Uncertainty-aware Analysis of RNA-Seq Data Using a Tree-based Framework

This Models, Inference and Algorithms (MIA) session features two presentations on RNA-Seq data analysis challenges and solutions. Begin with Rob Patro's primer on the complexities of abundance inference in high-throughput sequencing data, exploring how read-to-target ambiguity creates uncertainty in count data. Learn about generative models developed to address these challenges, methods for statistical inference, and approaches for estimating and propagating quantification uncertainty. Then, follow Noor Pratap Singh's main presentation introducing TreeTerminus, a novel data-driven tree-based framework that incorporates uncertainty into RNA-seq analysis by constructing hierarchical structures where uncertainty decreases as you ascend the tree. Discover mehenDi, a complementary differential testing method designed to maximize signal extraction while controlling for uncertainty, enabling identification of features missed by conventional testing methods.