Development of high-throughput nanoscale vesicle analysis at the single-particle level using super-resolution fluorescence microscopy

다국어 초록 (Multilingual Abstract)

The analysis of membrane vesicles at the nanoscale level is crucial for advancing the understanding of intercellular communication and its implications for health and disease. Despite their significance, the nanoscale analysis of vesicles faces challenges owing to their small size and the complexity of biological fluids. This study introduces a novel approach that employs artificial intelligence (AI) to enhance the capabilities of super-resolution fluorescence microscopy for high-throughput single-vesicle analysis. By comparing classical clustering methods (K-means, DBSCAN, and SR-Tesseler) with deep-learning-based approaches (YOLO, DETR, Deformable DETR, and Faster R-CNN) for the analysis of super-resolution fluorescence images of exosomes, we identified the Deformable DETR algorithm as the most effective, with superior accuracy and a reduced processing time. Our findings demonstrate that AI-enhanced methods significantly outperform traditional clustering techniques in identifying individual vesicles and resolving the challenges related to misidentification and computational demands. Moreover, the application of the combined Deformable DETR and ConvNeXt-S algorithms to differently labeled exosomes revealed its capability to differentiate between them, indicating its potential to dissect the heterogeneity of vesicle populations. This breakthrough in vesicle analysis suggests a paradigm shift towards the integration of AI into super-resolution imaging, which is promising for unlocking new frontiers in vesicle biology, disease diagnostics, and the development of vesicle-based therapeutics.

목차 (Table of Contents)

• Abstract
• Chapter 1. Background
• 1.1. Introduction
• 1.2. Conventional light microscopy with diffraction limitation
• 1.3. Development of super-resolution fluorescence microscopy.
• 1.3.1 Patterned illumination methods
• 1.3.2 Single molecule localization microscopy (SMLM)
• 1.4. Stochastic Optical Reconstruction Microscopy (STORM)
• 1.4.1. Three-dimensional STORM
• 1.4.2. Multi-color STORM
• Chapter 2. Development of high-throughput nanoscale vesicle
• analysis at the single-particle level using super-resolution
• fluorescence microscopy
• 2.1. Introduction
• 2.2. Materials and methods
• 2.2.1. Sample preparation for exosome imaging
• 2.2.2. STORM imaging
• 2.2.3. Identification of exosomes from STORM images using
• classical methods
• 2.2.4. Identification of exosomes from STORM images using deep-
• learning-based methods
• 2.2.5. Classification of exosomes from STORM images using deep-
• learning-based methods
• 2.2.6. Quantitative analysis of identified exosomes from STORM
• images.
• 2.3. Results and discussion
• 2.3.1. Classical clustering analysis methods for super-resolution
• images of exosomes
• 2.3.2. Quantitative evaluation of classical clustering analysis
• methods.
• 2.3.3. Evaluation of deep-learning-based clustering analysis
• methods for super-resolution images of exosomes
• 2.3.4. Quantitative evaluation of deep-learning-based clustering
• analysis methods
• 2.3.5. Application of optimized deep-learning-based cluster
• analysis method to various exosome samples
• 2.4. Conclusion
• References
• 국문 요지
• 감사의 글