A proof-of-concept study demonstrates the potential of MALDI-TOF MS for bacterial strain typing across four key pathogens, and reveals that center- and instrument-specific variation can substantially affect performance.
The Extra Trees Classifier consistently achieved the highest average accuracy and F1-score in both Gram type classification and species-level identification, demonstrating superior generalization across datasets.
Georgios Dolias, O. Bragina, Andres Udal et al.· Scientific Reports· 0 citations
A novel multi-label, multi-bacteria classification framework (MLMBC) that simultaneously predicts AMR across multiple bacterial species and antibiotics using MALDI-TOF mass spectrometry data and Convolutional Neural Networks (CNNs).
Leila Aro-Sati, Xaviera A. López-Cortés, José M. Manrıquez Troncoso et al.· Neural computing & applicati...· 0 citations
OXA-48 carbapenemases are among the most widespread and important resistance mechanisms in Enterobacterales. Yet detecting carbapenemases by conventional workflows necessitates additional testing, thus delaying optimization of therapy and implementation of infection control measures. Here, we present a machine learning approach that identifies the conserved pOXA-48 plasmid directly from routine MALDI-TOF spectra acquired for species identification. The model detects pOXA-48 carriers with an AUROC of 0.96–0.98 across two independent hospital cohorts and instrument platforms, indicating near-perfect discrimination. Using bottom-up proteomics, plasmid conjugation, and plasmid curing, we link the discriminative MALDI-TOF spectral features to proteins encoded on pOXA-48, with DUF1496 domain-containing protein producing the most discriminative spectral feature. Our approach reframes the resistance prediction task from inferring a resistance phenotype to detecting a conserved plasmid through its expressed proteomic signature and has the potential to enable rapid MALDI-TOF MS-based diagnostics for a wide range of plasmid-based resistance determinants.
Janko Sattler, J. Mueller-Reif, Dexiong Chen et al.· bioRxiv· 0 citations
The feasibility of portable Raman spectroscopy combined with machine learning for bacterial recognition under standardized sample conditions is supported, and pairwise binary classification showed variable performance among bacterial pairs.
Shisheng Cao, Ran Pang, Yongqiang Chen et al.· Frontiers in Microbiology· 0 citations
Accurate identification of Gram-positive bacteria by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) may be limited by the complex structure of their cell walls. This study evaluated the impact of on-plate formic acid (FA) pretreatment on MALDI-TOF MS identification performance in a routine clinical microbiology laboratory. A total of 148 non-repetitive Gram-positive clinical isolates were analyzed using two preparation methods: direct transfer (DT) and on-plate FA pretreatment. Identification scores were compared with an emphasis on high-confidence species-level identification (log score ≥ 2.0). Statistical analyses were performed using the Wilcoxon signed-rank and McNemar tests. Overall identification rates were significantly higher with FA pretreatment than with DT (99.3% vs. 79.7%). The proportion of isolates achieving species-level identification increased from 27.1% with DT to 76.2% with FA pretreatment. Mean MALDI-TOF MS log scores for all isolates were significantly higher with FA (2.104 ± 0.186) than with DT (1.861 ± 0.265; p < 0.001). Notably, FA pretreatment also resulted in a significant increase in mean log scores among isolates that had already achieved species-level identification using DT (p = 0.001). The most pronounced improvement was observed among Staphylococcus spp., particularly coagulase-negative staphylococci, with complete concordance between DT and FA-assisted species-level identifications. In conclusion, on-plate FA pretreatment significantly enhances identification accuracy and diagnostic confidence for Gram-positive bacteria in routine MALDI-TOF MS workflows. Building on these findings, we propose a pragmatic, score-based, and species-aware diagnostic algorithm in which FA pretreatment is reflexively applied to Gram-positive isolates, yielding a DT log score < 2.0, with primary FA application recommended for staphylococci (particularly coagulase-negative species) and for less commonly encountered Gram-positive organisms.
Abdurahman Gülmez, A. N. Ceylan, Selda Kömeç et al.· OSMANGAZİ JOURNAL OF MEDICIN...· 0 citations
ABSTRACT Rapid and accurate characterization of antimicrobial resistance is essential for effective patient treatment and outcomes. Infection‐causing microorganisms often harbour multi‐drug resistance, requiring multiple tests for identification. Here, we present a multiplex functional assay using liquid atmospheric pressure (LAP) matrix‐assisted laser desorption/ionization (MALDI) as next‐generation MALDI biotyping technology, which can accurately determine antibiotic resistance/susceptibility within three hours from <5 µL of bacterial culture, employing a beta‐lactam antibiotic panel. Strains with common resistance genes, including OXA‐48, KPC‐3, and VIM‐1, as well as susceptible isolates, are easily and reliably classified. Concurrent with multi‐drug testing, the same bacterial sample provides species‐identifying lipid and protein profiles (up to 100% classification accuracy) and characterization through tandem mass spectrometry (MS/MS) protein sequencing, facilitated by LAP‐MALDI's ability to generate multiply charged protein ions. Detection of multiple antibiotics, their degradation products and lipids/proteins with high mass accuracy, along with the possibility of protein sequencing, offers new diagnostic possibilities for clinical microbiology and antimicrobial stewardship that are less probability‐based than conventional MALDI biotyping.
Lily R. Adair, Shabnam Iyer, Ian M. Jones et al.· Advancement of science· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
