Genome wide association study in antimicrobial resistance of disease-causing Streptococcus pneumoniae in Singapore

Abstract

Pneumococcal disease caused by the bacterium Streptococcus pneumoniae is responsible for substantial mortality and morbidity worldwide, making it a global public health concern. S. pneumoniae has the exceptional ability to adapt to its surroundings to overcome selection pressures and maintain survival; this includes adaptation in response to host immune defences and antibiotics.
Increasing levels of resistance to commonly used antimicrobials have been identified in many countries and threaten the effectiveness of present and future treatment options. A longitudinal collection of S. pneumoniae isolates in Singapore offered the opportunity to track evolution. The serotypes most associated with invasive pneumococcal disease were serotypes 4, 8, 20 7A, 19A and 3, and many of these serotypes predominate in adult infection; serotype 6B predominates in paediatric infection. Resistance of isolates was highest to cotrimoxazole (63%), erythromycin (58%), tetracycline (58%) and doxycycline (58%).
While basic molecular epidemiology is a cornerstone of this analysis, the collection of ~2000 isolates, accompanied with some basic clinical information and bacterial antibiotic resistance phenotypes, provided the opportunity to perform a genome wide association study (GWAS) on antimicrobial non-susceptibility. High levels of recombination were identified in this dataset therefore the impact of this on population structure was assessed to ensure an accurate correction for the GWAS. A comparison in the ability of Gubbins and ClonalFrameML to identify recombination from this dataset showed ClonalFrameML was more conservative and concise in recombination calls and could process larger numbers of isolates. These findings showed how it was possible to perform a complete recombination analysis on large, diverse datasets. The recombination sites identified by ClonalFrameML were not a major contributor in the population clustering of isolates.
GWAS’ were performed to identify single nucleotide polymorphisms (SNPs) associated with non-susceptibility to penicillin, cotrimoxazole, erythromycin, clindamycin, chloramphenicol, tetracycline and doxycycline. This showed good association with drug resistance, with SNPs identified in genes involved in the peptidoglycan pathway, folate metabolism, protein synthesis and DNA synthesis. Many of these confirmed previous findings in the field. In addition, SNPs were associated with erythromycin resistance in new loci involved in DNA synthesis (dnaG, sigA). Although validation is required, these additional results provide a new opportunity to develop insights into the biological mechanisms that underlie the important clinical outcomes of drug resistance.

More information

Identifiers

ORCID for Stuart C. Clarke: orcid.org/0000-0002-7009-1548

Catalogue record

Export record