In a study published in Journal of Biophotonics, a research group led by Prof. LI Bei and his Doctoral student LIU Bo from Changchun Institute of Optics, Fine Mechanics and Physics (CIOMP) of the Chinese Academy of Sciences (CAS) proposed a novel method for classification of pathogenic bacteria, which combines Raman spectroscopy, Variational Auto‐encoder (VAE) and Long Short-Term Memory network (LSTM).

Rapid and accurate pathogen identification is the key to precise anti infection treatment. Biochemical and mass spectrometry identification is the most prevalent approach for pathogen identification. These methods based on culture is time-consuming, it is not suitable for rapid pathogen screening. As a result, it's necessary to develop a simple, rapid and label-free pathogen detection technology for effectively screening of pathogens, drug resistance analysis, public safety bacterium monitoring, and food safety inspection.

Experimental validations show that this method enhances machine learning classification accuracy while also reducing the demand for a considerable amount of training data. The clinical samples were used for verification, and the classification effect was very good.

Raman spectroscopy is a technology that, in contrast to other culture-free approaches (such as fluorescence labeling, magnetic labeling, single-cell sequencing, etc.), can identify bacteria without the need for specialized label creation and is straightforward to adapt to other samples. With just slight adjustments from its use in pathogen identification, this method can be adapted to other spectroscopic techniques (such as mass spectrometry and infrared spectroscopy) and material identification issues.

In the future, new algorithms will continue to be developed to improve the spectral generation speed and accuracy as well as the classification accuracy, and will be combined with sorting techniques (such as optical tweezers technology, laser induced forward transfer technology) to identify and sort single cells.