DIMPL: a bioinformatics pipeline for the discovery of structured noncoding RNA motifs in bacteria

Abstract

Summary Recent efforts to identify novel bacterial structured noncoding RNA (ncRNA) motifs through searching long, GC-rich intergenic regions (IGRs) have revealed several new classes, including the recently validated HMP-PP riboswitch. The DIMPL discovery pipeline described herein enables rapid extraction and selection of bacterial IGRs that are enriched for structured ncRNAs. Moreover, DIMPL automates the subsequent computational steps necessary for their functional identification. Availability and Implementation The DIMPL pipeline is freely available as a Docker image with an accompanying set of Jupyter notebooks. Full instructions for download and use are available at https://github.com/breakerlab/dimpl.

The Biochemical Landscape of Riboswitch Ligands

“…As additional classes are discovered, it is likely that many others already on the list will prove to be abundant relative to those discovered later. This is expected to be true because the bioinformatics methods used to discover novel candidates − have a higher probability of encountering abundantly represented classes. From this incomplete collection, it is striking to see that the ligands for these riboswitches are clustered near important chemical and metabolic processes.…”

“…The closest riboswitches come to affecting these common energy metabolism pathways is the glycine riboswitch, which commonly regulates genes for the glycine cleavage system that directs carbon from glycine into the formation of acetyl-CoA that could then enter the citric acid cycle. Given how obvious these potential riboswitch ligands appear to be, it seems likely that at least some of these foreseen riboswitch classes might reside among the long list of “orphan” riboswitch candidates (likely more than 100) whose ligands remain to be verified. − …”

Key players in regulatory RNA realm of bacteria

“…However, considering the time needed for manual analysis and the lack of well-defined techniques such as support vector machine (SVM) in the GC-IGR approach to map genomic regions that are enriched with noncoding RNAs, a bioinformatics pipeline called Discovery of Intergenic Motifs PipeLine (DIMPL) that automates the process of total genome analysis has now been introduced by the Breaker research group [ 181 ]. This pipeline consists of 2 stages, where in the initial stage a graph is generated using the IGRs that are extracted from the genome considering its length and %GC and including labels for IGRs from known RNA families.…”

“…This pipeline consists of 2 stages, where in the initial stage a graph is generated using the IGRs that are extracted from the genome considering its length and %GC and including labels for IGRs from known RNA families. Next, the machine learning algorithm SVM classifier is used identify a contiguous region of a genome's %GC versus length plot, before moving to the most computationally intensive steps [ 181 ]. In the second stage homology analysis, secondary structure prediction, statistics and finally the visualization of genetic context are performed to identify the candidate ncRNAs.…”

“…In the second stage homology analysis, secondary structure prediction, statistics and finally the visualization of genetic context are performed to identify the candidate ncRNAs. This approach is thought to accelerate the discovery of novel ncRNAs [ 181 ].…”