Background: Group B streptococcus (GBS) capsular polysaccharide is one of the major virulence factors underlying invasive GBS disease and a component of forthcoming vaccines. Serotype classification of GBS is based on the capsule polysaccharide of which ten variants are known to exist (Ia, Ib, II-IX). Current methods for GBS serotype assignment rely on latex agglutination or PCR while more recently a whole genome sequencing method was reported. In this study, three distinct algorithms for serotype assignment from genomic data were assessed using a panel of 790 clinical isolates.
Methods: The first approach utilised the entire capsular locus coupled with a mapping methodology. The second approach continues from the first and utilised a SNP-based methodology across the conserved cpsD-G region to differentiate serotypes Ia-VII and IX. Finally the third approach used the variable cpsG -K region coupled with a mapping methodology. All three approaches were assessed for typeability (percentage of isolates assigned a serotype) and concordance to the latex agglutination methodology.
Results: Following comparisons, the third approach using the variable cpsG-K region demonstrated the best performance with 99.9% typeability and 86.7% concordance. Overall, of the 105 discordant isolates, 71 were resolved following retesting of latex agglutination and whole genome sequencing, 20 failed to assign a serotype using latex agglutination and only 14 were found to be truly discordant on re-testing. Comparison of this final approach with the previously described assembly-based approach returned identical results.
Conclusions: These results demonstrated that molecular capsular typing using whole genome sequencing and a mapping-based approach is a viable alternative to the traditional, latex agglutination-based serotyping method and can be implemented in a public health microbiology setting.
Keywords: Group B Streptococci; Serotyping; Streptococcus agalactiae; Whole genome sequencing.