Computational Tool Predicts Mammalian Messenger RNA Degradation Rates

For a paper in Genome Biology, a pair of investigators at the San Francisco-based Calico Life Sciences considers the genetic and biochemical factors influencing messenger RNA (mRNA) degradation rates in humans and mice. Using transcriptome-wide mRNA half-life patterns found in dozens of human or mouse datasets, the team used available genetic and biochemical clues to put together and train statistical models such as Saluki — a hybrid machine learning-based tool for predict mRNA half-life based on RNA splice site, codon, RNA-binding, and other annotated mRNA features. "Saluki predicts the impact of RNA sequences and genetic mutations therein on mRNA stability, in agreement with functional measurements derived from massively parallel reporter assays," the authors explain, noting that the approach was also assessed with in silico saturation mutagenesis experiments. Together, these and other results suggest that Saluki "succinctly captures many of the known determinants of mRNA half-life and can be rapidly deployed to predict the functional consequences of arbitrary mutations in the transcriptome."