Deep Learning-Enabled Rapid Metabolic Decoding of Small Extracellular Vesicles via Dual-Use Mass Spectroscopy Chip...

Area of Science:

• Biochemistry
• Biotechnology
• Medical Diagnostics

Background:

• Small extracellular vesicles (sEVs) are increasingly utilized in liquid biopsy for disease detection.
• Current sEV isolation and analysis methods require significant improvements in speed, efficiency, and data handling.
• There is a need for high-throughput technologies to support the vigorous development of sEV-based liquid biopsy.

Purpose of the Study:

• To develop a high-throughput dual-use mass spectroscopic chip array (DUMSCA) for rapid isolation and detection of plasma sEVs.
• To evaluate DUMSCA's performance in terms of speed, storage, reuse, desorption/ionization efficiency, and metabolite quantification.
• To demonstrate the utility of DUMSCA-derived metabolic data for disease diagnosis using deep learning.

Main Methods:

• Development of a high-throughput dual-use mass spectroscopic chip array (DUMSCA).
• Isolation and detection of plasma sEVs using DUMSCA.
• Metabolite quantification and generation of a metabolic data matrix from sEVs.
• Application of a deep learning model for disease diagnosis (Crohn's disease).
• Biomarker discovery using feature sparsification and tandem mass spectrometry.

Main Results:

• DUMSCA achieved over a 50% increase in speed compared to traditional methods.
• DUMSCA demonstrated proficiency in robust storage, reuse, high-efficiency desorption/ionization, and metabolite quantification.
• A deep learning model utilizing the sEV metabolic data achieved high-performance diagnosis of Crohn's disease.
• Discovered biomarkers showed remarkable diagnostic performance.

Conclusions:

• DUMSCA offers a rapid and valid approach for disease diagnosis through sEV analysis.
• This technology enables disease diagnosis without requiring prior knowledge of specific biomarkers.
• DUMSCA provides a high-throughput platform that will significantly advance sEV-based liquid biopsy.