mobileRNA: an integrative, automated tool for accurate identification of mobile RNA molecules based on multiple genome mapping

RNA molecules can act as intercellular signals, with evidence of transport for both messenger RNA (mRNA) and small RNA over long distances. Such molecules that move from the cell in which they originated are termed mobile RNA. In grafted plants, when different genotypes are joined together, the movement of RNAs between tissues can be revealed by approaches that exploit natural polymorphisms between the genomes involved. However, current methods suffer from a high level of technological noise, and the lack of standardized approaches makes it challenging for researchers to robustly identify mobile RNA molecules. Here, we introduce an automated pipeline, mobileRNA, which uses a multi-genome approach based on simultaneous alignment of RNA sequencing data to merged reference genomes. We validated our approach using simulated grafting combinations and compared to other methods. When applying mobileRNA to real data, we find poor overlap to existing mobile mRNA annotations, consistent with recent studies that question the reliability of previously reported mobile mRNAs. Introducing a genome mappability index as a criterion for the likely success of detecting mobile RNAs, we find that SNP-based or two-step mapping approaches are inherently limited due to errors and poor mappability compared to cross-species grafting with large evolutionary distances.