Genomics of medical importance

Abstract

For nearly two decades, microbial genomics has been mainly performed in research settings, providing much information on the virulence and pathogenic mechanisms at play during bacterial infections, shedding light on the evolution of microorganisms, and allowing the identification of new targets for diagnosis, treatment, and/or vaccines [1]. Thanks to the recent availability of different high-throughput sequencing methods, bacterial genomes may now be sequenced in less than a week, at a cost similar to that of other medical diagnostic procedures routinely used in university hospitals, such as magnetic resonance imaging and positron emission tomography. Thus, bacterial genomics is now entering diagnostic laboratories, and it is time to summarize, in a special issue, the challenges, difficulties and opportunities for clinical microbiologists during this shift towards microbial genomics. As outlined by Bertelli et al. [2], this shift, mainly triggered by novel technologies, is mainly based on the ‘dirty genome’ concept, i.e. that raw genome sequences generally include >90% of all genes, allowing downstream proteomics to be performed on >90% of the encoded proteins [3]. Moreover, most of the genes missed with unassembled genomes are repeated elements, such as transposases and recombination hot spots, that are of little relevance for medical applications of bacterial genomics. The proof of the ‘dirty genome’ concept was provided during the recent enterohaemorrhagic Escherichia coli outbreak, as combined efforts demonstrated the feasibility of rapid genome sequencing in less than a week for the E. coli O104:H4 epidemic strain [16]. It also demonstrated the usefulness of the obtained raw genomic sequence for: (i) developing a specific molecular diagnostic test; (ii) obtaining information on virulence factors encoded on the genome; (iii) indirectly assessing the antibiotic susceptibility by looking at its resistome (antibiotic resistance-encoding genes); and (iv) understanding the evolutionary history of the strain. All of these different applications of rapid genome sequencing (Fig. 1) are extensively discussed in the first article of this special issue, after a brief description of currently available sequencing technologies. Moreover, future challenges of clinical bacterial genomics, as well as clinical metagenomics and single-cell genomics, are also discussed in this extensive review [2]. In the second review article, Sentausa and Fournier describe the advantages of genomics for bacterial taxonomy in terms of the quality and quantity of information that can be derived from genome sequences [4]. However, it is important to: (i) define which genes should be use to infer consensus trees that are taxonomically reliable; and (ii) precisely assess the added value of indexes such as the average nucleotide identity in comparison with the current reference standard, DNA–DNA hybridization. As outlined by Sentausa and Fournier, the exponentially increasing number of genomes will ease the use of the average nucleotide identity and other whole taxogenomic indicators to assign a given species [4]. Nevertheless, despite the amount of information provided by genome sequencing, bacterial genomics should clearly not be considered as providing complete strain characterization, and taxonomy should still rely on some complementary phenotypic tests. In the article by Price et al., the different applications of genomics in clinical microbiology are described by using Staphylococcus aureus as an example [5]. For this pathogen, genomics is very promising, given the large number of S. aureus genomes available for comparison, and will probably provide rapid simulatneous access to data on the presence of virulence factors such as superantigen-encoding genes, on the presence of mecA genes, and on the genotype. Information on the last of these has already been derived from genomic sequences to infer epidemiological relationships between different S. aureus carriers and to analyse methicillin-resistant S. aureus outbreaks [6,7]. However, for S. aureus, such outbreak investigations have, to date, been only retrospective, owing to the amount of data to be analysed and the current lack of automated bioinformatic pipelines allowing rapid inference of possible transmission events. The investigation of outbreaks by the use of bacterial genome sequencing is also reviewed extensively in the fourth article of this special issue [8], with Mycobacterium tuberculosis as an example. Whole genome sequencing showed better resolution than all typing approaches previously used by mycobacteriologists, including the more recent mycobacterial interspersed repetitive-unit variable-number tandem repeat typing approach [9,10]. In the review article by Walker et al., the usefulness of single-nucleotide polymorphisms to infer the directionality of transmission events within a given outbreak is nicely presented, allowing readers who are unaccustomed with this field to become familiar with the ‘evolution by descent’ paradigm [10]. Finally, after these four review articles, infectious disease specialists and clinical microbiologists will be highly interested