Bacteria detection and species identification at the single-cell level using super-resolution fluorescence imaging and AI analysis

Abstract

The skin microbiome is thought to play a critical role in maintaining skin health and protecting against infection. While most microorganisms that live on the skin are harmless or even beneficial, some can cause skin infections or other health problems, emphasizing the importance of diagnosis of the composition and diversity of the skin flora. However, conventional diagnostic methods for evaluation of the skin microbiome are not sensitive enough to detect bacteria at low concentrations and suffer from poor specificity, thus limiting early diagnosis of bacterial infections. In this study, we developed novel approaches for bacterial species detection and identification methods with single-cell sensitivity using super-resolution microscopy and AI-based image analysis: a protein quantification-based method and an AI-based bacterial image analysis method. We demonstrate that these methods can differentiate between common bacterial members of the skin flora, including Staphylococcus aureus and Staphylococcus epidermidis, and different ribotypes of Cutibacterium acnes, both in purified bacterial samples and in scaling skin samples. The advantages of these methods, including the lack of time-consuming amplification or purification steps and single-cell level detection sensitivity, allow early diagnosis of bacterial infections, even from bacterial samples at extremely low concentrations, thus showing promise as a next-generation platform for microbiome detection as single-cell diagnostics.

Keywords: Bacteria classification; Deep learning; Microbiome; Protein quantification; Super-resolution fluorescence microscopy.