Characterization of Clinically Relevant Fungi via SERS Fingerprinting Assisted by Novel Chemometric Models

Nonculture-based tests are gaining popularity and upsurge in the diagnosis of invasive fungal infections (IFI) fostered by their main asset, the reduced analysis time, which enables a more rapid diagnosis. In this project, three different clinical isolates of relevant filamentous fungal species were discriminated by using a rapid (less than 5 min) and sensitive surface-enhanced Raman scattering (SERS)-based detection method, assisted by chemometrics. The holistic evaluation of the SERS spectra was performed by employing appropriate chemometric tools-classical and fuzzy principal component analysis (FPCA) in combination with linear discriminant analysis (LDA) applied to the first relevant principal components. The efficiency of the proposed robust algorithm is illustrated on the data set including three fungal isolates (Aspergillus fumigatus sensu stricto, cryptic A. fumigatus complex species, and Rhizomucor pusillus) that were isolated from patient materials. The accurate and reliable discrimination between species of common fungal pathogen strains suggest that the developed method has the potential as an alternative, spectroscopic-based routine analysis tool in IFI diagnosis.