Application of Raman spectroscopy and machine learning for Candida auris identification and characterization.

Area of Science:

• Medical Mycology
• Spectroscopy
• Machine Learning

Background:

• Candida auris is a multidrug-resistant fungal pathogen causing severe nosocomial infections globally.
• Accurate and rapid identification and characterization of Candida auris are critical challenges in clinical settings.
• Existing methods for Candida auris detection and resistance profiling are often time-consuming and lack single-cell resolution.

Purpose of the Study:

• To develop a novel, rapid, and precise method for identifying Candida auris and its clades using Raman spectroscopy and machine learning.
• To predict antifungal resistance (fluconazole, amphotericin B) and key virulence factors (aggregating, filamentous cells) of Candida auris at the single-cell level.
• To establish a proof-of-concept for a valuable medical diagnostic tool to combat Candida auris infections.

Main Methods:

• Employing Raman spectroscopy for single-cell analysis of Candida auris isolates.
• Utilizing machine learning algorithms for species identification, clade differentiation, and antifungal resistance prediction.
• Developing predictive models for Candida auris phenotypic characteristics linked to virulence.

Main Results:

• Achieved an average identification accuracy of 93.33% across Candida species, with 98% accuracy for clinically simulated samples.
• Demonstrated high accuracy in predicting drug susceptibility: 99% for fluconazole and 94% for amphotericin B.
• Successfully predicted phenotypic characteristics with 100% accuracy for aggregating cells and 97% for filamentous cells.

Conclusions:

• Raman spectroscopy combined with machine learning offers a precise and rapid method for identifying Candida auris at the clade-specific level.
• This approach effectively predicts antifungal resistance and key virulence factors, addressing critical diagnostic needs.
• The developed methodology shows significant promise as a future medical diagnostic tool for managing multidrug-resistant Candida auris infections.