Differential expression (alternately abundance) analysis is regularly a core tool in identifying and quantifying differences between and across groups in -omics data. In this workshop session with follow-along analysis scripts we will take a deeper look at the models underlying differential expression analysis with the particular example being the DESeq2 framework. We will examine questions around design specification, the proper use of pre-computed offsets like normalization corrections, parameter estimation and testing, and robust false discovery rate correction through post-hoc shrinkage. Examples highlighting how these approaches differ across datasets will be drawn from bulk RNAseq, single-cell RNAseq, and ChIPseq.