Image created by Dr. Michael J. Miller
Aerosols play a critical role in the transmission of airborne pathogens by acting as carriers that transport pathogens between the environment and humans. Timely and accurate detection of these pathogens is essential for containing infectious disease outbreaks at their source and reducing hospital-acquired infection rates.
Traditional pathogen detection methods face several challenges, including slow detection speeds, low sensitivity, and complicated procedures.
In contrast, single-cell Raman spectroscopy has emerged as a promising technology for rapid pathogen identification. This method characterizes the phenotypic features of microorganisms and offers advantages such as speed, culture-free operation, and multi-target detection.
Recent advancements in artificial intelligence (AI) have significantly enhanced the resolution of Raman spectroscopy by improving its ability to distinguish subtle spectral variations between bacteria.
However, in real-world environments with extensive microbial diversity and many unculturable species, establishing comprehensive Raman spectral databases for all environmental microorganisms remains a considerable challenge. Fast and accurate identification of pathogens within complex airborne microbial communities is particularly difficult due to the high complexity of these environments.
To address this challenge, a research team led by Prof. Zhu Yongguan, CAS member, and Prof. Cui Li from the Institute of Urban Environment of the Chinese Academy of Sciences (IUE, CAS) developed and published an innovative airborne pathogen detection technology.
This technology, which combines single-cell Raman spectroscopy with an open-set deep learning algorithm, was reported in Science Advances.
The researchers developed a Raman spectral database for detecting pathogens in aerosols. They incorporated an open-set loss function and defined optimization thresholds within a deep learning algorithm. This innovative approach enables the identification of five key airborne pathogens in complex, real-world environments.
For air samples with pathogen abundances exceeding 1%, the entire process—including sample collection, preprocessing, identification, and reporting—takes just one hour. The system achieves an average identification accuracy of 93% for the five target pathogens and reduces false positive rates by 36% compared to traditional closed-set algorithms.
Its detection sensitivity is capable of identifying pathogens at the single-cell level. Additionally, this method can effectively target pathogens in real-world air samples that contain over 4,600 microbial species, showing significant resistance to interference when compared to traditional methods.
The method has been validated in various real-world settings, including hospitals, shopping malls, dining halls, kitchen waste plants, microbiology laboratories, and public restrooms.
These validations demonstrated its effectiveness in addressing the challenges of slow detection and inadequate identification associated with traditional techniques. By enabling efficient monitoring of environmental biosafety and providing early warnings, this approach plays a vital role in preventing airborne pathogen transmission.
More information: Longji Zhu et al, Open-set deep learning–enabled single-cell Raman spectroscopy for rapid identification of airborne pathogens in real-world environments, Science Advances (2025). DOI: 10.1126/sciadv.adp7991
Abstract
Pathogenic bioaerosols are critical for outbreaks of airborne disease; however, rapidly and accurately identifying pathogens directly from complex air environments remains highly challenging. We present an advanced method that combines open-set deep learning (OSDL) with single-cell Raman spectroscopy to identify pathogens in real-world air containing diverse unknown indigenous bacteria that cannot be fully included in training sets. To test and further enhance identification, we constructed the Raman datasets of aerosolized bacteria. Through optimizing OSDL algorithms and training strategies, Raman-OSDL achieves 93% accuracy for five target airborne pathogens, 84% accuracy for untrained air bacteria, and 36% reduction in false positive rates compared to conventional close-set algorithms. It offers a high detection sensitivity down to 1:1000. When applied to real air containing >4600 bacterial species, our method accurately identifies single or multiple pathogens simultaneously within an hour. This single-cell tool advances rapidly surveilling pathogens in complex environments to prevent infection transmission.