Statistical analysis supports pervasive RNA subcellular localization and alternative 3’ UTR regulation

Abstract

Targeted low-throughput studies have previously identified subcellular RNA localization as necessary for cellular functions including polarization, and translocation. Further, these studies link localization to RNA isoform expression, especially 3’ Untranslated Region (UTR) regulation. The recent introduction of genome-wide spatial transcriptomics techniques enable the potential to test if subcellular localization is regulated in situ pervasively. In order to do this, robust statistical measures of subcellular localization and alternative poly-adenylation (APA) at single cell resolution are needed. Developing a new statistical framework called SPRAWL, we detect extensive cell-type specific subcellular RNA localization regulation in the mouse brain and to a lesser extent mouse liver. We integrated SPRAWL with a new approach to measure cell-type specific regulation of alternative 3’ UTR processing and detected examples of significant correlations between 3’ UTR length and subcellular localization. Included examples, Timp3, Slc32a1, Cxcl14, and Nxph1 have subcellular localization in the brain highly correlated with regulated 3’ UTR processing that includes use of unannotated, but highly conserved, 3’ ends. Together, SPRAWL provides a statistical framework to integrate multi-omic single-cell resolved measurements of gene-isoform pairs to prioritize an otherwise impossibly large list of candidate functional 3’ UTRs for functional prediction and study. SPRAWL predicts 3’ UTR regulation of subcellular localization may be more pervasive than currently known.