Abstract
Klebsiella pneumoniae is a primary bacterial pathogen responsible for various human infections with the emergence of carbapenem-resistant K. pneumoniae (CRKP) as a significant clinical challenge. Although previous studies have classified K. pneumoniae based on homology of conserved genes, the biological implications of these classifications remain unclear. In this study, we extracted 47,222 centroid coding sequences (CDSs), representative sequences selected from clusters of highly similar CDSs, from 3,592 complete genomes and analysed the genomic characteristics of 4,458 CRKP strains using these centroid CDSs. This approach revealed substantial genomic diversity among CRKP strains, with classification results largely aligning with the multilocus sequence typing (MLST) method. Associations between centroid CDSs and MLST sequence types (STs) were identified, and the biological functions enriched in each ST including those linked to resistance and virulence mechanisms were highlighted. Moreover, the coexistence of CDSs annotated as antibiotic resistance and virulence genes was observed, particularly between a group of carbapenem resistance genes and those encoding proteins for type I and III fimbriae, efflux pumps, type II secretion system and siderophore-mediated iron acquisition. These findings provide an alternative approach for classifying CRKP strains, offering insights into their genomic characteristics.