we could identify an average of about 46,000 genes, 1.3-fold more genes than T2T-CHM13. Similarly, we identified about 185,000 isoforms with GRCh38 and 140,000 with T2T-CHM13, finding similar differences for full splice match (FSM) and incomplete splice match (ISM) transcript isoforms. There were about 90,000 novel isoforms for GRCh38 and 70,000 for T2T-CHM13, 47% and 50% of the total number of identified isoforms, respectively. Differences in isoform numbers between GRCh38 and T2T-CHM13 were identified for the subcategories "Genic Genomic", "Intergenic", and "Genic Intron". Using GRCh38, we generally identified a higher number of non-coding isoforms, as well as a higher number of isoforms aligning within intron and intergenic regions. Nonetheless, GRCh38 might incur false positive results, and T2T-CHM13 is likely more accurate for genome sequences in the repetitive regions.

LrRNA-seq is a valid method for the identification of novel isoforms in blood, and this study is a first step toward the creation of a comprehensive database of the structure and expression of transcript isoforms for optimized predictive medicine.