Enhancing efficiency in detection of COVID-19 through AI-driven colorimetric isothermal detection with multiplex primers.

COVID-19 has afflicted millions of lives worldwide. Although there are many rapid methods to detect it based on colorimetric loop-mediated isothermal amplification, there remains room for improvement. This study aims to 1) integrate multiple primers into a singleplex assay to enhance the diagnostic sensitivity, and 2) utilize a high-throughput smartphone-operatable AI-driven color reading tool to enable a rapid result analysis. This setup can improve the sensitivity by 10-100 times and can analyze approximately 6700 samples per minute. The assay is simpler than RT-qPCR, with a turnaround time of less than 75 min. It can detect various types of SARS-CoV-2 by targeting 3 genes, increasing the likelihood that it will remain effective even if the virus undergoes mutations in any single target gene. In summary, it affords potential for adaptation to detection of new/re-emerging diseases with the visual readout for maximum assay simplicity and AI-operated mode for large-scale testing.