Abstract
Klebsiella species, including Klebsiella pneumoniae and Klebsiella quasipneumoniae, present significant challenges in clinical microbiology due to their genetic similarity, which complicates accurate species identification using established methods, including matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) on the protein/peptide level. Although the treatment choice for infections caused by these pathogens is often similar, precise species characterization enhances our epidemiological understanding. While whole-genome sequencing can accurately distinguish Klebsiella species accurately, those analyses are time-consuming, requiring specialized expertise, and are not currently used in routine clinical laboratories. Therefore, developing a timely and accurate pathogen characterization method is essential for effective treatment, management, and infection control measures. This study combined MALDI-TOF MS in negative ion mode with machine learning techniques to identify potential lipid biomarkers as a novel method to distinguish between K. pneumoniae and K. quasipneumoniae. Using this method, we identified discriminative features between the species, with peaks at m/z 2157, m/z 1931, m/z 1964, m/z 2042, and m/z 1407 highlighted as potential biomarkers for species identification. Our findings suggest that the lipid profiles of the species obtained from MALDI-TOF MS can serve as effective biomarkers for distinguishing Klebsiella species. Further research should focus on the structural identification of these biomarkers and expand the data set to include more isolates for each of the species. This approach holds promise for developing more cost-effective and rapid diagnostic tools in clinical microbiology, ultimately improving patient outcomes and infection control.