Search

Characterization of expression profile of organisms in response to antimicrobials provides important information on the potential mechanism of action of the drugs. The special expression signature can be used to predict whether other drugs act on the same target. Here, the common response of Shigella flexneri to two inhibitors of RNA polymerase was examined using gene expression profiling. Consistent with similar effects of the two drugs, the gene expression profiles indicated that responses of the bacteria to these drugs were roughly the same, with 225 genes affected commonly. Of them, 88 were induced and 137 were repressed. Real-time PCR was performed for selected genes to verify the microarray results. Analysis of the expression data revealed that more than 30% of the plasmid-encoded genes on the array were up-regulated by the antibiotics including virF regulon, other virulence-related genes, and genes responsible for plasmid replication, maintenance, and transfer. In addition, some chromosome-encoded genes involved in virulence and genes acquired from horizontal transfer were also significantly up-regulated. However, the expression of genes encoding the beta-subunit of RNA polymerase was increased moderately. The repressed genes include those that code for products associated with the ribosome, citrate cycle, glycolysis, thiamine biosynthesis, purine metabolism, fructose metabolism, mannose metabolism, and cold shock proteins. This study demonstrates that the two antibiotics induce rapid cessation of RNA synthesis resulting in inhibition of translation components. It also indicates that the production of virulence factors involved in intercellular dissemination, tissue invasion and inflammatory destruction may be enhanced through derepressing horizontal transfer genes by the drugs.

Next generation sequencing (NGS) technologies have greatly changed the landscape of transcriptomic studies of non-model organisms. Since there is no reference genome available, de novo assembly methods play key roles in the analysis of these data sets. Because of the huge amount of data generated by NGS technologies for each run, many assemblers, e.g., ABySS, Velvet and Trinity, are developed based on a de Bruijn graph due to its time- and space-efficiency. However, most of these assemblers were developed initially for the Illumina/Solexa platform. The performance of these assemblers on 454 transcriptomic data is unknown. In this study, we evaluated and compared the relative performance of these de Bruijn graph based assemblers on both simulated and real 454 transcriptomic data. The results suggest that Trinity, the Illumina/Solexa-specialized transcriptomic assembler, performs the best among the multiple de Bruijn graph assemblers, comparable to or even outperforming the standard 454 assembler Newbler which is based on the overlap-layout-consensus algorithm. Our evaluation is expected to provide helpful guidance for researchers to choose assemblers when analyzing 454 transcriptomic data.