The paper introduces *DESeq2*, an advanced statistical method for analyzing RNA-seq data to identify differentially expressed genes. *DESeq2* addresses the challenges of small replicate numbers, discreteness, large dynamic range, and outliers in RNA-seq data by using shrinkage estimation for dispersions and fold changes. This method improves the stability and interpretability of estimates, enabling a more quantitative analysis focused on the strength of differential expression rather than just its presence. The package employs a generalized linear model (GLM) with a negative binomial distribution to model count data, allowing for the estimation of dispersion and fold change using empirical Bayes shrinkage. This approach helps to reduce the noise in low-count genes and provides more reliable estimates of fold changes. *DESeq2* also includes features such as automatic independent filtering, hypothesis testing with thresholds on effect size, and detection of count outliers. The method is evaluated through simulations and real data benchmarks, demonstrating its high sensitivity, precision, and control of false positives compared to other methods like *DESeq*, *edgeR*, *DSS*, and *EBSeq*. The package is available as an R/Bioconductor package.The paper introduces *DESeq2*, an advanced statistical method for analyzing RNA-seq data to identify differentially expressed genes. *DESeq2* addresses the challenges of small replicate numbers, discreteness, large dynamic range, and outliers in RNA-seq data by using shrinkage estimation for dispersions and fold changes. This method improves the stability and interpretability of estimates, enabling a more quantitative analysis focused on the strength of differential expression rather than just its presence. The package employs a generalized linear model (GLM) with a negative binomial distribution to model count data, allowing for the estimation of dispersion and fold change using empirical Bayes shrinkage. This approach helps to reduce the noise in low-count genes and provides more reliable estimates of fold changes. *DESeq2* also includes features such as automatic independent filtering, hypothesis testing with thresholds on effect size, and detection of count outliers. The method is evaluated through simulations and real data benchmarks, demonstrating its high sensitivity, precision, and control of false positives compared to other methods like *DESeq*, *edgeR*, *DSS*, and *EBSeq*. The package is available as an R/Bioconductor package.