Condition-Specific Mapping of Operons (COSMO) Using Dynamic and Static Genome Data

An operon is a set of adjacent genes which are transcribed into a single messenger RNA. Operons allow prokaryotes to efficiently circumvent environmental stresses. It is estimated that about 60% of the Mycobacterium tuberculosis genome is arranged into operons, which makes them interesting drug targets in the face of emerging drug resistance. We therefore developed COSMO - a tool for operon prediction in M. tuberculosis using RNA-seq data. We analyzed four algorithmic parameters and benchmarked COSMO against two top performing operon predictors. COSMO outperformed both predictors in its accuracy and in its ability to distinguish operons activated under distinct conditions. Overall design: The RIF samples were obtained from two Mtb patients and RNA-sequenced. The lineages and drug resistance profiles were confirmed using the TBProfiler tool (Phelan et al., 2019) in Galaxy (https://galaxy.sanbi.ac.za/) (Jalili et al., 2020). Except for the WT family, all other strains were RIF resistant. The 64 samples consisted of six biological replicates from L2 (Beijing family) and six biological replicates from L4 (Family X). For each of the aforementioned samples, half of the biological replicates were high MIC strains (150 µg/ml) and the other half, were low MIC strains (40 µg/ml). We also isolated three biological replicates of naturally produced rpOB mutants (L2) and three wildtype (WT) strains (L2). Most biological replicates had a minimum of three technical replicates, with the exception of the WT strains. All strains were grown under RIF stress (experimental) and no RIF-stress conditions (control) – with the exceptions again of the WT strains, which were only exposed to control conditions. Under RIF-stress conditions, the low and high MIC strains were exposed to a ¼ MIC treatment.