A machine-learning-integrated portable electrochemiluminescence sensing platform for the visualization and...

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Machine Learning

Background:

• Electrochemiluminescence (ECL)-based point-of-care testing (POCT) offers sensitive and selective disease identification.
• High-throughput ECL devices face challenges with energy consumption and sample cross-contamination.

Purpose of the Study:

• Develop a high-throughput ECL biosensor platform integrated with machine learning for disease screening.
• Address limitations of traditional high-throughput ECL systems, including cross-contamination and signal interference.

Main Methods:

• Fabricated a microcolumn array electrode modified with gold nanoparticles for enhanced ECL signal.
• Utilized a microelectrochemical workstation and smartphone with custom software for data acquisition and analysis.
• Employed Support Vector Machine (SVM) and Backpropagation (BP) neural network algorithms for signal analysis and protein quantification.

Main Results:

• Achieved quantitative detection of SARS-CoV-2 nucleocapsid protein with a linear range of 0.001-10 ng/mL and a limit of detection of 0.86 pg/mL.
• Demonstrated high accuracy in classifying and predicting protein concentration using SVM and BP algorithms with RGB feature extraction.
• Significantly reduced cross-contamination and signal interference compared to traditional high-throughput ECL systems.

Conclusions:

• The developed machine learning-assisted ECL-POCT platform offers a promising solution for large-scale, simultaneous disease screening.
• The integration of advanced algorithms enhances the reliability and efficiency of ECL-based diagnostics.
• This approach has substantial potential for improving infectious disease outbreak surveillance and management.