AI 'Self-Driving' Microscope Predicts and Analyzes Protein Aggregation
July 24, 2025
EPFL researchers created an AI-driven microscopy system that predicts and analyzes misfolded protein aggregation in real time.
Key Details
- Developed a self-driving imaging system combining multiple microscopy methods and deep learning.
- System predicts and detects protein aggregation—a hallmark of neurodegenerative diseases—in living cells.
- Uses label-free microscopy to minimize sample alteration and maximize imaging efficiency.
- Upon aggregation detection, system triggers Brillouin microscope to analyze biomechanical properties of aggregates.
- Aggregation onset detection achieved 91% accuracy using a specialized deep learning algorithm.
- Published in Nature Communications, with potential impact on drug discovery and precision medicine.
Why It Matters
This work showcases how AI-driven, label-free imaging can reveal dynamic, disease-relevant cellular processes in unprecedented detail, opening new pathways for diagnostic research and drug development in neurodegenerative diseases.