Bioconductor Europe Conference 2022

Abstract

I discuss my current research project, in which I am extending the software package tradeSeq for performing differential expression tests at the isoform level. tradeSeq is a highly flexible and scalable tool that allows for testing both within-lineage and between-lineage changes in gene expression. One drawback of tradeSeq is that it only allows for performing differential expression analyses, either at the level of genes (differential gene expression, DGE) or transcripts (differential transcript expression, DTE). However, it has been shown extensively that most multi-exon genes are subject to alternative splicing and can thus produce a variety of functionally different transcripts from a single genomic locus. In this context, the biological question of interest also lies in assessing the change in relative usage of transcripts within the same gene (differential transcript usage, DTU). Several Bioconductor packages like DEXSeq, DRIMSeq and our package satuRn implement differential transcript usage tests, but are limited to comparisons between discrete groups of sample or cells. We here extend tradeSeq for testing DTU along trajectories. We model transcript usage as a smooth function of (pseudo-)time by means of a quasibinomial generalized additive model.