Department of Biomedical Engineering, Medical School, Shenzhen University, Shenzhen 518060, China.
Smart Medical Imaging, Learning and Engineering Lab, Department of Biomedical Engineering, Medical School, Shenzhen University, Shenzhen 518060, China.
Proc Natl Acad Sci U S A. 2024 Jan 9;121(2):e2314030121. doi: 10.1073/pnas.2314030121. Epub 2024 Jan 2.
Multiplex, digital nucleic acid detections have important biomedical applications, but the multiplexity of existing methods is predominantly achieved using fluorescent dyes or probes, making the detection complicated and costly. Here, we present the StratoLAMP for label-free, multiplex digital loop-mediated isothermal amplification based on visual stratification of the precipitate byproduct. The StratoLAMP designates two sets of primers with different concentrations to achieve different precipitate yields when amplifying different nucleic acid targets. In the detection, deep learning image analysis is used to stratify the precipitate within each droplet and determine the encapsulated targets for nucleic acid quantification. We investigated the effect of the amplification reagents and process on the precipitate generation and optimized the assay conditions. We then implemented a deep-learning image analysis pipeline for droplet detection, achieving an overall accuracy of 94.3%. In the application, the StratoLAMP successfully achieved the simultaneous quantification of two nucleic acid targets with high accuracy. By eliminating the need for fluorescence, StratoLAMP represents a unique concept toward label-free, multiplex nucleic acid assays and an analytical tool with great cost-effectiveness.
多重数字核酸检测在生物医学领域有重要应用，但现有方法的多重性主要是通过荧光染料或探针实现的，这使得检测既复杂又昂贵。在这里，我们提出了一种基于沉淀副产物视觉分层的无标记、多重数字环介导等温扩增方法（StratoLAMP）。StratoLAMP 指定了两组具有不同浓度的引物，以便在扩增不同的核酸靶子时产生不同的沉淀产量。在检测中，深度学习图像分析用于对每个液滴内的沉淀进行分层，并确定封装的核酸定量目标。我们研究了扩增试剂和过程对沉淀生成的影响，并优化了检测条件。然后，我们实现了用于液滴检测的深度学习图像分析管道，总体准确率达到 94.3%。在应用中，StratoLAMP 成功地实现了两种核酸靶标同时进行高准确度的定量。通过消除对荧光的需求，StratoLAMP 代表了一种无标记、多重核酸检测的独特概念，也是一种具有高成本效益的分析工具。
