Exploring the value of MALDI-TOF MS for the detection of clonal outbreaks of B. contaminans

Abstract

Background: Microbial diagnostics aims to identify the causative bacterial, fungal and parasitic agents of infectious diseases. Usually this is a lengthy process relying biochemical profiling of metabolic traits. Although molecular genetics has risen in both output and affordability to become the gold standard in diagnosis, it is not yet available for most routine clinical microbiology due to cost and the level of skill it requires. MALDI-TOF MS approaches may be useful in bridging the gap between low resolution phenotypic methods and bulky genotypic methods in the goal of epidemiological source-typing of microbes. Burkholderia cepacia complex has been shown to be identifiable at the species level using MALDI-TOF MS by the optimization of the database by adding reference Mass Spectra Profiles (MSPs) for B. contaminans on its commercial database. There have not yet been studies assessing the ability of MALDI-TOF MS to source-type Burkholderia contaminans isolates into epidemiologically relevant outbreak clusters Therefore, the aim of the present study was to compare the performance of MALDI-TOF MS as a typing method with the gold standard method pulsed-field gel electrophoresis (PFGE) Methods & Materials: 55 well characterised B. contaminans isolates were used to create a panel for analysis of MALDI-TOF MS biomarker peaks and their relation to outbreak strains, location, diagnosis and typeability. Previously all isolates were subjected to PFGE of XbaI-digested genomic DNA.The macrorestriction profiles of strains were distinct and classified into different clades on the basis of epidemiological information available Classification models were generated using biostatistical analysis software ClinProTools 3.1. Three algorithms were available: Genetic Algorithm (GA), Supervised Neural Network (SNN) using 3-nearest neighbours and the QuickClassifier (QC). Results: B. contaminans spectra derived from MALDI-TOF MS were of sufficiently high resolution to identify 100% of isolates. Classification algorithms showed promise for discrimination, Genetic Algorithm models showing 100% recognition capability for all outbreak models and the typeable vs non-typeable model, and 96.63% recognition for the location model. Conclusion: MALDI-TOF MS shows promise in the discrimination of B. contaminans isolates into epidemiological clusters. Further work should investigate this capability, and include peptide studies and genomic sequencing to identify individual proteins or genes responsible for this.