Simple and Multiplexed Detection of Nucleic Acid Targets Based on Fluorescent Ring Patterns and Deep Learning Analysis

Simple diagnostic tests for nucleic acid targets can provide great advantages for applications such as rapid pathogen detection. Here, we developed a membrane assay for multiplexed detection of nucleic acid targets based on the visualization of two-dimensional fluorescent ring patterns. A droplet of the assay solution is applied to a cellulose nitrate membrane, and upon radial chromatographic flow and evaporation of the solvent, fluorescent patterns appear under UV irradiation. The target nucleic acid is isothermally amplified and is immediately hybridized with fluorescent oligonucleotide probes in a one-pot reaction. We established the fluorescent ring assay integrated with isothermal amplification (iFluor-RFA = isothermal fluorescent ring-based radial flow assay), and feasibility was tested using nucleic acid targets of the receptor binding domain (RBD) and RNA-dependent RNA polymerase (RdRp) genes of SARS-CoV-2. We demonstrate that the iFluor-RFA method is capable of specific and sensitive detection in the subpicomole range, as well as multiplexed detection even in complex solutions. Furthermore, we applied deep learning analysis of the fluorescence images, showing that patterns could be classified as positive or negative and that quantitative amounts of the target could be predicted. The current technique, which is a membrane pattern-based nucleic acid assay combined with deep learning analysis, provides a novel approach in diagnostic platform development that can be versatilely applied for the rapid detection of infectious pathogens.