Monitoring Amyloidogenesis with a 3D Deep-Learning-Guided Biolaser Imaging Array

Area of Science:

• Biophotonics
• Neuroscience
• Biochemistry

Background:

• Amyloidogenesis is central to neurodegenerative diseases and drug screening.
• Early identification of intermediate protein aggregate states remains a significant challenge.

Purpose of the Study:

• To develop a peptide-encapsulated droplet microlaser for monitoring amyloidogenesis.
• To evaluate the efficacy of anti-amyloid drugs using this novel biosensing platform.
• To employ deep learning for sensitive detection of spectral shifts.

Main Methods:

• Fabrication of peptide-encapsulated droplet microlasers.
• Monitoring lasing wavelength shifts correlated with amyloid peptide folding and nanostructure.
• Development of a 3D deep-learning strategy for analyzing spectral shifts from far-field images.
• Extraction of 1D color and 2D features from laser images for progression monitoring.

Main Results:

• The microlaser's lasing wavelength accurately reflects amyloid peptide folding behaviors and nanostructure.
• A deep-learning model achieved over 95% classification accuracy on training, validation, and test sets.
• The system successfully monitored the progression of amyloidogenesis using microdroplet laser arrays.

Conclusions:

• Deep-learning-empowered peptide microlasers offer a powerful tool for studying protein misfolding.
• This technology demonstrates potential for high-throughput imaging in cavity biosensing applications.
• The developed method provides a sensitive approach for evaluating anti-amyloid drug efficacy.