DEXSeq is a statistical method designed to detect differential exon usage in RNA-seq data. It uses generalized linear models (GLMs) to account for biological variation and provides reliable control of false discoveries. The method is sensitive and specific, detecting genes and exons that are subject to differential exon usage. DEXSeq is demonstrated through its application to several datasets, including those from Brooks et al., Brawand et al., and the ENCODE Project Consortium. The method facilitates the study of alternative exon usage regulation on a genome-wide scale and is available as an R/Bioconductor package. The article also discusses the importance of modeling overdispersion, compares DEXSeq with other tools like Cuffdiff, and highlights the advantages of per-exon analysis over transcript-level analysis.DEXSeq is a statistical method designed to detect differential exon usage in RNA-seq data. It uses generalized linear models (GLMs) to account for biological variation and provides reliable control of false discoveries. The method is sensitive and specific, detecting genes and exons that are subject to differential exon usage. DEXSeq is demonstrated through its application to several datasets, including those from Brooks et al., Brawand et al., and the ENCODE Project Consortium. The method facilitates the study of alternative exon usage regulation on a genome-wide scale and is available as an R/Bioconductor package. The article also discusses the importance of modeling overdispersion, compares DEXSeq with other tools like Cuffdiff, and highlights the advantages of per-exon analysis over transcript-level analysis.