With the increase in mortality caused by pathogens worldwide and the subsequent serious drug resistance owing to the abuse of antibiotics, there is an urgent need to develop versatile analytical techniques to address this public issue. Vibrational spectroscopy, such as infrared (IR) or Raman spectroscopy, is a rapid, noninvasive, nondestructive, real-time, low-cost, and user-friendly technique that has recently gained considerable attention. In particular, surface-enhanced Raman spectroscopy (SERS) can provide a highly sensitive readout for bio-detection with ultralow or even trace content. Nevertheless, extra attachment cost, nonaqueous acquisition, and low reproducibility require the conventional SERS (C-SERS) to further optimize the conditions. The emergence of dynamic SERS (D-SERS) sheds light on C-SERS because of the dispensable substrate design, superior enhancement and stability of Raman signals, and solvent protection. The powerful sensitivity enables D-SERS to perform only with a portable Raman spectrometer with moderate spatial resolution and precision. Moreover, the assistance of machine learning methods, such as principal component analysis (PCA), further broadens its research depth through data mining of the information within the spectra. Therefore, in this study, D-SERS, a portable Raman spectrometer, and PCA were used to determine the phenotypic variations of fungal cells Candida albicans (C. albicans) under the influence of different antifungals with various mechanisms, and unknown antifungals were predicted using the established PCA model. We hope that the proposed technique will become a promising candidate for finding and screening new drugs in the future.
Comparative Analysis of Machine Learning Algorithms on Surface Enhanced Raman Spectra of Clinical Staphylococcus Species
