Integrated detection and quantification of aberrant transcripts with novel splicing events

Splicing misregulation, such as the inclusion of previously unknown cryptic exons, is implicated in numerous diseases. Recent methods have increased accurate and efficient detection of such splicing alterations occurring in disease phenotypes. However, the quantification and differential analyses of non-canonical splicing alterations remains focused at a splice event level, thus preventing a complete view of the effects on the downstream transcriptomic landscape. Here, we present a novel and integrated pipeline, SpliCeAT, that (1) detects and quantifies differential non-canonical splicing events from short-read bulk RNA-seq data, (2) augments the canonical transcriptome with novel transcripts containing these non-canonical splicing events, and (3) performs transcript-level differential analysis to identify and quantify aberrant cryptic exon-containing transcripts based on this augmented transcriptome. Using TDP-43, an ALS/FTD-associated RNA-binding protein as an example, we identified and catalogued aberrant splicing events in embryonic mouse brains from deletion of TDP-43 in neural progenitor cells. The accuracy of our integrated pipeline was further confirmed and validated with long-read isoform sequencing. Furthermore, by comparing neuronal TDP-43 knockouts in mice with a publicly available human dataset with TDP-43 pathology, we identified and validated 4 common genes, namely, Kalrn/KALRN, Poldip3/POLDIP3, Rnf144a/RNF144A, and Unc13a/UNC13A, with cryptic exons. In summary, our integrated pipeline, novel splice events are identified, incorporated and quantified at the transcript level, thereby enabling more complete transcriptome profiling of well-annotated genomes in in the case of pathological splicing misregulation.