Comparison of performance of cgMLST to wgMLST and SNP analysis in C. difficile. cgMLST performance in C.difficile

Clostridioides difficile is the most common cause of antibiotic-associated gastrointestinal infections. Capillary-electrophoresis (CE)-PCR ribotyping is currently the gold standard for C. difficile typing, but lacks discriminatory power to study transmission and outbreaks in detail. New molecular methods have the capacity to differentiate better, but backward compatibility with CE-PCR ribotyping must be assessed. We investigated whether core genome multilocus sequence typing (cgMLST) was able to predict CE-PCR ribotypes. We also compared the discriminatory power of cgMLST (SeqSphere & EnteroBase) and whole genome MLST (wgMLST) (EnteroBase) with single nucleotide polymorphism (SNP) analysis, using a well-characterized collection of diverse strains (N=630; 100 unique ribotypes [RTs]). Mean intra-RT allele differences were calculated to assess the diversity within a ribotype. Discriminatory power of cg/wgMLST in comparison with SNP analysis (threshold: ≤6 allelic differences or SNPs) was assessed in two outbreak settings (RT078 & RT181). A unique cgMLST profile (>6 allele differences) was observed in 82/100 ribotypes, indicating sufficient backward compatibility. Intra-RT allele difference varied per ribotype and MLST clade. Application of cg/wgMLST and SNP analysis in two outbreak settings with ribotypes with a low intra-ribotype allele difference showed no distinction between outbreak- and non-outbreak strains. We conclude that cgMLST has the potential to be an alternative to CE-PCR ribotyping. The method is reproducible, easy to standardize and offers higher discrimination. However, in some monomorphic ribotypes adjusted cut-off thresholds and epidemiological data are necessary to resolve outbreaks. We propose to decrease the current threshold to 3 alleles resulting in better identified outbreaks without excluding outbreak strains.