Abstract
In Dual RNA-Seq experiments the simultaneous extraction of RNA and analysis of gene expression data from both interacting organisms could be a challenge. One alternative is separating the reads during in silico data analysis. There are two main mapping methods used: sequential and combined. Here we present a combined approach in which the libraries were aligned to a concatenated genome to sort the reads before mapping them to the respective annotated genomes. A comparison of this method with the sequential analysis was performed. Two RNA-Seq libraries available in public databases consisting of a eukaryotic (Zea mays) and a prokaryotic (Herbaspirillum seropediceae) organisms were mixed to simulate a Dual RNA-Seq experiment. Libraries from real Dual RNA-Seq experiments were also used. The sequential analysis consistently attributed more reads to the first reference genome used in the analysis (due to cross-mapping) than the combined approach. More importantly, the combined analysis resulted in lower numbers of cross-mapped reads. Our results highlight the necessity of combining the reference genomes to sort reads previously to the counting step to avoid losing information in Dual RNA-Seq experiments. Since most studies first map the RNA-Seq libraries to the eukaryotic genome, much prokaryotic information has probably been lost.