Trapnell, C. et al. Nat. Biotechnol. 31, 46–53 (2013).

Turning raw RNA sequencing data into a genome-wide view of transcription requires overcoming a number of difficulties. Many unique but related transcripts can be produced by alternative exon splicing, promoters or termination, which makes it difficult to identify or quantify isoforms. To compare expression between samples, variability across replicates needs to be addressed. Trapnell et al. describe Cuffdiff 2, software that provides accurate quantification and differential expression analysis at gene and individual transcript resolutions. The approach models variability in the number of fragments associated with each transcript across replicates and can be used on data from a variety of experimental designs. Using Cuffdiff 2, the authors found transcripts that are differentially expressed between wild-type and HoxA1-deficient human cells.