Comparative genomics of a drug-resistant Pseudomonas aeruginosa panel and the challenges of antimicrobial resistance prediction from genomes

• Published in: FEMS microbiology letters 2017-10, Vol.364 (18)
• Main Authors: Jeukens, J., Kukavica-Ibrulj, I., Emond-Rheault, J. G., Freschi, L., Levesque, R. C.
• Format: Article
• Language: English
• Subjects: Antibiotic resistance; Antibiotics; Antiinfectives and antibacterials; Antimicrobial agents; Antimicrobial resistance; Drug resistance; Drug Resistance, Multiple, Bacterial - genetics; Gene sequencing; Genes, MDR; Genomes; Genomics; Genomics - methods; Genotype; Genotypes; Gram-negative bacteria; Humans; Metabolomics; Microbiology; Microorganisms; Multidrug resistance; Mutation; Opportunist infection; Pathogens; Phenotypes; Predictions; Proteomics; Pseudomonas aeruginosa; Pseudomonas aeruginosa - genetics; Pseudomonas Infections - microbiology; Strains (organisms); Whole Genome Sequencing - methods
• ISSN: 1574-6968; 0378-1097; 1574-6968
• DOI: 10.1093/femsle/fnx161

Abstract Antimicrobial resistance (AMR) is now recognized as a global threat to human health. The accessibility of microbial whole-genome sequencing offers an invaluable opportunity for resistance surveillance via the resistome, i.e. the genes and mutations underlying AMR. Unfortunately, AMR prediction from genomic data remains extremely challenging, especially for species with a large pan-genome. One such organism, for which multidrug-resistant (MDR) isolates are frequently encountered in the clinic, is Pseudomonas aeruginosa. This study focuses on a commercially available panel of seven MDR P. aeruginosa strains. The main goals were to sequence and compare these strains’ genomes, attempt to predict AMR from whole genomes using two different methods and determine whether this panel could be an informative complement to the international P. aeruginosa reference panel. As expected, the results highlight the complexity of associating genotype and AMR phenotype in P. aeruginosa, mainly due to the intricate regulation of resistance mechanisms. Our results also urge caution in the interpretation of predicted resistomes regarding the occurrence of gene identity discrepancies between strains. We envision that, in addition to accounting for the genomic diversity of P. aeruginosa, future development of predictive tools will need to incorporate a transcriptomic, proteomic and/or metabolomic component. The challenges of predicting antibiotic resistance from genomes in opportunistic pathogen Pseudomonas aeruginosa.