edger

edgeR is a Bioconductor software package for differential expression analysis of count data, notably RNA-Seq. It models read counts with a negative binomial distribution and uses empirical Bayes methods to shrink dispersion estimates, allowing reliable inference even with small sample sizes. The package provides a practical workflow to identify genes whose expression differs between conditions while accounting for biological variability and library size differences.

A core feature is normalization with the trimmed mean of M-values (TMM), which adjusts for compositional differences between libraries. edgeR uses a DGEList data container to hold counts, sample information, and normalization factors. Dispersion estimation supports common, trended, and tagwise dispersions, with empirical Bayes shrinkage to stabilize per-gene estimates. Users can fit simple two-group tests with exactTest or fit generalized linear models (GLMs) for complex designs using glmFit and perform tests with glmLRT or glmQLFTest (and related quasi-likelihood methods glmQLFit, glmQLFTest).

Outputs are typically summarized with topTags and can be further refined using decision procedures such as decideTestsDGE. The package is applicable to various count-based sequencing assays beyond RNA-Seq, including CAGE and SAGE, and supports experiments with replicates and factorial designs.

edgeR was developed by Robinson, McCarthy, and Smyth as part of Bioconductor. It remains widely used for differential expression analysis in genomic count data, valued for its robust dispersion modeling and flexible testing strategies.