Contrastive Models for Case-Control Changes in Sequencing - Models, Inference and Algorithms

Explore advanced techniques for analyzing case-control sequencing experiments in this comprehensive lecture from the Models, Inference and Algorithms series at the Broad Institute. Delve into contrastive latent variable models presented by Didong Li from the University of North Carolina at Chapel Hill, designed to uncover complex transcriptional changes in different conditions. Learn about a novel model-based hypothesis testing framework for count data, capable of detecting global and gene subset-specific expression changes. Discover how these methods are validated through simulations and real-world gene expression data analyses. Additionally, gain insights into the contrastive regression extension, applicable to continuous covariate data in case studies. Complement this with Sarah Nyquist's primer on current high-throughput transcriptomics techniques and the growing need for contrastive methods in RNA-sequencing analysis. Understand the limitations of traditional differential expression approaches and the importance of identifying low-dimensional structures that capture variation exclusive to case data.