Applying Machine Learning with Localized Surface Plasmon Resonance Sensors to Detect SARS-CoV-2 Particles

Area of Science:

• Nanotechnology
• Biomedical Engineering
• Artificial Intelligence

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
• Current methods like PCR and immunoassays have limitations in speed, cost, or convenience.
• A need exists for direct virus detection without extensive sample preparation.

Purpose of the Study:

• To develop a novel, rapid, and accurate method for direct SARS-CoV-2 detection.
• To integrate localized surface plasmon resonance (LSPR) sensors, optical imaging, and AI for virus detection.
• To achieve qualitative and quantitative detection of SARS-CoV-2 particles.

Main Methods:

• Utilized localized surface plasmon resonance (LSPR) sensors for optical detection.
• Integrated advanced optical imaging techniques.
• Applied artificial intelligence (AI) and machine learning models for analysis and prediction.
• Developed a workflow for direct virus particle detection without sample preparation.

Main Results:

• Achieved direct detection of SARS-CoV-2 virus particles within 12 minutes.
• Demonstrated qualitative and quantitative detection in the range of 125.28 to 10^6 viral particles (vp)/mL.
• Established a limit of detection (LOD) of 100 vp/mL.
• Attained over 97% accuracy in SARS-CoV-2 positive/negative assessment.
• Developed a regression machine learning model with R2 > 0.95 for virus concentration prediction.

Conclusions:

• The developed LSPR-based method offers a rapid, accurate, and convenient approach for SARS-CoV-2 detection.
• This integrated system demonstrates significant potential as a low-cost diagnostic tool for COVID-19.
• The AI-driven analysis enhances the reliability and predictive power of the diagnostic method.