Staphylococcus aureus is a versatile bacterial opportunist responsible for a wide spectrum of infections. Several genomes of this major human pathogen have been publicly available for almost 10 years, but comprehensive links between virulence or epidemicity and genome content of the bacterium are still missing. This project aims at characterizing a set of small transcribed molecules currently ignored by standard automated annotation algorithms. We assessed the NanoString's nCounter Analysis System for its ability to determine the orientation and quantity of the expressed small RNA (sRNA) molecules that we recently detected with RNA-Sequencing (RNA-Seq). The expression of approximately seventy small RNAs, including sRNA localized in pathogenic islands, was assessed at 5 time points during growth of the bacterium in a rich medium. In addition, two extraction strategies were tested: RNA was either purified on columns or simply prepared from crude lysates in the presence of a chaotropic buffer. The nCounter System allowed us to perform these 64 measurements in a single experiment, without any enzymatic reaction, thus avoiding well-known technical biases. We evaluated the reproducibility and reliability of the nCounter compared to quantitative RT-PCR (RT-qPCR). By using two different designs for the two coding strands, we were able to identify the coding strand of 61 small RNA molecules (95%). Overall, the nCounter System provided an identification of the coding strand in perfect concordance with RNA-Seq data. In addition, expression results were also comparable to those obtained with RT-qPCR. The sensitivity and minimal requirements of the nCounter system open new possibilities in the field of gene expression analysis, for assessing bacterial transcript profiles from complex media (i.e. during host-pathogen interactions) or when starting from poorly purified RNA or even directly from lysed infected tissues.