Abstract
Colorimetric loop-mediated DNA isothermal amplification-based assays have gained momentum in the diagnosis of COVID-19 owing to their unmatched feasibility in low-resource settings. However, the vast majority of them are restricted to proprietary pH-sensitive dyes that limit downstream assay optimization or hinder efficient result interpretation. To address this problem, we developed a novel dual colorimetric RT-LAMP assay using in-house pH-dependent indicators to maximize the visual detection and assay simplicity, and further integrated it with the artificial intelligence (AI) operated tool (RT-LAMP-DETR) to enable a more precise and rapid result analysis in large scale testing. The dual assay leverages xylenol orange (XO) and a newly formulated lavender green (LG) dye for distinctive colorimetric readouts, which enhance the test accuracy when performed and analyzed simultaneously. Our RT-LAMP assay has a detection limit of 50 viral copies/reaction with the cycle threshold (Ct) value ≤ 39.7 ± 0.4 determined by the WHO-approved RT-qPCR assay. RT-LAMP-DETR exhibited a complete concordance with the results from naked-eye observation and RT-qPCR, achieving 100% sensitivity, specificity, and accuracy that altogether render it suitable for ultrasensitive point-of-care COVID-19 screening efforts. From the perspective of pandemic preparedness, our method offers a simpler, faster, and cheaper (∼$8/test) approach for COVID-19 testing and other emerging pathogens with respect to RT-qPCR.
【저자키워드】 SARS-CoV-2, machine learning, Colorimetric RT-LAMP, LAMP-DETR, 【초록키워드】 COVID-19, pandemic, feasibility, point-of-care, DNA, RT-LAMP, sensitivity, specificity, RT-qPCR, COVID-19 testing, Concordance, pathogen, Accuracy, cycle threshold, Interpretation, Detection limit, Analysis, isothermal, observation, Perspective, offer, downstream, Complete, approach, limit, ENhance, analyzed, performed, exhibited, majority, faster, diagnosis of COVID-19, 【제목키워드】 COVID-19, point-of-care, Disease diagnosis, automated, platform, isothermal,