A Proof-of-Concept Study of Near-Infrared Spectroscopy and DD-SIMCA for Candida auris Strain Authentication

Candida auris is an emerging fungal pathogen associated with healthcare outbreaks, invasive infections, environmental persistence, and antifungal resistance. Rapid strain characterization may support epidemiological surveillance, but conventional identification workflows can require specialized instrumentation and updated databases. This proof-of-concept study evaluated near-infrared (NIR) spectroscopy combined with Data-Driven Soft Independent Modeling of Class Analogy (DD-SIMCA) for authenticating a reference strain representing Clade I. Four reference strains representing Clades I–IV were cultured under standardized conditions, and 200 diffuse-reflectance spectra were acquired from 900 to 2600 nm. The Clade I model yielded sensitivities of 91.43% for the training set and 100% for the test set under spectrum-level partitioning. To assess performance using independent culture plates, a plate-separated analysis was performed using seven plates for training and three separate plates for testing, yielding test sensitivities ranging from 46.7% to 80%. Specificity was 100% for the strains representing Clades III and IV and 64% for the strain representing Clade II under spectrum-level partitioning, whereas plate-separated specificity for the Clade II representative ranged from 42% to 54%. These findings are consistent with greater spectral overlap between the Clade I and II representatives and support the analytical feasibility of NIR/DD-SIMCA for rapid, culture-based screening. Because only one reference strain represented each clade, validation with multiple independent isolates, culture batches, instruments, and laboratories is required before clade-level or clinical use.