Abstract
Introduction. Klebsiella pneumoniae, Klebsiella quasipneumoniae and Klebsiella variicola are closely related species (collectively referred to as 'PQV') but exhibit distinct clinical presentations and epidemiological profiles. These distinctions are important for proper diagnosis and treatment of infections and other Klebsiella-related diseases. The ability to differentiate between these species is becoming more important as we better understand their unique roles in disease.Hypothesis/Gap statement. Despite the importance of accurate taxonomic classifications, both traditional and modern laboratory techniques fail to accurately delineate between the PQV species, which can lead to treatment failures. Conversely, taxonomy via whole-genome sequencing is highly accurate but is both costly and resource-intensive. Thus, there is a lack of a widely adopted method that balances both cost-effectiveness and accuracy for differentiating PQV species.Aim. To assess the taxonomic accuracy of existing genome databases and classification tools for PQV species and to develop a rapid, cost-effective method for accurate species differentiation.Methodology. To address these challenges, we extracted 78 representative PQV genomes from the Integrated Microbial Genomes and Microbiomes database. We used multiple comparative genomic comparison techniques, phylogenetic tree constructions and pangenome profiling to accurately phylogenetically and subsequently taxonomically classify the genomes. After establishing accurate taxonomic classifications, we identified species-specific marker genes (SSMGs) represented by KEGG Orthologies (KOs).Results. The Genome Taxonomy Database (GTDB) classifications were consistent with our comparative genomic benchmarks, while the National Center for Biotechnology Informationdatabase originally contained six misclassified genomes. The GTDB-Tk tool also showed reliability for systematic classification of the genomes. Our KO-based screening identified 22 candidate SSMGs. Four markers (K05306, K07507, K13795 and K09955) exhibited significant specificity. ATP-binding proteins showed slightly higher maximum percentage identity values due to conserved domains but were still valuable within multi-locus SSMG panels.Conclusion. Our findings establish the GTDB as the gold standard taxonomic reference for PQV classification when complete genomes are available. Additionally, we developed a practical panel of genetic markers that enables rapid, cost-effective and accurate species differentiation. These SSMGs represent practical tools that can be implemented in diagnostic laboratories for both clinical specimens and environmental surveillance, addressing a critical gap in clinical microbiology.