Identification performance of MALDI-ToF-MS upon mono- and bi-microbial cultures is cell number and culture proportion dependent

Abstract

Biotyping using matrix-assisted laser desorption ionization-time of flight (MALDI-ToF) mass spectroscopy (MS) has revolutionized microbiology by allowing clinicians and scientists to rapidly identify microbes at genus and species levels. The present study extensively assesses the suitability and reliability of MALDI-ToF biotyping of 14 different aerobic and anaerobic bacterial species as pure and mixed cultures. Reliable identification at species level was possible from biomaterial of older colonies and even frozen biomaterial, although this was species dependent. Using standard instrument settings and direct application of biomaterial onto the MALDI-ToF target plates, it was determined that the cell densities necessary for completely reliable identification of pure cultures varied between 2.40 × 10⁸ and 1.10 × 10¹⁰ viable cell counts (VCCs) per mL, depending on the species. Evaluation of the mixed culture algorithm of the Bruker Biotyper^® software showed that the performance of the algorithm depends greatly on the targeted species, on their phylogenetic distance, and on their ratio of VCC per mL in the mixed culture. Hence, the use of MALDI-ToF-MS with incorporation of the mixed culture algorithm of the software is a useful pre-screening tool for early identification of contaminants, but due to the great variability in performance between different species and the usually unknown percentage of the possible contaminant in the mixture, it is advisable to combine this method with other microbiology methods.

Keywords: Biotyping; MALDI-ToF; Mixed culture; Viable cell count.

Fig. 1

(a) Number of Identifications with an ID score > 2.0 (n = 24) at different concentrations of VCC per milliliter (in logarithmic scale). Number of Identifications with an ID score > 1.7 and > 2.0 (n = 24) of strain (b) E. faecalis, (c) C. sporogenes, (d) E. gallinarum, (e) B. breve

Fig. 2

(a) Concentration (VCC per milliliter) for a reliable identification (ID score > 1.7 and > 2.0, n = 24 successful identifications) of bacteria calculated with a linear regression model. The standard error of the predicted value was determined by a parametric bootstrap estimation (10,000 re-samplings). Linear regression models were calculated using the linear part of the whole range of concentrations measured counting all samples with a reliable ID score (> 1.7 or > 2.0, n = 24, # n = 20, * n = 16). Graph shows E. gallinarum (b) and C. sporogenes (c) All concentrations with trend line of the linear range and 95% CI

Fig. 3

Global success rate (Gsr) of combinations containing two different species. Different ratios of two bacteria (n = 6) were counted, if positively identified as mixed culture (a) within a mixR ± 0.48, (b) within a mixR ± 0.95, (c) all measured ratios without constrictions. Example graphs of representative combinations are highlighted by a “+” in the heatmaps and displayed on the left. Grey area displays the range used to calculate Gsr values of the respective heatmap, which is defined by the minimum number of mixed culture IDs (> 4 of 6) and the range of mixR