NCM 2023 Houston: Haplotypes, isoforms, and fusions: towards a richer cancer transcriptome

Long-read sequencing allows for the quantification of many complex aspects of RNA and so it can be used to profile biological causes of complex problems such as cancer drug resistance. We developed FLAIR, a tool for detecting alternative splicing of genes from long reads, as well as the haplotype of the reads supporting each isoform. This allows us to detect isoform-level allele bias, particularly of clinically relevant mutations. To further examine the mutational state of the transcriptome, we also developed FLAIR-fusion, which identifies the expressed isoforms of gene fusions, which are prevalent in many cancer types. We used long-read single cell sequencing and these tools to investigate the total mutational state of expressed transcripts in a cancer cell line both before and after drug treatment. We identified known and novel gene fusions and somatic mutations as well as distinct subpopulations with changes in splicing of KRAS, an important cancer driver.