An automated approach for global identification of sRNA-encoding regions in RNA-Seq data from <italic>Mycobacterium tuberculosis</italic>

Abstract

Deep-sequencing of bacterial transcriptomes using RNA-Seq technology has made it possible to identify small non-coding RNAs, RNA molecules which regulate gene expression in response to changing environments, on a genome-wide scale in an ever-increasing range of prokaryotes. However, a simple and reliable automated method for identifying sRNA candidates in these large datasets is lacking. Here, after generating a transcriptome from an exponential phase culture of Mycobacterium tuberculosis H37Rv, we developed and validated an automated method for the genome-wide identification of sRNA candidate-containing regions within RNA-Seq datasets based on the analysis of the characteristics of reads coverage maps. We identified 192 novel candidate sRNA-encoding regions in intergenic regions and 664 RNA transcripts transcribed from regions antisense (as) to open reading frames (ORF), which bear the characteristics of asRNAs, and validated 28 of these novel sRNA-encoding regions by northern blotting. Our work has not only provided a simple automated method for genome-wide identification of candidate sRNA-encoding regions in RNA-Seq data, but has also uncovered many novel candidate sRNA-encoding regions in M. tuberculosis, reinforcing the view that the control of gene expression in bacteria is more complex than previously anticipated.

Comprehensive Essentiality Analysis of the Mycobacterium tuberculosis Genome via Saturating Transposon Mutagenesis

“…There was modest overlap between the sRNAs identified here and the sRNAs previously identified in studies enumerating intergenic transcripts or putative sRNAs (Fig. S3b) (Wang et al [42], Arnvig et al [48]). Many of the published putative sRNAs failed to reach our depth and boundary criteria designed to distinguish them from mRNA degradation products, but differences also likely reflect alterations in culture conditions, growth phases, and methods of library preparation.…”

“…To date, these genomic features have not been included in TnSeq analyses because they contain only a few TA sites. In addition, there has been little consensus between studies describing the identification of sRNAs and their boundaries (42–45). …”

The Mycobacterium tuberculosis transcriptional landscape under genotoxic stress

“…From the 120 ncRNAs identified in the data from the MMC stressed sample, we found 93 to be novel (Additional file 6 ). Among all the 202 ncRNAs identified 44 were previously reported (Additional file 6 , highlighted in yellow) [ 25 – 29 ]. The length of the ncRNAs varied between 50 and 270 bases.…”

Genome-wide identification of the context-dependent sRNA expression in Mycobacterium tuberculosis

“…Experimentally validated sRNAs and sRNAs expressed in all conditions are highlighted in orange and red strokes respectively. Green strokes represent the expressed sRNAs and the blue strokes represent the absence of the sRNA expression bacterial genomes using expression data [18,19]. However, the challenges while using such an approach include discriminating between sRNA expression signal and the noise arising from IGRs, and to systematically eliminate the signals which are associated with the neighbouring gene expression.…”

“…In E. coli, IGRs were identified as sRNAs if they showed significant expression compared to the upstream and downstream protein-coding genes [18]. In M. tuberculosis, expression data corresponding to the log-phase growth was utilised to identify sRNAs by considering the read depth at a given position in the genome excluding the UTRs [19]. However, none of these genome-wide analyses focused on the conditional expression of the sRNAs in different stress environments.…”