Smirnov, A. et al. Proc. Natl. Acad. Sci. USA 113, 11591–11596 (2016).

An RNA's sequence is a poor predictor of its function, particularly for noncoding RNAs, which come in many different lengths and structures. To improve annotation, Smirnov et al. profiled the association of RNA with RNA-binding proteins by Grad-seq (gradient profiling by sequencing). They first partitioned cell lysates from Salmonella enterica by glycerol gradient centrifugation to sort complexes by size and shape. Each fraction's RNA was then sequenced and proteins were analyzed by liquid chromatography–tandem mass spectrometry. Clustering by principal-component analysis revealed two major branches: coding RNAs associated with larger, ribosomal components, and noncoding RNAs associated with low-molecular-weight complexes. The approach allowed the researchers to home in on the particular protein associated with a given RNA in order to uncover its function.