Machine Learning-Driven and Smartphone-Based Fluorescence Detection for CRISPR Diagnostic of SARS-CoV-2

Area of Science:

• Biotechnology
• Medical Diagnostics
• Computational Biology

Background:

• Accurate and affordable detection of SARS-CoV-2 is essential for managing COVID-19 transmission.
• Existing diagnostic methods can be costly or time-consuming, hindering widespread implementation.

Purpose of the Study:

• To develop a cost-effective, smartphone-based diagnostic system for SARS-CoV-2 detection.
• To integrate CRISPR diagnostics with machine learning for quantitative viral detection.

Main Methods:

• A 3D-printed, smartphone-coupled device was engineered for fluorescence signal evaluation.
• CRISPR diagnostic assays were used to detect SARS-CoV-2 genetic material.
• Machine learning software was developed to analyze fluorescence images and quantify viral presence.

Main Results:

• The system achieved a limit of detection of 6.25 RNA copies/μL in laboratory samples.
• Clinical evaluation on 96 nasopharyngeal swabs demonstrated 95% accuracy and 97% sensitivity.
• A quantitative fluorescence score correlated strongly with RT-qPCR Ct values, indicating viral load.

Conclusions:

• The developed smartphone-based system provides a rapid, accurate, and low-cost method for SARS-CoV-2 detection.
• This technology offers a valuable tool for COVID-19 diagnostics, potentially aiding in pandemic control.
• The quantitative readout provides additional clinical information regarding viral load, surpassing nonquantitative methods.