Zinc-mediated RNA fragmentation allows robust transcript reassembly upon whole transcriptome RNA-Seq

Abstract

Whole transcriptome RNA-Seq has emerged as a powerful tool in transcriptomics, enabling genome-wide quantitative analysis of gene expression and qualitative identification of novel coding or non-coding RNA species through transcriptome reassembly. Common protocols for preparation of RNA-Seq libraries include an RNA fragmentation step for which several RNA sizing techniques are commercially available. To date, there is no global information about their putative bias on transcriptome analysis. Here we compared the effects of RNase III- and zinc-mediated RNA fragmentation on transcript expression measurement and transcriptome reassembly in the budding yeast Saccharomyces cerevisiae. We observed that RNA cleavage by RNase III is heterogeneous along transcripts with a striking decrease of autocorrelation between adjacent nucleotides along the transcriptome. This had little impact on mRNA expression measurement, but specific classes of transcripts such as abundant non-coding RNAs were underrepresented in the libraries constructed using RNase III. Furthermore, zinc-mediated fragmentation allows proper reassembly of more transcripts, with more precise 5' and 3' ends. Together, our results show that transcriptome reassembly from RNA-Seq data is very sensitive to the RNA fragmentation technique, and that zinc-mediated fragmentation provides more robust and accurate transcript identification than cleavage by RNase III.

Keywords: RNA-Seq; RNase III; Segmentation; Zinc fragmentation; ncRNA.