Leveraging short and long-read sequencing to comprehensively characterize alternatively polyadenylated and alternatively spliced transcripts

Abstract

Pre-mRNA processing through alternative splicing and alternative polyadenylation generates diverse transcripts from the same gene. Alternative splicing involves use of different exons and specific 3'Untranslated regions (3'UTRs). When limited to the same terminal exon, 3'end formation through alternative polyadenylation gives rise to transcripts containing 3'UTRs of different sizes which affect their subcellular localization, stability, and translatability. The full repertoire of transcripts arising from alternative pre-mRNA processing under different biological and disease states remains unknown. Using the blood cancer, mantle cell lymphoma's (MCL) as our experimental model we found that two genes, SOX11 and ATM, that contribute to the disease pathology, express transcripts that have short 3'UTRs. Apart from select transcripts, our analysis of short-read sequencing data did not uncover widespread global shortening of 3'UTRs in MCL. However, we discovered that both the canonical and non-canonical polyadenylation signals (PASs), were used for 3'end formation. We experimentally verified the PASs used by specific transcripts. Combining short-read and long-read sequencing data enabled us to characterize, classify and identify alternatively splicing in MCL. We identified splice variants for several genes including PTEN, EZH2 and validated the expression of GLS1 at the protein level and by using GLS1 inhibitors. Collectively, our approach provide insights into how to identify alternatively spliced and alternatively polyadenylated transcripts.