Feng, H. et al. Nat. Commun. 6, 7816 (2015).

The enormous number of public gene expression data sets can turn into research gold when mined for any number of biological questions. But digging into the data of other researchers is often fraught by a lack of metadata. In particular, missing electrophoresis-based RNA quality scores make it impossible to cull poor-quality samples from analysis or to take computational steps to account for RNA degradation. Feng et al. developed the mRNA integrity number (mRIN), a metric based on quantitative modeling of the degradation-dependent 3′ bias in read coverage, to assess mRNA integrity for transcriptomes or individual transcripts directly from sequence data. The authors used mRIN to quantify degradation in large-scale RNA-seq data from postmortem brain tissue in the BrainSpan and Genotype-Tissue Expression (GTEx) projects and found that degradation has a reproducible gene-specific component.