Multilevel genome typing: genomics-guided scalable resolution typing of microbial pathogens

Abstract

Background Both long- and short-term epidemiology are fundamental to disease control and require accurate bacterial typing. Genomic data resulting from implementation of whole genome sequencing in many public health laboratories can potentially provide highly sensitive and accurate descriptions of strain relatedness. Previous typing efforts using these data have mainly focussed on outbreak detection. Aim We aimed to develop multilevel genome typing (MGT), using consecutive multilocus sequence typing (MLST) schemes of increasing sizes, stepping up from seven-gene MLST to core genome MLST, to allow examination of genetic relatedness at multiple resolution levels. Methods The system was applied to Salmonella enterica serovar Typhimurium. The MLST scheme used at each step (MGT level), defined a given MGT-level specific sequence type (ST). The list of STs generated from all of these increasing MGT levels, was named a genome type (GT). Using MGT, we typed 9,096 previously characterised isolates with publicly available data. Results Our approach could identify previously described S. Typhimurium populations, such as the DT104 multidrug resistance lineage (GT 19-2-11) and two invasive lineages of African isolates (GT 313-2-3 and 313-2-752). Further, we showed that MGT-derived clusters can accurately distinguish five outbreaks from each other and five background isolates. Conclusion MGT provides a universal and stable nomenclature at multiple resolutions for S. Typhimurium strains and could be implemented as an internationally standardised strain identification system. While established so far only for S. Typhimurium, the results here suggest that MGT could form the basis for typing systems in other similar microorganisms.

Elucidation of global and local genomic epidemiology of Salmonella enterica serovar Enteritidis through multilevel genome typing

“…Both SNP address and HierCC assign a series of cluster types for each isolate based on a range of genetic distance thresholds. These cluster types, which are generated based on the single-linkage clustering, can merge when additional data is added [14]. Thus, cluster typing systems may not be stable when new isolates are added, which is problematic for long term surveillance, especially for large databases in which the number of isolates is expanding quickly.…”

“…Thus, cluster typing systems may not be stable when new isolates are added, which is problematic for long term surveillance, especially for large databases in which the number of isolates is expanding quickly. In contrast, sequence types (STs) are based on exact matching of genes between isolates, and thus are stable, so that once a ST is assigned to an isolate it will never change [14]. It would be useful to develop a surveillance system where each isolate is assigned with both stable STs and cluster types at varied subtyping resolution.…”

“…Multilevel genome typing (MGT) is a novel genome sequence based typing method developed by Payne et al [14]. Based on the well-established multi locus sequence typing (MLST) concept [15,16], MGT is defined by multiple MLST schemes containing different numbers of loci which result in different resolutions.…”

“…Based on the well-established multi locus sequence typing (MLST) concept [15,16], MGT is defined by multiple MLST schemes containing different numbers of loci which result in different resolutions. An MGT scheme and a public database were established for Salmonella Typhimurium where 7 to loci were included in MGT1 to 9 [14]. MGT1 corresponds to the classic 7-gene MLST scheme of S. enterica [16].…”

“…MGT1 corresponds to the classic 7-gene MLST scheme of S. enterica [16]. For each isolate, sequence types (STs) and clonal complexes (CCs) were called at each MGT level [14]. Thus, the subtyping resolution was gradually increased, which allowed isolates to be grouped together at different scales [14].…”